What Congress hasn’t done is establish safeguards that could have prevented or curtailed the fraud sooner.

Congress has been trying for years. The House twice passed market structure legislation that would have given regulators the authority to stop FTX, most recently in July 2025 by a bipartisan vote of 294 to 134. The Senate Banking Committee and Senate Agriculture Committee cleared their own version of the legislation — the Digital Asset Market Clarity Act — earlier this year. It’s about to get its first shot on the Senate floor.

Tomorrow, on September 15, the Senate votes on whether to open debate. If the bill becomes law, exchanges serving U.S. consumers will have to adopt safeguards that FTX lacked. If the bill doesn’t, those gaps will remain.

FTX did not fail because regulators missed a sophisticated scheme. There was nothing sophisticated about it. FTX simply hid that they misappropriated customer assets, because there were no independent custodian, segregated assets, or disclosure requirements. And there was no regulatory oversight ensuring those types of safeguards were followed. The firm collapsed when their fraud was uncovered and customers tried to withdraw their funds, exposing an $8 billion hole. 

The safeguards that could have prevented this have existed for nearly a century, but they simply do not extend to spot digital asset markets. Congress required futures commission merchants to segregate customer property under the Commodity Exchange Act of 1936. Broker-dealers holding customer assets are subject to custody, reserve, capital, disclosure, and examination requirements, including the Securities and Exchange Commission’s Customer Protection Rule. The Securities Investor Protection Act of 1970 provides an additional framework when a broker-dealer fails. 

Far from exotic, these are the established rules of every well-regulated market.

Which means the status quo — the thing everyone in Washington has spent years complaining about — is not a neutral state. The question is no longer whether digital asset markets will exist. It is what rules will govern them.

The CLARITY Act brings digital commodity brokers, dealers, and exchanges inside the regulatory perimeter and applies the boring, proven machinery of traditional financial regulation to them: segregation of customer property, qualified custody, restrictions on conflicts involving affiliates, mandatory disclosure, listing standards, limits on insider selling, and a named compliance officer answerable for the firm’s adherence to the law.

It settles the jurisdictional question between the Securities and Exchange Commission and the Commodity Futures Trading Commission, which today is open to expansive and weaponizable interpretation. It replaces a project’s unfalsifiable claim to be “sufficiently decentralized” with a statutory test based on control. And it imposes issuer disclosure obligations, as well as lockup periods and insider-trading restrictions, much as we already do for public stocks.

In short, CLARITY means that digital asset markets and intermediaries would have to follow rules similar to those followed by traditional markets and intermediaries already.

Three objections have kept CLARITY off the floor: first, that the idea that legislation “deregulates” crypto; second, that officials who hold crypto stand to benefit from it; and finally, that stablecoin rewards will drain deposits out of the banking system. Each deserves a fair hearing, but none is an argument for preserving the status quo.

First, that CLARITY is “deregulation” that will let crypto run wild. This rests on a false premise: that all of crypto is already subject to the securities laws. It is not, and courts have said so repeatedly. The reach of the securities laws over digital assets is uncertain, and it is not seriously disputed that many digital assets, bitcoin and ethereum among them, are not securities. 

CLARITY settles a fight between two untenable positions — that everything onchain is a security, and that nothing is. Neither has ever been true, and the cost of leaving the question unresolved falls on consumers. The Wild West that opponents fear is the thing we already have. 

The absence of a clear rulebook is also what helped FTX pass for a legitimate business. An offshore exchange with no meaningful disclosure obligations competed directly against domestic firms trying to comply with a patchwork of state requirements, uncertain asset classifications, and shifting enforcement positions. That is not a market. That is a penalty on good behavior.

Second, that officials who hold crypto stand to benefit from it. The concern is legitimate. Public officials should not profit from industries they oversee. But that is a question of government ethics, and it applies to every asset an official can own. Whether a multitrillion-dollar market should operate under federal rules is a question of financial regulation. Collapsing the two means answering the second badly to make a point about the first — and, in the process, leaving millions of market participants unprotected.

Nothing about the concern is specific to crypto. Officials trade stocks, hold real estate, and own stakes in private companies, and the conflict-of-interest rules that govern them do not turn on asset class. An ethics regime written one asset at a time invites whack-a-mole, since anyone determined to self-deal can simply route around it. If Congress believes existing rules are too weak, the fix is to strengthen them for everyone, not to hold a market structure bill hostage to a rider that would reach one asset and leave the rest untouched.

As proposed, CLARITY contains unprecedented constraints. Voting it down doesn’t constrain anyone’s holdings. Rather, it leaves them unsupervised. The bill would impose on token issuers the disclosure obligations, lockups, and insider-selling limits that public stock already carries. Today none of that exists. The market opponents describe — opaque assets with no rules —  is the status quo they are voting to keep.

Third, that stablecoin rewards will drain deposits out of the banking system. Banks argue that paying interest-like returns on stablecoin balances would create unregulated savings accounts and pull funding away from lending to households and small businesses.

That objection isn’t supported by evidence, but even if it were, the concern has already been met. After months of  negotiation, the resulting bill text bars passive yield — any return economically or functionally equivalent to deposit interest — while preserving  rewards tied to genuine activity. And the latest draft of CLARITY enables the Treasury Department to add restrictions if evidence of deposit flight does materialize. But this isn’t really about deposits. The White House Council of Economic Advisers put the lending effect of a total ban at roughly $2.1 billion — about two hundredths of one percent of bank lending. This is an anti-competitive argument masquerading as a financial stability one.

Should a version of CLARITY reach the President, it will be a compromise, because that is what legislation is. In exchange for a statutory foundation and clear rules, the crypto industry will come within the U.S. regulatory perimeter. In no case, can anyone argue that this trade is not already significantly better than the status quo. 

This isn’t a problem the regulatory agencies can solve on their own. Any commission rule is a rule only until the next commission reverses it, and one that anyone with standing can tie up in court for years. We just watched an entire industry’s legal treatment swing with a change of administration. Firms holding other people’s money should not have to build compliance systems on a framework that may not survive the next election, and institutions will not commit to investing in critical infrastructure under those conditions. Certainty is a product. Only a statute will deliver it.

If the Senate doesn’t act now, the next failure will be bigger.

When FTX collapsed, crypto was still largely a retail market at the edge of the financial system. That is no longer true. Congress passed the GENIUS Act in July 2025 and gave dollar-denominated stablecoins a federal framework. Since then, supply has passed $300 billion, transaction volumes have risen sharply, and stablecoin issuers now sit among the largest holders of U.S. government debt. But GENIUS only provides rules for the dollars moving onchain while leaving all the blockchain rails they travel on untouched.

Everything else onchain has scaled too. Tokenized assets have passed $30 billion in market value and are diversifying well beyond crypto-native products. The Depository Trust and Clearing Corporation, whose depository subsidiary custodies more than $114 trillion of securities, processed its first production transactions involving tokenized assets in July and launches its full tokenization service next month. BlackRock, Fidelity, Franklin Templeton, and Goldman Sachs all run live digital asset businesses, and all have publicly backed the bill.

Bipartisan negotiators in both chambers have already done the hard work. The Senate should finish it. Every week without a rulebook is another week in which an exchange can hold Americans’ assets without the safeguards they take for granted everywhere else.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Franklin Templeton has kept the official share register of its onchain U.S. government money fund on permissionless blockchains since 2021 and added the Solana network in February 2025. BlackRock has issued shares of its tokenized money market fund on Ethereum since March 2024. And in January 2025, Apollo made tokenized access to its Diversified Credit Fund available across six permissionless networks. Additional announcements touting traditional financial institutions’ deployment of products on permissionless networks continue on a near-weekly basis.

Yet, some TradFi institutions still treat permissionless networks as somehow unavailable. Instead, many banks, broker-dealers, and asset managers have been trending toward permissioned networks — systems in which a gatekeeper or consortium decides who may validate transactions, who may use or participate in the network, and for what purpose. These institutions are currently choosing permissioned networks because they believe — wrongly —  they must. The premise underlying that choice is that a known, verified set of actors is a precondition to compliance with financial integrity laws: the Bank Secrecy Act (BSA), with its anti-money laundering (AML) and countering the financing of terrorism (CFT) requirements; and U.S. sanctions laws. 

To put it simply, institutional compliance departments think permissionless networks are  irreconcilable with the BSA and sanctions laws.

Our new paper, “The Compatibility of Permissionless Networks and Financial Integrity: A Practical Guide for Financial Institutions,” explains that financial institutions can build and transact on permissionless blockchain networks. Concerns about financial integrity laws should not deter such usage and today can be addressed under existing law. Institutions can satisfy their obligations through appropriate, risk-based compliance frameworks that place controls at the layer the institutions operate. 

There is no precondition — regulatory or otherwise — for financial institutions to own, screen, or censor the underlying infrastructure over which their financial transactions, and the communications about them, travel. In fact, regulators explicitly have recognized that financial institutions can adapt their financial integrity compliance programs to technological innovations like permissionlessness.

What financial integrity laws require

The BSA and sanctions laws require financial institutions to have reasonable control over risks, and put controls in place to mitigate such risks. They do not require the total elimination of risk — an impossible threshold. 

Under the BSA, financial institutions’ AML/CFT programs must focus on detection, documentation, and deterrence of illicit financial activity. These programs are not intended to prevent money laundering or terrorist financing wholesale, nor can they. The federal banking regulators and FinCEN have explicitly stated that the key to financial integrity is a “reasonably designed” AML program that includes “effective processes to identify, measure, monitor and control risks.”

FinCEN stated the point more plainly in its August 2020 Statement on Enforcement, describing its approach to BSA enforcement as something other than a “gotcha” game. Treasury’s report on de-risking addressed the underlying fear directly, observing that while banks believe any failure in banking controls exposes them to substantial fines, the regulators note that such fines are rare and follow the collapse of an entire AML/CFT program, rather than the limited shortcomings a risk-based approach will sometimes produce.

The sanctions regime works on similar logic. OFAC’s Framework for Compliance Commitments identifies five essential components of an effective risk-based sanctions compliance program: 

1. management commitment, 
2. risk assessment, 
3. internal controls, 
4. testing and auditing, and
5. and training.

OFAC scales its expectations on implementation to an institution’s size, products, customers, and geography.

The Economic Sanctions Enforcement Guidelines weigh willfulness, awareness of the conduct, harm to sanctions objectives and the adequacy of the compliance program when the agency decides how to address an apparent violation. 

The agencies’ respective enforcement regimes and history support the “risk-balanced-not-zero-tolerance” approach. FinCEN and OFAC focus enforcement efforts on reasonably knowable, systemic deficiencies rather than isolated slips. This posture bears directly on the inadvertent-violation concern institutions raise about permissionless networks. 

Both AML/CFT and sanctions regimes ask for control proportionate to identified risk, which institutions can achieve on a permissionless network. Any attenuated or inadvertent violations of those regimes should not create risk for financial institutions.

Attenuated, protocol-mediated contact at the network layer

Financial institutions should understand the use of permissionless networks as akin to the use of infrastructure, much as they already treat the public internet and the telephone network. Both of these are shared systems whose other users and operators they neither know nor screen. They should calibrate their compliance approach accordingly. 

Instead, financial institutions’ hesitation to engage with permissionless networks focuses on potential inadvertent or unknowing interactions with sanctioned or illicit actors — for example, paying network fees to a validator operated by a sanctioned actor, unknowingly transacting with a sanctioned actor, or receiving or transacting with cryptoassets that may have touched illicit actors at some point in their history.

But inadvertent, unknowing interactions with validators or other network participants in a sanctioned jurisdiction are not the kind of activity the sanctions laws were written to address. The concern is broader than geography, since a validator could be a designated person operating from anywhere: The institution has not selected, contracted with, exported to, financed, or otherwise dealt with the operator, and the fee reaches it through rules that apply identically to every user of the network.

Regulators have confirmed this. For example, in November 2025, the OCC addressed this concern by issuing Interpretive Letter 1186, confirming that a bank may pay network fees on blockchain networks and may hold as principal the crypto-assets needed to pay them. The letter reasons from Interpretive Letter 1174 of January 2021, in which the OCC concluded that a bank may validate, store, and record payment transactions by serving as a node. Accepting the fee paid to a node follows from that conclusion. The letter uses the example of the Ethereum blockchain, a permissionless network whose protocol selects validators pseudo-randomly. None of the letters in this line distinguishes permissioned networks from permissionless ones.

When an institution submits a transaction via a permissionless network, the protocol assigns the right to propose the block that will include that transaction to a unique validator, generally pseudo-randomly and in proportion to stake. Protocol-defined rules set the fee as a function of network demand and the computational resources the transaction consumes. As a result, an institution does not choose the validator that will process its transaction. It cannot bargain over the fee, and it has no way to learn who the validator is before or after the transaction. Every other user of the network transacts under identical rules.

This bears some resemblance, as mentioned above, to the relationship between an email sender and the owners of the routers that carry the message, or between a caller and the owners of the switches that complete the call. A U.S. financial institution whose internet protocol packets traverse infrastructure in a sanctioned jurisdiction is not considered to violate sanctions on that basis: The same analysis of neutral, protocol-mediated carriage applies to a permissionless network’s consensus layer. That distinction is codified in the BSA’s own regulatory definition, which expressly excludes those who “merely provide the delivery, communication, or network access services used by a money transmitter to support money transmission services.” The BSA also distinguishes neutral carriage from transacting, and the sanctions analysis turns on the same feature of the relationship, the absence of selection, direction or dealing.

Although an institution transacting on a permissionless network does have some contact with operators it has not screened, that contact is different from what sanctions laws police: In the latter case, no party involved selected any other. To put this in perspective, in the nearly five years since OFAC published its Sanctions Compliance Guidance for the Virtual Currency Industry, no enforcement action has been predicated on a validator having proposed a block that happened to contain a sanctioned party’s transaction, and none has rested on a market participant’s payment of protocol-level fees.

Privacy and compliance are compatible

The second concern institutions raise is privacy: Can a bank transact on a public ledger without exposing client positions, counterparties, and strategies to its competitors?

The early case for permissionless ledgers rested on total transparency as a compliance asset. Financial integrity requires something narrower: that the necessary information be verifiable by the institution, its counterparty, and its regulator or supervisor. Cryptography has advanced far enough that an institution can prove a compliance-relevant proposition without publishing the data that establishes it: that a counterparty sits outside the Specially Designated Nationals (SDN) List, for instance, or that reserves exceed liabilities, with the book and the counterparty’s identity remaining undisclosed. Proofs of provenance permit a party to show that an asset never came from an identified illicit set without exposing its transaction graph. Confidential-transfer designs encrypt amounts and balances on the ledger while retaining a viewing key an institution can furnish to an examiner. 

Together, these cryptographic advances give a supervisor better assurance than a closed system provides. And it does so while giving a competitor nothing at all, which turns privacy from an objection into a reason to build on permissionless networks.

Some of these techniques are in production today; some are still in the research and development phase. Address rotation and account abstraction are in production, as are omnibus and tiered custody structures that keep client-level detail off the ledger and the messaging protocols that carry Travel Rule data alongside an onchain transfer. While confidential transfers with an auditor key are shipping, they are currently thinly used at institutional scale. Proofs of unsanctioned status and provenance proofs against a designated set remain in pilot testing and research. Yet solutions do exist: Privacy Cash, for example, is a privacy protocol that sits on top of Ethereum and Solana and uses zero-knowledge proofs to enable confidential transfers and swaps. 

A risk-management framework for permissionless networks

We set out nine components of a financial integrity program adapted to permissionless-network activity: Transaction-level controls that apply to the institution’s customers and counterparties in largely the same form they take today, and network-level controls address the infrastructure itself.

These controls do not require identification of validators, service-level agreements with a protocol, or petitions to a gatekeeper for membership, among other features of permissioned networks: The current financial integrity laws do not require any of this. 

The proposed framework is also consistent with recent U.S. legislation, the GENIUS Act. GENIUS likewise adopts a framework where AML/CFT and sanctions controls sit at the application layer, operated by entities with knowledge of customers and control over assets. GENIUS requires permitted payment stablecoin issuers, as identifiable regulated entities at the application layer, to certify the existence of AML programs and sanctions compliance programs, and to maintain the technical capability to execute lawful orders to freeze or burn outstanding stablecoins. These obligations run to the issuer, not to the permissionless networks on which stablecoins circulate.

***

A generation ago, regulated financial institutions confronted an open, global, permissionless network that anyone could join and that carried the traffic of legitimate and illegitimate users alike. Institutions moved their businesses onto the open protocols of the internet and built their controls at the application layer. The same can be done now for permissionless networks.

Avoiding permissionless networks is not a financial integrity strategy; it is an abdication of the role that U.S. institutions have always played in making the dollar-based financial system resilient, data-enriched, and risk-based. And dollar-denominated activity on permissionless networks is already happening and will continue,whether or not U.S. institutions take part. The reach of U.S. financial enforcement rests on line of sight into financial flows, and the architecture behind the financial integrity laws — their implementation and enforcement —  depends on U.S. institutions observing the activity they are asked to monitor.  To not participate in the innovation and benefits borne by permissionless networks — based on a misinterpretation of relevant laws or concerns based on past regulators’ positions — unnecessarily limits the choices that traditional financial institutions make in serving the needs of their customers. 

***
Rebecca Rettig is the Chief Operating Officer and Chief Legal Officer of Jito Labs, a technology startup that develops Solana blockchain software and infrastructure.

***

This article is adapted from The Compatibility of Permissionless Networks and Financial Integrity: A Practical Guide for Financial Institutions, by Rebecca Rettig, Omid Malekan, and Michael Mosier.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

• It makes Jolt post-quantum.
• It makes the prover and verifier 2x-3x faster.
• It gives Jolt the shortest proofs of any post-quantum zkVM: under 100 KB today, with a path to further reductions. Proofs get posted on-chain and sent over networks, so smaller proofs mean cheaper verification everywhere.

Those properties hold across the whole range of settings where zkVMs get used. The same prover that handles billions of CPU cycles on GPUs can also prove millions of cycles on a phone, and in both cases developers write ordinary programs rather than hand-crafted circuits requiring special expertise. That is what we mean when we call Jolt the “everything SNARK”.

But the bigger story is what Lattice Jolt means for SNARK design and adoption. Nearly every post-quantum SNARK in production today is hash-based. Lattice Jolt shows that lattice-based SNARKs can be both faster and more compact. Digital signatures are going through the same transition: Hash-based schemes are the conservative choice, but lattice-based ones are what the world is mostly deploying. We expect SNARKs to follow, and the second half of this post explains why.

Replacing elliptic curves with lattices

Jolt’s previous polynomial commitment scheme, Dory, was the one component of the system that depended on elliptic-curve cryptography. Lattice Jolt replaces Dory with Akita, a new polynomial commitment scheme based on the lattice assumption Module-SIS. Lattice Jolt targets a full 128 bits of security based on this standard, well-studied assumption.

This is the same family of assumptions — Module-SIS and its sibling Module-LWE — that the world’s digital infrastructure is now migrating to. These assumptions underlie not only the digital signature standard ML-DSA, but also the key-establishment standard ML-KEM, which is already the most widely deployed post-quantum primitive in the world.

Akita’s development and implementation was led by researchers and engineers at LayerZero, in collaboration with researchers at Carnegie Mellon University, the University of Southern California, and our engineering and research teams here at a16z crypto.

How Lattice Jolt is faster

Not only is Lattice Jolt post-quantum, it’s faster than the elliptic curve-based version it replaces.

The speedup primarily comes from a simple source. Elliptic curves forced Jolt to work over 256-bit fields; lattice cryptography achieves comparable security over 128-bit fields. The Jolt prover’s work is dominated by multiplying field elements (essentially, very large numbers) — so halving their size makes each multiplication several times faster.

Jolt with Dory was already fast: Our last performance update on Jolt reported about 700,000 RISC-V (RV64IMAC) cycles per second on a laptop, and optimizations since then have pushed curve-based Jolt past 1 million cycles per second.

Lattice Jolt proves over 2 million cycles per second on the same machine.

Much of the past six months went not only into developing Akita and integrating it into Jolt, but also into rewriting the Jolt codebase from scratch. Jolt already ran well on GPUs, but the rewrite makes GPU implementations far easier to build and optimize.

The first payoff is an Apple Metal implementation, which brings large speedups to Apple hardware. (Metal is Apple’s framework for running code on the GPUs built into its devices, including MacBooks and iPhones.) —

• GPU-accelerated Lattice Jolt proves over 10 million RV64IMAC cycles per second on a MacBook.
• CPU-only Lattice Jolt proves over 2 million cycles per second on the same machine.
• Even curve-based Jolt now runs at about 4 million cycles per second on a MacBook with Metal.

In other words, in a single release, Jolt on a MacBook has gone from about 1 million cycles per second (curve-based, CPU-only) to over 10 million (lattice-based, with Metal).

To put those numbers in perspective: When we first wrote about SNARK prover overheads four years ago, proving a computation cost millions of times more than simply running it. Lattice Jolt brings that overhead down to roughly 10,000x. And this is still not the end; clear engineering and protocol optimizations remain on the table.

Proof size matters just as much as prover speed. At under 100 KB, Lattice Jolt proofs are already substantially smaller than those of other post-quantum zkVMs, whose proofs range from over 200 KB to roughly 600 KB or more.

The switch to lattices also improves Jolt’s already best-in-class memory usage, dropping the prover’s space usage from about 300 bytes per cycle to 200. This means you can prove millions of RISC-V cycles on a mobile phone.

A companion paper, coming shortly, adds zero knowledge to Lattice Jolt, a property required for privacy applications.

The case for lattices over hashing

For years, the SNARK community’s attention — and essentially all of its production deployments — have gone to hash-based SNARKs as the path to post-quantum security.

But there has also been a sustained line of work on lattice SNARKs and lattice commitments — spanning LaBRADOR, Greyhound, LatticeFold, SuperNeo, and Akita’s immediate predecessor, Hachi. Lattice Jolt builds on that research, bringing a lattice commitment layer into the architecture of a high-performance zkVM while showing that lattice-based SNARKs have unmatched speed and compactness.

This should not be surprising. As mentioned, the same pattern has already played out for digital signatures.

Cryptographers have built signatures from many assumptions. Hash-based signatures are often viewed as the most conservative: their security assumptions are simple and old. But the world is moving primarily toward lattice-based signatures because they are shorter and faster —

• ML-DSA signatures are on the order of a few kilobytes.
• The NIST-standardized hash-based alternative, SLH-DSA, is several times larger.
• For encryption and key exchange, the case is even clearer: No hash-based option exists (provably so), and post-quantum deployment is overwhelmingly lattice-based. ML-KEM — the primary key-establishment standard NIST finalized in 2024 — is already deployed by default in major browsers, messaging apps, and a large share of TLS connections on the web.

The analogy between SNARKs and signatures isn’t superficial. A digital signature is essentially a proof of knowledge of a secret key, tied to an authorized message. SNARKs extend this paradigm from one narrow statement to arbitrary computation. So it would be surprising if the long-term cryptographic landscape for SNARKs looked completely different from the landscape for signatures and encryption.

There is also a misconception worth clearing up here: Hash-based SNARKs are often described as the conservative post-quantum choice because they are “just based on hashes”. That is true only if the underlying hash function is non-algebraic.

Today, most hash-based SNARK deployments rely on SNARK-friendly algebraic hash constructions like Poseidon to make it cheap to prove that a hash was evaluated correctly, which is especially important for recursion (note here recursion means proving you hold a valid SNARK proof). Those constructions are far more structured than standard hash functions, and their cryptanalysis is much less mature.

In short, we do not have confidence in the security of algebraic hash functions. Despite this, they are used pervasively in production SNARK systems today. (Though, in a sign of progress, the Ethereum Foundation recently announced that it is moving away from them.)

Algebraic hashes are not the only hidden assumption in deployed hash-based SNARKs: Many have historically set concrete security levels using conjectured proximity-gap bounds rather than fully proven ones. Some of the strongest have turned out to be false.

Even hash-based SNARKs that avoid the above conjectures often target fewer than 128 bits of security, because a full 128 bits comes with significant performance overheads. Why? Hash-based SNARKs can’t reach 128 bits of security over a 128-bit field, owing to soundness error that scales like n/|F|, where n is roughly the size of the statement being proved and |F| is the field size. So over a 128-bit field, proving a billion-step statement loses about 30 bits of security, dropping under 100 bits. Instead, Lattice Jolt’s soundness error scales like log(n)/|F| — keeping nearly the full 128 bits of security over the same field (the small log(n) loss can be restored with standard techniques).

Ironically, some systems marketed as the “conservative” post-quantum choice in fact rely on algebraic hash functions, conjectural proximity-gap bounds, and a sub-128-bit target security level all at once. So while hash-based SNARKs are an important direction, they are not automatically the low-risk option that many believe them to be.

One Jolt, three foundations: curves, lattices, and hashes

Our view has always been that Jolt should not be tied to one cryptographic foundation. We should have mature, performant SNARKs based on curves, hashes, and lattices. Different assumptions and performance profiles will be appropriate for different settings.

But if digital signatures are any guide, lattice-based SNARKs will become the most deployed post-quantum option.

Jolt is unusually well positioned for this transition. The original Jolt design exploited properties of elliptic curves that are especially useful for commitments, including fast commitments to sparse vectors. Lattice commitments have the same property: Committing to a vector is cheap when most of its entries are zero or small, and Jolt commits almost exclusively to such vectors. This property made it possible to swap out Dory for Akita while leaving the rest of Jolt intact.

We will build a hash-based version of Jolt. But a hash-based version — compared to curve-based and lattice-based Jolt — will have worse space efficiency, bigger proofs, and a variety of complications. This is because the most promising hash-based SNARKs work over binary fields, number systems that are convenient for proving hash evaluations but mismatched with how CPUs do arithmetic. That mismatch makes proving ordinary CPU multiplications expensive. Even so, the ecosystem should have zkVMs under every major assumption family, just as it has for digital signatures.

The everything SNARK

Lattice Jolt is everything builders need from a zkVM at once: post-quantum, transparent, fast, compact, and space-efficient. It brings the lattice-SNARK line of work, from LaBRADOR to Hachi, into a production zkVM without giving up anything that made Jolt fast in the first place.

Our goal is to not only release the most performant zkVM as open source for anyone to use, but to largely eliminate the need to hand-tune SNARKs for particular applications. This doesn’t require Jolt to be as fast as hand-tuned provers — an impossible goal analogous to demanding a CPU match a purpose-built ASIC at every task. It only requires Jolt to be fast enough for an acceptable user experience.

For the “small” statements relevant to client-side proving (where hand-optimized circuits dominate today), the key bar to clear is generating proofs on a phone in under a second or so. Jolt is on the edge of achieving this already, with plenty more speedups in the pipeline.

The era of lattice SNARKs is upon us.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Business was static. It started; it stopped. 

The internet changed that. Amazon transformed your local bookstore from a physical shop with business hours to an always-open global storefront for every imaginable product. Slack, mobile phones, and Zoom transformed the commute and the business day into relics of the past. Stripe turned starting a business and accepting payments from a complex legal process into a simple API. You can now start a business with just a credit card and an internet connection. Consumer fintech products like PayPal made payments feel faster and new investing platforms like Coinbase and Robinhood made it easier for customers to participate in economic growth. 

Through all the positive changes that the internet has brought to business in the last 30 years, banking is more or less the same. Your bank still opens and closes. Money still moves at the speed of ACH and wire transfers. Some of the most essential financial services aren’t available on weekends, despite the global economy moving faster than ever before. Finance is stuck in the past. 

While fintech changed the front end and UX of financial services, banking remains the bottleneck on the back end. Most fintech products are built on the same small group of sponsor banks, many of which run on outdated systems and processes. These banks do their best to accommodate the demands placed on them, but they struggle to provide the programmability and feature set that modern consumers and businesses expect from their software.

We need a bank that’s always on, always open, always available. A bank that never closes. A continuous bank. 

Banks are slow and static — keeping limited hours — for a reason. They run on much the same atrophying technology stack that they did before the internet changed our world. Banks are fundamentally ledgers, and the internet gave us a better kind of ledger. The internet gave us blockchains. 

We need a bank built like a blockchain, the financial infrastructure with the strongest resilience, uptime, and accessibility on the planet. We need a bank with instant settlement and support for atomic transactions. 

We also need a bank with the experience, compliance, and security to stand the test of time. 

OpenReserve is building this bank: the bank for the internet age. Traditional banks create credit, but cannot operate natively onchain. Stablecoins move money continuously, but do not create credit. OpenReserve combines continuous money with regulated credit. 

Venture capital hasn’t historically focused on banks. The VC model typically relies on technology to increase productivity growth and drive outlier returns. Historically, banking was not driven by technological innovation. We believe that’s changed. Stablecoins are better money, and can enable faster, cheaper, and more transparent money movement. We also believe there is greater operating leverage in banking today due to blockchains as more efficient infrastructure and AI agents as faster and lower cost intelligence. Bankers should be forward deployed, spending time with their customers instead of buried in excel or even literal paperwork for processes that are better automated in software. 

While stablecoins can solve money movement, they don’t address credit. Part of the reason we believe OpenReserve is purpose-fit to improve banking is its combination of deep lending expertise, customer empathy, and knowledge of onchain finance. We are excited about the potential to launch new onchain vaults backed by the guarantees only a federally chartered U.S. bank can make. We’re also excited about a bank core built like an L1, and the potential for developers to build new protocols and products on top of OpenReserve in the same way Ethereum and Solana served as a substrate for onchain finance. 

It isn’t easy to start a new bank. After the great financial crisis the number of de novo bank charters plummeted, and has only recently begun to increase. We believe Americans and companies worldwide will benefit from more banking options, and the technological progress that new banks like OpenReserve represent. 

OpenReserve is a bank powered by a programmable core ledger. It is designed for the continuous settlement that tokenized deposits and stablecoins make possible. Continuous settlement means both always being open, and the support for instant money movement.  OpenReserve represents a regulated balance sheet capable of creating and distributing credit onchain. And it’s run by an experienced team who have run financial services at scale. 

OpenReserve was founded by Dee Choubey. Dee immigrated to America from India as a small child, his ascension up the ranks of traditional finance culminated when he founded his first company MoneyLion in 2012. He took MoneyLion public and ultimately sold it to Gen Digital (formerly Symantec) in 2025. Dee has a rare combination of true depth in financial services and the technological expertise to change the future of finance. He is an experienced leader and lender. He innovated first in consumer lending based on alternative data and then scaled the operation to lend billions in total during his time leading MoneyLion. Dee is joined by Rick Correia, who held group COO roles at both Merill Lynch and Citadel before working with Dee at MoneyLion for almost a decade.

When I first met Dee through my friend and fellow a16z General Partner, David Haber, it was immediately clear he was inherently motivated to succeed. Our partner Alex Rampell will sometimes talk about investing in a “Count of Monte Cristo,” or a second time founder who got close to their ultimate goal, and had strong success, but ultimately had greater ambitions. 

We believe Dee’s second act could eclipse his first, but it has also become clear Dee is inspired by more than personal success. At MoneyLion most of their customers were households with $50-150,000 in annual income, and 95% of investment accounts were first time investors. OpenReserve represents not just a more accessible financial system, but a chance to make core services like banking and lending more efficient for all customers. 

We are proud to partner with Dee, Rick, and the entire OpenReserve team, and to lead their $25M seed round. 

***

Finance is one of the most obvious use cases for blockchains. It’s also one of the most demanding. A benefit of blockchains, often referenced in the context of financial use cases, is that they bring assets, ownership records, and execution rules into a shared system that no single counterparty controls. This does not, however, make blockchains automatically ready for the scale and constraints they’d have to contend with in global financial markets.

For years, debates around blockchain readiness focused on cost and throughput (how many transactions a chain can process per second) because financial applications can’t scale without that capacity. The conversation is changing, however, as throughput is becoming less differentiated across many production systems.

As financial institutions start executing trades onchain and issuing stablecoins and tokenized assets, they need to know whether these networks can actually handle the demands and expectations of a real financial market. This includes reliable access, predictable transaction processing rules, and control over if and when sensitive information becomes visible. Those guarantees must also hold through congestion, outages, and attacks.

Blockchain designers have worked on these issues for years. This post looks at what they’ve solved, what remains unresolved, and why blockchains are now better equipped to support financial markets.

Predictability: guarantees on inclusion and ordering

For years, capacity was an obvious bottleneck. Table-stakes issues like too few transactions, too much congestion, and too-high fees made many real-world financial use cases hard to take seriously. Without capacity, routine actions, like placing a trade or updating an order, get expensive or unreliable the second activity picks up.

This constraint has shifted over the last few years, and especially in the last cycle of infrastructure upgrades. Across the industry, aggregate blockchain throughput has grown more than 100x in five years, and some production systems can now process tens of thousands of transactions per second. This is only the beginning as current stress tests and new architectures may push these numbers even higher.

The challenge now is that throughput alone doesn’t say much about when a transaction will land. This is where predictability comes in.

Transactions need to land on schedule, and under rules participants can understand ahead of time, because finance is full of actions that are only useful within a very specific window. For everyday payments, a second of uncertainty is probably fine, but for markets, every millisecond can decide whether participants trust the venue at all. Take an onchain order book, for instance. If a trader’s cancel order lands too late, someone can trade against a stale price. Market makers price in that risk by widening spreads, which means worse prices for everyone.

In other words, predictability requires two kinds of guarantees: (1) guarantees on inclusion — or censorship resistance — and (2) guarantees on ordering.

Resilience: reliable access to execution

Even with good ordering rules, what if one actor can still control access to execution? Imagine we selected a randomly staked node worldwide to control transaction ingress into NYSE every second. Clearly these nodes could exert extreme market power over traders on the exchange.

Sequencing determines an order’s place relative to other orders, once it hits the chain, while resilience starts a step earlier. So, can a participant submit an order in the first place — and can they do so without relying on a single gatekeeper or operator whose decisions could affect price, risk, or trade outcomes?

This is why blockchain developers are moving toward stronger guarantees that emphasize immediate, timely inclusion rather than eventual inclusion. The goal is that if a valid transaction reaches the network on time, it should be included instantaneously. That is a much stronger standard than “your transaction will get in soon,” and it is what financial markets need.

Our researchers introduced the concept and measure of Strong Chain Quality, where access to the next round of execution doesn’t depend on one operator’s queue. The premise is that a blockchain can reserve portions of each block for transactions surfaced through other parts of the network, giving valid transactions more than one route in. The result is less single-operator control over who gets processed next.

The goal here is to make blockspace feel like shared infrastructure with multiple routes in, as opposed to a single queue. This requires more complicated protocols than we run in production today, but the benefits may be worth it, so participants can be confident their transactions will land even when the network is under pressure.

Ordering: rules markets can reason about

Predictability is also about how a venue decides the order of transactions. Once transactions are ready to be processed, markets need clear rules for sequencing them.

On many blockchains, transactions don’t become final one at a time. Instead, they’re gathered into batches, or “blocks,” and added to the ledger by a block builder or proposer. In many of today’s systems, that party has quite a bit of discretion over which transactions are included and in what order. This design — meant to simplify the protocol’s life — may be acceptable for some applications. But in financial markets, ordering determines who gets priority, what price someone receives, and whether execution is fair.

Moreover, if one party controls the next block, they may be able to react to transactions before everyone else. In crypto, this is often called MEV — or value obtained by including, excluding, or reordering transactions in a block, in addition to standard block rewards. MEV is often explained through examples like “sandwich” attacks, where a trader sees a pending transaction, trades ahead of it to move the price, and then trades immediately after to profit from the resulting price movement. The institutional version is simpler: whoever controls the flow can potentially make decisions based on it before anyone else can.

This starts to look a lot like the flaws of existing financial market structure, where centralized intermediaries can benefit from privileged access to information about order flow and trading intent. If onchain markets cannot provide strong guarantees around ordering and execution rules, they risk recreating the same asymmetries.

Blockchain protocol designers are working on more explicit ordering rules, primarily including deterministic rules based on priority fees and application-specific rules set by a trading venue. The exact mechanisms are still an active area of research. Still, the principle is familiar to TradFi: market participants have precise guarantees on how their transactions will be ordered and executed, similar to price-time priority offered by traditional venues.

Inclusion and ordering have to work together. A fair sequencing rule isn’t helpful if an auction bid or a liquidation can be selectively delayed before the rule is ever applied. Likewise, timely inclusion is less useful if participants can’t reason about how their transaction will be treated once it gets in.

High throughput makes onchain finance plausible. Predictable inclusion and clear ordering rules make it usable when timing matters and under real market pressure.

Privacy: protecting intent before execution

A participant’s trades can divulge what they know or intend to do. The size and direction of an order, for instance, can show whether a fund is building or unwinding its position. If this information becomes public before the order is fully executed, other traders will have a chance to move the price.

This is especially relevant onchain because many blockchains expose pending transactions before they’re finalized. So for finance, we need pre-execution privacy, even when post-execution privacy isn’t technically achievable.

The goal for today’s protocol designers is to keep transaction details private while it’s still possible for people to exploit them, and also to support a wider range of privacy models after execution. Once the trade is complete, the disclosure rules can change. A public market might publish the full record, while a bank might make it visible only to the counterparties and regulators. People who can’t see every detail should still be able to verify that the trade followed the rules.

To that end, protocol designers are exploring different types of encryption (e.g., timelock encryption and threshold encryption), which can hide transaction contents until a deadline passes or until a committee confirms that the transaction’s place in the block is final. So by the time the market can see the transaction, it should already be too late to front-run it.

Markets can be transparent without giving every participant a live preview of everyone else’s unexecuted orders. Onchain finance needs the same sort of distinction, enabling both auditability after execution and confidentiality before execution.

If blockchains are going to become core financial infrastructure, they’ll need to support markets that participants can trust.

Throughput was just the first test. Without enough capacity, onchain financial applications could never scale beyond niche use cases. Now that speed is the baseline, the harder test is how blockchains can meet the standards financial institutions already expect from market infrastructure. These include predictable execution, unambiguous priority rules, resilient access, and confidentiality before execution without sacrificing transparency after the fact.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Moving money shouldn’t be this hard. The scale of the problem: $161 billion was remitted to Latin America and the Caribbean in 2024, roughly 80% of it from the U.S. A lot of this still moves through physical cash networks. 

Stablecoins offer a solution. Dollars can now move onchain in seconds, for lower fees than the banks charge, with none of the waiting. Because settlement is near-instant, companies don’t have to park capital in accounts in every country where they pay out. Stablecoins are uniquely suited to making cross-border transactions because the rails are better. But consumers don’t care about rails. They care about products. The winner here will likely be whoever makes sending money home feel as easy as sending a message.

Félix is doing just that. They have built an AI financial companion. They have looked to recreate the comfort and familiarity that comes with knowing your local banker and providing that same connection inside of one of our most personal apps on your phone – WhatsApp. 

Why WhatsApp? Well simply put, it’s the dominant messaging app in Latin America and one that more than half of Hispanic adults in the U.S. already use. Félix’s conversational AI agent handles the details — onboarding, transaction processing, customer support — and helps recreate the familiar experience that many immigrants grew up with when sending money through a person at a counter. On the backend, Félix settles most transactions in USDC and converts to local currency through a network of payout partners. 

Customers do not need to interact with crypto directly. They care that the money arrives quickly, the price is clear, and help is there when they need it.

Manuel Godoy and Bernardo García met as MBA students at Wharton, where they pitched an early version of the WhatsApp remittance flow at a startup competition and won. Both are immigrants who lived the problem they are now solving. Godoy, a Venezuelan engineer trained at Caltech, started Félix after watching his mother struggle to get money to family in Venezuela. García, from Mexico, had worked on market expansion at Uber and in banking analytics at McKinsey. They have since assembled a leadership team drawn from Google, Lyft, Airbnb, and Uber.

The product already has excellent traction – including processing over $8 billion, and serving six million people in eleven markets. Most of Félix’s new users arrive organically, through word of mouth and referrals rather than paid marketing. In consumer fintech, there’s no clearer sign that a product is working better than what came before it.

Remittances are just the first act toward a larger opportunity. U.S. Latinos produce roughly $4 trillion in economic output a year; if they were their own country, they would be the world’s fifth-largest economy. Despite this, U.S. Latinos remain underserved by the traditional finance industry. Félix is poised to expand into areas like credit and savings, layering a full financial platform onto its existing relationships. Stablecoin infrastructure makes this even easier to do, allowing the service to add lending, savings, and yield products in a more efficient and consumer friendly way, as compared to the current banking system.  

We are proud to lead Félix’s Series C equity raise, and to partner with Manuel, Bernardo, and the entire Félix team as they build it.

***

It happened fast. Downloads of Argentina’s 15 leading crypto apps nearly doubled in 2024, rising 93% from the previous year.

Argentina’s preference for dollars predates crypto. Back in 2001–2002, the government froze bank deposits and forcibly converted dollar-denominated deposits and loans into pesos via Decree 214/2002. After the dollar peg ended, the exchange rate fell from one peso per dollar to nearly four, wiping out roughly three-quarters of the peso’s dollar value. The crisis deepened distrust of the peso and reinforced the habit of keeping savings in physical dollars outside the banking system — think bills under mattresses or in safe-deposit boxes.

Stablecoins started to pick up steam in Argentina after the government reintroduced currency controls in 2019. Within months, the government had limited Argentines to $200 in official dollar purchases each month, while additional eligibility rules excluded many people entirely. Dollar-pegged stablecoins became another way to save in dollars without relying on the official market.

More recently, stablecoins have become a larger part of contractor pay. Year-over-year inflation reached 289% in April 2024, and the share of Argentina-based contractors receiving pay in USDC rose over the same period.

The above data comes from a16z portfolio company Deel, which helps run payroll in more than 160 countries. Using this data as an indicator, we can see both the share of Argentina-based contractors paid in USDC each month and year-over-year inflation.

Since the two metrics are indexed to January 2024, we’re seeing how each metric changed from that point, rather than the raw values. For a while, the two metrics appeared to move together. Then inflation eased, and stablecoin usage appeared to fall with it. As of July 2026, both were holding at about a fifth of their respective peaks.

“Buying crypto” with pesos means “buying dollars” in Argentina. 94% of peso crypto trading goes to stablecoins — the highest stablecoin share of any major currency Artemis tracks.

What’s interesting is that for several years, the crypto dollar cost much more than a dollar bought at the official exchange rate.

By 2023, the country’s capital controls made official dollars inaccessible to many Argentines and pushed the gap between the official and parallel exchange rates above 100%. Stablecoins became an alternative because they were available around the clock and weren’t affected by the controls. After Argentina lifted most restrictions on individuals buying dollars in April 2025, the rates largely converged.

As of August 28, 2026, a digital dollar costs about 4% more than one bought through the official market.

The economic crisis in Argentina seems to be cooling for now. Inflation is down, dollars are legal to buy, and the pressures that originally drove many Argentines to stablecoins have eased. You’d expect usage to fade with them. But it hasn’t.

Paycheck usage leveled off instead of disappearing. And downloads of Lemon, one of Argentina’s largest crypto wallets, climbed every quarter, even as monthly inflation fell from 25.5% to 2.1%.

Stablecoins may no longer be just a hedge against inflation for Argentines — they could be becoming a habit.

Acknowledgments: Thanks to Ryan Holloway for data analysis and feedback.

Alejandro Flores is an editorial intern at a16z crypto.

***

This article was translated into Korean by Token Post, available here.

Amid the narrow lanes teeming with hack writers and failed scribblers, you might find someone like Samuel Boyse — a drunken Irish poet who had fled Edinburgh because of his debts. His specialty, doggerel. He would happily compose you some for a shilling or two. If you had deeper pockets, he could even gin up a pamphlet that really laid into your target.

Boyse lived in a garret. He struggled with alcoholism. Once he found himself in such dire straits that he was forced to pawn the shirt off his back to buy the paper necessary for his craft. (Ah, the life of the writer!) Boyse’s situation was not unique — a writer as highly regarded as Samuel Johnson is alleged to have spent some time on Grub Street — although Boyse was one of the more prolific practitioners of his craft.

Grub Street produced mountains of what we today might call “slop”: single verses, pamphlets, periodicals, newspapers — all for sale in both its production and consumption. Inevitably, perhaps, this drew comparison of such writers to strumpets, and critics equated writing for pamphleteers to whoring. No respectable person would take part in such a dreadful enterprise … except that lots of respectable people did, even while decrying it to their equally titillated friends.

What I’m saying is, slop is hardly a new phenomenon. Neither is the moral panic around it. AI-generated content is the latest iteration of an age-old trend: Whenever technology dramatically lowers the cost of producing media, it unleashes a flood of cheap, derivative work — and a chorus of critics convinced that culture has reached a new low. After Grub Street came the penny press, then pulp fiction, television (a “vast wasteland” if you’re feeling fancy, or, more colloquially, the boob tube), followed by blogging, social media, and now AI.

Each time, the cost of execution falls, and the masses jump in. Each time, criticism comes hot and fast. And each time, the answer from the self-proclaimed cognoscenti is obvious: Claims to taste become their refuge, and taste achieves apotheosis. While AI may differ in degree and perhaps even in kind, the cultural response to falling production costs has recognizable historical precedents.

But this immediately raises the question many are debating: What exactly is taste? If taste consists of stable criteria that experts consistently apply, then a modern AI can almost certainly parrot taste. A model trained on enough examples of Bauhaus design can infer the principles that make an object feel Bauhaus. A model trained on a literary agent’s acquisitions can approximate that agent’s editorial preferences. Taste as the consistent application of discoverable rules is precisely the kind of pattern-recognition problem modern AI excels at.

If, on the other hand, “taste” isn’t reducible to stable criteria, then we’ve been using the word far too casually. The idea that taste simply means knowing what’s “good” borders on the tautological, while reducing it to personal preference — I like this; I don’t like that — is too subjective to explain why some people repeatedly identify promising ideas, products, or creators before everyone else.

So what are we really talking about when we say someone has taste? Prediction? Judgment? Social position? Experience? Luck? “Taste” is sometimes all of these: it’s a folk category into which we’ve collapsed separate mechanisms. AI forces us to disaggregate them. Drawing on social science can help, not in giving us a single theory of taste but in teasing disparate ideas apart. Once we identify the separate strands, the question shifts from whether machines have taste to which mechanisms that produce good human judgment remain scarce when production becomes cheap.

Success is socially contingent

One body of research that may help illuminates how songs become hits.

In the early 2000s, Columbia University sociologists Duncan Watts and Matt Salganik, together with a team of researchers, created an artificial online “music lab.” Thousands of participants were invited to listen to and download songs by unknown bands. The researchers divided listeners into eight separate groups, each with a different condition of how the songs were presented. In some, participants could see how many times each song had already been downloaded; in others, they could not. The team also manipulated the information in some “worlds” (what they called their experimental conditions), showing, for instance, a particular song at #1 when it was not.

When social information was hidden, downloads were evenly distributed. But when people could see what others were choosing, outcomes diverged. In these manipulated conditions, a song that became a runaway hit in one world performed poorly in another.

The larger lesson is that once quality clears a basic threshold, success is heavily shaped by social influence and path dependence. This makes it almost impossible, even for informed observers, to predict in advance which products will rise to the very top. Put another way, experts in a domain — TV executives, for example — are pretty good at filtering out obvious failures before they happen. But they cannot reliably predict breakout successes among the remaining strong candidates. Or as Salganik has said, the success of Harry Potter was a fluke.

Success is not necessarily dictated by quality alone. Therefore taste cannot simply mean predicting success.

Taste is brokerage

Another possibility is that what we call taste isn’t prediction at all. Maybe it’s the function of a particular social position: standing somewhere that gives you a unique vantage point to see how the world could fit together, or how ideas and products could translate from one arena to another.

This is why some of the most innovative employees in companies tend to be the ones most connected to different groups — or that’s what Ron Burt, another sociologist, found. Burt calls this “brokerage.” Brokerage isn’t the clairvoyance necessary to pick winners but a structural position that generates novel recombinations.

Burt’s conclusions grew out of decades of research mapping social networks inside organizations. Rather than focusing on individual brilliance or expertise, he asked how people’s positions within those networks affected their performance. First, Burt reconstructed who communicated with whom. He then compared those patterns with measures like promotion, compensation, and managers’ evaluations. He found that the most innovative employees were rarely embedded in a single, tightly knit community. Instead, innovators (might we call them those with taste?) occupied “structural holes”: the gaps between otherwise disconnected groups.

Because they moved in different circles, brokers encountered information that others didn’t and, even more importantly, saw problems from multiple perspectives. Their position allowed them to recombine ideas that no one network could have produced on its own. Here, innovation isn’t a matter of exceptional intelligence but of extraordinary position.

This suggests that taste may not be an aesthetic faculty at all but rather a structural advantage. We often imagine taste as residing inside someone’s head — as a refined sensibility or intuitive aesthetic judgment. Burt suggests that what looks like taste may instead arise from where someone sits in a social network. The people who seem to have an uncanny feel for what’s interesting, original, or promising are often those who spend their lives moving between otherwise disconnected social worlds. Their “taste” emerges not from mysterious intuition but from repeated exposure to combinations that no one else is able to see.

An LLM trained on, well, everything has, in a sense, access to all the networks simultaneously — so if structural holes were purely about information access, AI might excel. But the broker’s advantage comes from living inside multiple social worlds, with the embodied, relational, contextual knowledge that comes from their network position. We might say the broker’s advantage is tacit. Brokers don’t merely possess information from multiple communities; they actively participate in multiple communities. They accumulate tacit knowledge, trust, and experience, and respond to novel incentives in ways that LLMs simply can’t. You can’t simulate being an outsider-insider from training data alone. That’s a defensible human edge — for now, at least.

Granted, Burt’s work is about organizational careers and competitive strategy, not cultural products. But perhaps what we call taste in other domains may be structurally equivalent to Burt’s brokerage — the advantage of the person who’s genuinely of multiple worlds.

Judgment is trained

So where can we find human judgment? We create it.

LLMs can perform the tasks often associated with entry-level roles: drawing up contracts like a junior law associate, performing basic financial analysis like a junior consultant, coding like an engineer, and so on. It’s been suggested that this may allow organizations to slash costs and increase output by replacing those jobs with an LLM. It also grants expertise to the common person: No longer do you have to rely on Big Law and its hourly rates for basic (or sometimes even complex) tasks.

The human role in all of this, argues economist Christian Catalini, is verification: making sure the LLM’s output didn’t drift too far. Humans check. Humans correct. Humans employ taste?

But this opens up the question of where humans learn taste. Good (human) evaluators develop their abilities with apprenticeship, feedback, and exposure.

Back in the day, when I was a junior editor at Princeton University Press, I used to have to photocopy the book manuscripts that my boss had edited with his green fountain pen before I mailed them to the author. We needed a copy for reference and also in case the original was lost. While the copy machine churned away, I took the time to review his edits, read his queries, understand both his green-inked penmanship (which was a nontransferable job-specific skill that hasn’t helped later in my career but was, at the time, crucial, especially because few others — including the authors — could decipher his writing) and, more importantly, his editorial style. This informed my own editing (which my mentor reviewed in turn) and created the foundation from which I still work. (This experience also points to another often underappreciated truism: Jobs can be messy.)

In other words: At least one element of what we call taste — judgment — is a kind of discrimination one accumulates and refines over time. It is not innate. If we are to continue to create humans with taste, we have to create spaces where they can develop it.

The obvious objection is that firms investing in apprenticeships may lose in the short term to competitors who aggressively automate. Maybe! But the tradeoff resembles overcutting a forest: Chopping down every sapling may improve one quarter’s balance sheet, but it also ensures that you’ll have no mature trees later on. And of course it’s incumbent on anyone looking for a job to make sure they train themselves in the absence of those roles.

Practice complicates things

Pushing each of these three findings toward some kind of practice, none of them break — but I don’t think any of them means quite what the obvious reading suggests, either.

Take the socially contingent hit first. Salganik’s conclusion doesn’t change: Success still isn’t dictated by quality alone, so taste still can’t simply mean predicting success. What AI adds is volume — more candidates clearing the “good enough” bar than any one curator can sort through. In other words, TV execs still filter the same way. It just means there are a lot more shots on goal, and quality alone determines even less of which shots land. Success depends on luck — or, more precisely, social or network position — in this environment more than ever. Whichever candidates already have an audience are the ones most likely to catch a wave before the field crowds them out.

Brokerage holds up better. An LLM can already do the cheap half of brokerage — surfacing what’s happening in an adjacent field to anyone who asks, which used to require actually knowing someone in that field. (Someone does still need to know to ask the right questions.) The structural hole that really matters isn’t “knows two worlds” in the informational sense. It’s inhabiting worlds that haven’t been written down somewhere an LLM could ingest them: subcultures, informal practice, arguments that happen in a room and never make it onto the record, and so on. If brokerage keeps paying off, it will be because someone sought out exactly that kind of undocumented and undocumentable territory.

Apprenticeship is the one where the standard advice — protect it — holds up. The apprenticeship model assumed junior people learned by doing grunt work under a senior person’s mentorship, the way I learned editing by photocopying my mentor’s marked-up pages and then having him review my own work, once I was allowed to start editing on my own. If AI now does that grunt work, the easy conclusion is that the apprenticeship disappears along with the job that trained it.

This may be backward: The process that mattered was “someone catches what you got wrong, repeatedly, until you stop getting it wrong.” That can survive if the junior person’s job becomes checking and correcting the AI’s draft under supervision, rather than drafting something from scratch themselves (although I’d argue that can still be invaluable). What changes isn’t whether firms keep junior humans around but rather what those humans spend their day doing. Firms that simply dispense with entry-level roles will tear out a lot of scaffolding with it. Firms that redirect entry-level roles toward supervised verification can maintain their scaffolding while still capturing some savings. (There is a Tragedy of the Commons possibility here, though, where all the firms dispense with junior jobs and simply lack senior judgment across the board later on, like the timber company that harvests all the young trees.)

***

Boyse pawned his shirt for paper and died on Grub Street. Samuel Johnson, who allegedly walked the same streets, didn’t stay. Not because he had better taste but perhaps because he had better scaffolding.

In the debate over AI, “taste” has become a kind of totem, invoked when someone wants to identify something humans can do but machines can’t. But taste doesn’t have to be mysterious. It is what good judgment looks like after it has been shaped by environments that reward experimentation, networks that expose us to novel combinations, and institutions that cultivate discrimination.

In the age of AI, the scarce resource isn’t taste. What threatens to become scarce is the social infrastructure that produces the people we describe as having it.

***

Acknowledgments: Many thanks to Drew Coffman, Sonal Chokshi, Steph Zinn, and especially Robert Hackett for helpful comments and edits.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Arguments about AI writing tend to rely on two assumptions: (1) that machine-generated prose can be identified by a conspicuous set of tells, and (2) that anything containing those tells is necessarily bad and embarrassing. Respectfully, these are terrible ways to think about writing.

We writers can pontificate about what AI will do to the craft all we want. The reality is that most writing isn’t some artful monomyth. It just needs to work. Does it explain things clearly? Can readers trust it? Is someone — anyone — behind the wheel, making decisions here? In short, is the writing doing its job? The extent to which a human was involved will matter in some cases and not others. The nice thing about these questions is that they apply equally well to writing produced by a person, a model, or some increasingly entangled combination of the two. They’re also useful in industries like ours, where scams (often written by AI) and jargon (too often written by humans) can easily distract from real goals and values.

What is actually useful about the AI detection discourse is that it’s made some longstanding writing problems much easier to see. So instead of using some percentage of “Human Written” as the relevant heuristic, we take a look at the hallmarks of machine-generated prose — its rhetoric, voice, structure, punctuation — as a taxonomy of problems that have always existed. Our goal as editors is to help founders, and anyone else in the business of communicating ideas, figure out when these habits get in the way of effective writing, regardless of provenance.

Rhetorical hallmarks: Insight-shaped writing

One of the most infuriating (then rewarding) experiences in editing is finding a sentence that sounds good but just kind of feels off. You rearrange it, move a phrase here and there, and slowly realize it never meant anything to begin with. Then you delete that slippery little pretender and move on with your life.

A quick example straight out of asking Fable for its most recent punctuation meta:

“Human punctuation has a body; it fidgets. My punctuation is uniformly deliberate, and uniform deliberateness is itself the rhythm.”

I would describe this tone of voice as “person at an otherwise nice party talking about jazz improvisation,” but I’m not going to pick these sentences apart here — you get the point. People say things about this sort of writing, like “it could just never have been written by a human.” And I take their point, but as an editor, ghostwriter, and lifelong reader of fantasy novels, I can assure you that people do indeed write these sorts of sentences without AI all the time.

Spotting this kind of language, however, isn’t always as straightforward as in the fidgeting-punctuation example. Unlike other AI hallmarks (like three consecutive sentences full of lists or the familiar Morse code of too many em dashes), these sentences don’t actually look like a problem. They just kind of look like sentences. Some of them even sound better than the other sentences around them: “Something real is happening.” “This matters for crucial reasons.” “The implications are significant.”

AI writing produces these sorts of phrases in bulk, because they’re plausible (if semantically vacant) connective tissue. Before it became associated with AI writing, we used to call this category of bland, meaningless language corporate, because it’s your typical, dispassionate business-blog speak. Can you think of a more spiritless way to say you’re excited about something than to start with, “We’re excited to announce…”? “We’re at an inflection point” or “this is the next chapter of our journey” are two more examples. They’re phrases that sound weighty but generally lack the specificity to survive paraphrase.

I’ve heard other editors miscategorize this as lazy or thoughtless writing. That’s uncharitable. It can be a matter of taste, but mostly, it’s not easy to swat away the generic, instantaneously retrievable phrases and idioms buzzing at the front of our minds and find something better. It is first-draft language. And honestly, it’s a perfectly fine place to start. We just don’t want to end there.

The approach I recommend for avoiding meaningless language is the same, whether you’ve used AI or not. If you spot a sentence that feels off — if it uses any of the phrases or phrasings below, if you don’t immediately understand what it means, or if it’s just got a bad vibe to it — start rearranging and try to restate it.

If there’s a paraphrase, keep the better version. If it boils down to “stuff exists” or “things are changing,” skip it, and see if your piece can live without it. If it can, it’s time to embrace the illicit thrill of deleting words from your page.

Some tells and when to edit them

• Naturally essenced profundity (the La Croix of insight). “Something real is happening.” “The stakes couldn’t be higher.” Sounds important; is actually filler.
• Empty contrasts. “It’s not just about X — it’s about Y,” where Y is fuzzier than X. This format works sometimes when the contrast is very clear. But before leaving it as is, check whether you can delete that first “it’s not about” phrase. You may get to your point faster.
• Hedges. “In many ways.” “At some level.” “Arguably.” Each of these exists to make the sentence impossible to be wrong about. Note that in some industries, like finance or healthcare, some hedges are necessary to soften claims and comply with the law. I think AI would call these “load-bearing” hedges.
• Too much parallelism. Bullets and sentence structures that mirror each other too tidily. Every list item is the same grammatical shape and the same length. These are AI tells, but they also make for terrifically boring writing.
• In-summation phrases. “At the end of the day.” “When the dust settles.” It’s often possible to delete these phrases with no impact on your paragraph.

Voice hallmarks: Alexa voice

There are roughly one million words in the English language. Default AI writing sounds like it’s working with about four hundred of them. But rather than trying to pinpoint which word of the day is an AI hallmark, the test here should be fungibility: could this sentence be taken word-for-word out of your writing and dropped into someone else’s essay on a different topic without anyone noticing?

For a while you could track this category of low-friction vocabulary empirically. Researchers famously noticed the word “delve” spiking in academic abstracts after ChatGPT launched in 2022, and an onslaught of banned-word lists followed: tapestry, testament, underscore; more recently, load-bearing, scaffolding, broader, etc.

In the meantime, thoughtful diction isn’t just a preoccupation of “literary” writers. It matters for anyone who starts or runs a company because people increasingly market on the strength of their personalities.

The best founder writing is individually expressive. Whether or not others would deem it “good,” it gives readers a sense of the person who believes it. It’s why we can so easily conjure what it means to write like Brian Armstrong (pithy, earnest, unironic); like Vitalik Buterin (technical, digressive, dense); or like Chris Dixon (austere, philosophical).

So, when using an LLM, please make sure you don’t cede your personality to it. Your own word choices and their imperfections add patina a machine could never.

For all of the writing available on finding the right words (two of my favorites are Theodore Bernstein’s The Careful Writer and George Saunders’ A Swim in a Pond in the Rain), there’s very little actionable advice on how to actually do it. Magic, alchemy, etc. It’s often an exercise in vibe curation, or emotional precision, rather than technical precision. Even quintessential writer of rulebooks E.B. White caveats that nobody can say for certain why some words “ignite” and others don’t.

Start by listing what’s good about the writing you like. Learn a new word, or use an old word in a new context. Match your word choices to the mood you’re trying to evoke — for example, use short, simple words to explain something opaque; or choose a word like “scheme” over “plan” when you want the reader to smell trouble. There’s a know-it-when-you-hear-it quality to diction that you can’t compensate for with rules.

Some tells and when to edit them

• Generic warmth. Friendly in a hold-message sort of way. “Great question!”
• Recyclable phrasing. Low-stakes phrases and sentences that could be transplanted into any other piece of writing with few consequences. “A useful way to think about it is,” “The key idea is,” “This can be understood as.”
• Low-friction vocabulary. Every word is precisely acceptable in an uncanny sort of way.
• Abstract nouns. The writing leans on words like “efficiency,” “complexity,” “society,” “communication,” and “innovation.” A timeless tactic for saying nothing in 3+ syllables that long predates ChatGPT.
• Insipid dynamism. “Navigate,” “leverage,” “unlock,” “foster,” both “power” and “empower,” “shape,” “elevate,” “streamline” — words that signal motion but fall flat.
• A beacon of something. “A testament to…” “Stands as a beacon of…” “Serves as a reminder that…”
• Gestures vaguely. “Landscape,” “space,” “journey,” “ecosystem,” “tapestry” — ambiguous, noncommittal words that point to the vicinity of things instead of identifying them directly.
• Vague intensifiers. Phrases like “very important,” “significant impact,” or “major role,” especially when there’s nothing concrete backing them up.

Structural hallmarks: All form, no function

AI writing tends to feel a bit overstructured when left to its defaults — too many H2s and lists, enough paragraph breaks to resemble a William Carlos Williams poem. But it’s not all bad. People have always drawn on a bank of familiar structures that work well and feel good on the brain.

We group ideas into three because three feels complete. We add signposts (e.g., “First,…” or “In other words, …”) to instantly situate readers within our writing. We create neat little taxonomies because they’re easier to skim.

As students, most of us learn some version of “hamburger” logic, where we tell our audience what we’re going to say, say it in discrete and supported chunks, and then restate it to conclude. It’s useful while forming an argument, because it forces us to articulate a thesis, collect evidence, and arrange our thoughts into something legible — a good thing!

A good, clear structure helps readers immediately understand their surroundings, whether that’s an op-ed, an explainer, or even a piece of fiction. The trouble is when pre-set structures pressure writers to draw and quarter their ideas into formats that don’t necessarily make sense.

Structure is a set of decisions — which container to use, which information matters most, which ideas go together, and what to call them. The right decisions depend on the job to be done. A narrative essay needs discovery and tension to keep readers moving. A product announcement needs to be brutally efficient to catch readers mid-scroll. An explainer needs an order that builds on concepts sequentially learned.

Ask yourself — anywhere in the writing process, but hopefully at the beginning — “what is the best format for my idea?” Then, borrow from something that already works. If you’re working on an argument piece, take a moment to understand how other writers structure op-eds and similar articles. If you’re writing a technical explainer, pick out the best explainer you’ve ever read, and look at how its author organizes information. Once you have a template, you can apply it to your own work.

Some tells and when to edit them

• Over-organization. Lots of subheads, bullets, numbered sections, or mini-frameworks.
• Ideas and lists that come in threes. Though this remains a best practice; more below.
• Familiar essay shapes. Broad introduction, explanation, examples, caveat, conclusion.
• Formulaic openings. “In today’s rapidly changing world…”
• A little too much signposting. “First,” “Next,” “Finally,” “In conclusion,” “Here’s a breakdown,” “Let’s unpack this.”
• Hard-pivot transitions. “To understand why this matters, we first need to look at…”
• Section previews. “There are three key reasons…”
• Bulleted lists. AI detectors often flag these unfairly; bullets are useful! A signature move, however, is the bold lead-in to a bullet point. See, for example, these exact bullets (but they’re so much easier to read this way).
• Punchy fragments. Seen in many a dramatic LinkedIn post. Short. Punchy. Often in threes.
• Ending on a restatement. In other words, the conclusion paraphrases the piece instead of expanding and hinting at further directions.
• Moral-of-the-story endings. A vague final sentence about progress, the future, or “what we can learn.”

Punctuation hallmarks: The em dash panic

The em dash — now a canonical AI tell — has always been divisive (😏). Strunk and White advise restraint more generally: “Use a dash only when a more common mark of punctuation seems inadequate.” But there’s a very easy, conversational feel to em dashes, in particular, that no other punctuation mark really has.

Em dashes are often seen as the enemy of efficient writing — it’s admittedly kind of distracting to jam this whole other phrase into the middle of a sentence — but as soon as they started signaling slop writing, they were an easy target for a ban.

The thing about punctuation, though, is that it has extremely specific rules, based on whatever 1,000+ page style guide you’re following. Whether AI favors em dashes is beside the point, because there are certain situations where you should probably use them. At the very least, they’re the most appropriate choice for setting apart longer parenthetical information, or for illustrating some abrupt changes in thought and emotion (as rhapsodized by horror writer R.L. Stine).

My take: shift + option + dash (on macOS) is humanity’s own troubled, distracted, often off-topic child, and I, for one, will never abandon it. Sometimes nothing hits quite like an em dash, so I urge you to avoid letting the current AI writing meta influence your choice of punctuation marks, not least because all metas must, by definition, change.

Now that people are bending over backward to avoid em dashes, AI itself is routing around them with colons. Shall we start using interrobangs to look appropriately human?! Since no punctuation mark is safe, choose the one that’s best for your writing.

So how do you tell what’s best for your writing? The short answer is: whatever is most correct and least distracting. The particulars of the rules can be nuanced and vary across taste and styles (Oxford vs. no Oxford comma is a popular religious debate among people who’ve made grammar and usage part of their personalities), so getting them perfect isn’t as important as ensuring they support your meaning and don’t infuriate your readers.

Another punctuation-related hallmark to look out for is sameness; that is, are all of your sentences looking and sounding the same? LLMs often use consecutive lists starting with a colon, which makes for an exhausting read. Likewise, several sentences interrupted by em dash asides will distract from your point, no matter who or what wrote it.

A good test is reading your writing aloud, using punctuation like stage directions. Anything that sounds unnatural will probably stand out to readers too.

Some tells and when to edit them

• Colon-heavy phrasing. “The issue is:”, “The result is:”, “The key point is:”… but especially colons followed by grocery lists of things.
• Em dash clustering. Not the dash itself, but the density, especially paired with lists and asides. There’s not really a numerical limit on the number of em dashes you can use in a sentence, but don’t use more than two.
• Unserious parentheticals (like this one). These usually carry the self-aware or joking register, while dashes carry the qualifying one.
• Performative semicolons. Writer Kurt Vonnegut once said that the only reason to use a semicolon is to show that you’ve been to college. It’s actually the least mean thing he said about semicolons, but it’s still unnecessarily crotchety. They have many legitimate use cases; however, these don’t come up all that often, honestly.

As more people use these tools, interrogating whether something is machine-generated is becoming kind of pointless. The answer is almost always going to be “to some degree.” Of course, the egregious cases of AI writing still need outing, whether by our own standards or Proof of Person technologies: for example, when disclosure is required, when the human on the byline is the whole point, or when one person is pretending to be a thousand.

For almost everything else, we can ask the same question we’ve always asked: is the writing doing its job?

AI helps people articulate and publish ideas they’d otherwise never write down. It can save time on organization and research. It may even help us become better writers. It’s mean-spirited to dismiss the ability to express oneself as “cringe” or “slop.” And, if a piece works as intended, does it really matter which parts of the process show through?

Finally, with all of the hand-wringing over whether our writing will be exposed as AI-assisted, one question worth asking is: why are we showing LLMs this kind of deference at all? We would relinquish an entire punctuation mark — an innovation conceived at the dawn of printing! — all to avoid looking like a machine helped us. It’s a ridiculous concession, especially given how often people already use, and will continue to use, the machine.

Acknowledgments: Thanks to the a16z crypto editorial team — Tim Sullivan, Robert Hackett, and Sonal Chokshi — for feedback on this post, and for the many years of edits and debates that informed it.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

More than 30 years ago, Gödel Prize winner Noam Nisan helped develop a proof technique that he never expected anyone to use in the real world. At the center of that breakthrough was the sum-check protocol — an elegant technique that emerged from Noam Nisan’s early work on interactive proofs. At the time, even Nisan assumed ideas like these were far too theoretical to become practical technology. Today, sum-check sits at the heart of some of the fastest modern SNARKs, proof systems that are increasingly important to blockchains and verifiable computation.

In this episode of First Principles, a16z crypto Head of Research Tim Roughgarden and research partner Justin Thaler sit down with Noam Nisan to trace that unlikely journey from theory to infrastructure.

They explore how interaction and randomness changed the very definition of a mathematical proof; the deceptively simple “two into one” insight behind sum-check; the rapid sequence of discoveries that culminated in IP = PSPACE; and why Justin believes sum-check may be close to the optimal tool for building practical SNARKs.

But Nisan’s career also tells a broader story about how research responds to technological change. When the web arrived in the 1990s, he deliberately left a field in which he was already a leading researcher to understand a new problem: How do you get independent actors on the internet to cooperate when they have different incentives? That question helped give rise to algorithmic game theory — and, decades later, brought Nisan back to questions around blockchain fees, token economics, and protocol design.

Highlights

0:00 — Intro
1:28 — Noam Nisan and the origins of verifiable computation
2:58 — Why Noam Nisan left complexity theory
7:52 — POPcorn, distributed computing, and early blockchain-like ideas
11:10 — The birth of algorithmic game theory
16:14 — Justin Thaler discovers the sum-check protocol
25:05 — From arithmetization to LFKN
27:06 — The story behind IP = PSPACE
30:40 — Why sum-check matters for modern SNARKs
31:15 — What is a SNARK?
38:52 — The key idea behind sum-check: turning two into one
44:03 — When SNARKs went from theory to practice
46:03 — Why blockchains were the breakthrough use case
50:36 — Noam Nisan’s move into blockchain economics
56:15 — EIP-1559, transaction fees, and efficient blockspace
1:02:07 — From theoretical computer science to real-world systems
1:08:49 — Boiling down the sum-check protocol: Two become one
1:10:58 — Is the sum-check protocol optimal?

About First Principles

First Principles is a series from a16z crypto exploring the foundational ideas behind modern computing, cryptography, and blockchains through conversations with the researchers who developed them — and the people extending those ideas today.

Subscribe for more First Principles episodes: https://www.youtube.com/@a16zcrypto

Links

Noam Nisan
Hebrew University: https://cris.huji.ac.il/en/persons/noam-nisan/
StarkWare: https://starkware.co/about-us/

Justin Thaler
Georgetown University: https://people.cs.georgetown.edu/jthaler/

Tim Roughgarden
https://timroughgarden.org/

a16z crypto
Website: https://a16zcrypto.com/
X: https://twitter.com/a16zcrypto
Newsletter: https://a16zcrypto.substack.com/
YouTube: https://www.youtube.com/@a16zcrypto

Crypto payment cards have gone from a novelty to more than $750 million in monthly spend. These cards let people pay with crypto anywhere traditional card networks are accepted. Behind the scenes, the crypto — stablecoins, overwhelmingly — gets converted to local currency at the point of sale, so the payments look like any other card transactions to merchants.

Crypto cardholders don’t require a traditional bank account. Depending on the program, users either deposit stablecoins with a card issuer, or hold them directly onchain through self-custody. Crypto cards expand people’s access to U.S. dollar accounts globally, and they offer a convenient way for stablecoin holders to transact.

Monthly crypto payment card volume reached $759 million in July, up (roughly 2.5x) from $306 million a year earlier — and up from less than $1 million when tracking began in Oct. 2023. These figures reflect onchain activity attributable to card programs tracked by Paymentscan. (For RedotPay, the largest program by volume, spending is self-reported by the issuer rather than observed onchain.)

Growth in the number of purchases made with crypto payment cards is trending similarly to the growth in crypto card volume. Nearly 9 million purchases were made using crypto payment cards during July, up from about 5.2 million a year ago.

That puts the average amount spent per transaction at around $86.

In early 2024, crypto card spending was concentrated in a single chain: Gnosis, home of Gnosis Pay (the first Visa card connected directly to a self-custodial wallet). As new card programs have launched, the number of card settlement chains has expanded. 

As of July, Optimism carries about 29% of crypto card spend volume, Solana about 19%, and Base about 19%, according to Paymentscan. Gnosis has dropped to about 2%.

Euro-backed stablecoins once dominated spending: In early 2024, around 88% of crypto card volume settled in EURe, much of it on Gnosis. As of July, EURe’s share has fallen to about 2%.

Dollar-backed stablecoins have since taken the lead. USDC handles about 58% of card spending and USDT about 26%, up from roughly 48% and 7% a year ago, respectively. Crypto payment card spending now happens overwhelmingly in digital dollars.

Crypto payment cards remain a small market next to traditional card networks, which process trillions of dollars per month. 

But the trend is growing as stablecoins make greater inroads into the global financial system, including by piggybacking on existing major card network rails. For the tracked programs, this is happening almost entirely through Visa.

Crypto payment cards are one piece of a broader crypto acceleration post-GENIUS, which we’ve been charting here. That includes the rapid adoption of stablecoins and tokenized assets.

***

Robert Hackett is features editor and head of special projects at a16z crypto.

Ryan Holloway is a data consultant.

***

This article was translated into Korean by Token Post, available here.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Here are nine key moments from Marc Andreessen and Chris Dixon’s conversation on what’s at stake. Or watch the full conversation here.

1. After a five-year regulatory onslaught, crypto is still standing

2. How compliance became a competitive disadvantage

3. What builders actually want from DC

4. Why blockchain transparency is good for national security

5. A “kill shot” for open source

6. How crypto rules became reliant on an ethics debate

7. The policy regime that gave us FTX

8. Marc Andreessen on American technological leadership: “It’s such an obvious argument”

9. What’s next: finance is only the first major use case for crypto

How can you prove that something is true without revealing why it is true?

That question gave rise to zero-knowledge proofs, one of the most important breakthroughs in modern cryptography. What began as an attempt to solve a seemingly playful problem — playing poker securely over the telephone — ultimately changed how computer scientists think about proofs, privacy, and verification.

In this episode of First Principles, Turing Award–winning cryptographer Shafi Goldwasser joins a16z crypto Head of Research Tim Roughgarden and Research Partner Justin Thaler to tell the origin story of zero-knowledge proofs and interactive proof systems.

Goldwasser recounts how she, Silvio Micali, and Charles Rackoff developed a new kind of proof involving interaction, randomness, and a small probability of error. Their work introduced the idea that a prover could convince a verifier that a statement is true while revealing no additional information.

The conversation follows those ideas through some of theoretical computer science’s deepest results, including interactive proofs, IP = PSPACE, probabilistically checkable proofs, the sum-check protocol, and SNARKs. These concepts now power blockchain rollups, privacy-preserving applications, and systems for verifying computation without repeating all the work.

They also explore why breakthrough ideas are often initially rejected, how toy problems can lead to foundational theories, why abstraction and narrative matter in scientific research, and whether AI systems should be required to prove their answers.

Highlights

• 0:00 — Intro
• 3:36 — How mental poker inspired zero-knowledge proofs: Proving that something is true without revealing the underlying information
• 6:30 — The simulation paradigm and the meaning of “zero knowledge”
• 8:33 — Why the original paper was repeatedly rejected
• 10:33 — How interactive proofs became more powerful than conventional proofs
• 13:36 — The road to modern SNARKs
• 19:02 — Why the sum-check protocol is so useful for verifiable computation
• 25:31 — Why many so-called “zk proofs” are not actually zero knowledge
• 34:06 — Why toy examples, playfulness, and narratives can produce deep theory
• 37:23 — The role of rigor and computational assumptions in cryptography
• 42:44 — Applying zero-knowledge proofs to law, evidence, and secret software
• 45:23 — Training AI systems to provide proofs alongside their answers
• 54:27 — Why genuinely new ideas are often difficult for experts to recognize

About First Principles

First Principles is a special limited series from a16z crypto about the scientific roots of modern computing — especially blockchains — told through rare conversations with the pioneers who helped shape the foundational ideas behind distributed systems, consensus protocols, economics, mechanism design, cryptography, zero knowledge, and more.

People often tell the story of the Bitcoin whitepaper as if it appeared out of nowhere. But the ideas behind Bitcoin — and blockchains more broadly — come from decades of computer science, economics, mathematics, and cryptography. First Principles is a guide to that lineage, as told by the people who helped build it.

Links

Shafi Goldwasser:

• MIT CSAIL: http://people.csail.mit.edu/shafi/
• Simons Institute: https://simons.berkeley.edu/people/shafi-goldwasser

Tim Roughgarden:

• a16z crypto: https://a16zcrypto.com/team/tim-roughgarden/
• Website: https://timroughgarden.org/
• X: https://x.com/Tim_Roughgarden

Justin Thaler:

• a16z crypto: https://a16zcrypto.com/team/justin-thaler/
• Website: https://people.cs.georgetown.edu/jthaler/
• X: https://x.com/SuccinctJT

a16z crypto:

• Subscribe: https://www.youtube.com/@a16zcrypto
• Website: https://a16zcrypto.com/
• X: https://twitter.com/a16zcrypto
• Newsletter: https://a16zcrypto.substack.com/

Crypto is no longer a niche market. Stablecoins move trillions of dollars, and major banks and payment companies are building onchain, but the federal rules governing much of that activity remain incomplete.

That is the problem the “CLARITY Act” (HR 3633) is meant to solve. The bill would establish federal oversight for crypto markets, define the respective roles of the SEC and CFTC, impose disclosures and restrictions on insiders, and bring exchanges and other intermediaries under rules familiar from traditional finance. If passed, this bill will establish clear rules of the road for blockchain systems — ending the years of uncertainty that have stifled innovation and exposed consumers to harm.

This post — based on a recent conversation with a16z Cofounder and General Partner Marc Andreessen and a16z crypto Founder and Managing Partner Chris Dixon — covers why crypto needs clear, lasting rules now; how CLARITY would protect consumers; why regulatory ambiguity rewards bad actors; and how the bill addresses illicit finance, privacy, and government ethics. It also explains what happens if the bill fails; why American technological leadership is at stake; and why the biggest risk is the status quo.

For more, watch the full conversation:

Why does crypto need rules now?

Crypto has come a long way since the Bitcoin whitepaper. What started as a technology used largely by hobbyists and enthusiasts is now an industry with mature infrastructure and growing institutional adoption.

A technology has become an industry: stablecoins now rival the size of the Visa network, with trillions of dollars transacted. Major financial institutions, including banks, asset managers, card networks, and fintech companies, are building products with stablecoins, tokenized stocks, tokenized deposits, and other digital assets. The underlying networks have also become faster and cheaper. Transactions that once cost several dollars can now settle in under a second for less than a penny on widely used blockchains.

For a variety of reasons, the regulation around crypto was broken into two components — stablecoins and the rest of the market. The GENIUS Act, which became law in July 2025, created a federal framework for stablecoins, but stablecoins depend on blockchain networks and markets that still lack a comprehensive federal structure. It is the equivalent of regulating cell phones while leaving the cell towers in legal limbo.

Agency guidance can fill some gaps, but it is not a substitute for legislation. Guidance can change with agency leadership or a new administration. Companies deciding whether to make investments that may take five or ten years need to know what the rules are, which regulator has authority, and whether the product they build today will remain lawful tomorrow.

CLARITY would provide a permanent framework for doing business responsibly.

How would CLARITY protect consumers?

The most basic consumer-protection problem in crypto is that exchanges do not operate under the same comprehensive federal framework that governs major securities and commodities exchanges.

The New York Stock Exchange and Nasdaq have federal regulators. Crypto exchanges do not have an equivalent market-wide system of registration, supervision, audits, disclosures, surveillance, and customer-asset protections. The CLARITY Act would provide a clear pathway for digital assets to transition from the Securities and Exchange Commission (SEC) to oversight by the Commodity Futures Trading Commission (CFTC).

A federally registered crypto exchange would be subject to audits and financial controls. It would have to safeguard customer assets, comply with anti-fraud and insider-trading rules, and provide regulators with information about its operations. A company that refused to meet those standards could not legally operate in the United States.

Those requirements help prevent the conditions that allowed FTX to collapse. FTX allegedly moved money among related entities, lacked adequate controls, and did not hold the customer assets it claimed to hold. Federal oversight cannot guarantee that fraud will never happen. It can make it much harder to hide, and it can give regulators the authority to intervene before a failure becomes a catastrophe.

The same principle applies to products marketed as stablecoins. Terra-Luna was presented as stable even though it was not backed by dollars or another stable reserve asset. Under the stablecoin framework, a compliant dollar-denominated stablecoin must be backed by corresponding reserves and subject to audit. CLARITY would bring comparable discipline to the rest of the market.

How would CLARITY stop ambiguity from rewarding bad actors?

Regulatory ambiguity creates a race to the bottom.

A U.S.-based company that takes compliance seriously may spend heavily on lawyers, controls, audits, sanctions screening, and customer protections. Those obligations cost money and can slow product development. An offshore competitor can avoid those expenses, copy the product, charge lower fees, and move faster precisely because it is not doing the compliance work.

The result is that uncertainty can punish the responsible company and reward its offshore counterpart. A noncompliant offshore exchange shouldn’t be able to serve Americans while a compliant U.S. exchange bears the full cost of following the law.

CLARITY would define the regulatory perimeter: which businesses are intermediaries, which rules apply to them, which agency supervises them, and what happens if they refuse to comply. A company that holds customer funds or facilitates financial transactions would be subject to the same kinds of anti-money-laundering, sanctions, and Treasury requirements that apply to comparable financial businesses such as payment providers and fintech companies.

Clear rules favor companies willing to meet a standard. Gray areas favor companies willing to exploit them.

How would CLARITY strengthen sanctions enforcement?

Privacy isn’t the same as secrecy. Public blockchains are often described as anonymous. In practice, many are highly transparent.

Transactions are recorded permanently on a public ledger. A wallet address may not immediately display a legal name, but investigators can follow the movement of funds and connect that activity to exchanges, accounts, devices, or other identifying information. The records remain available even years later, which can give law enforcement evidence that did not exist at the time of the transaction.

Unlike payment methods that may leave no public trace, blockchains produce trails. That is why some national security officials have described crypto activity as creating “prosecution futures”: transactions recorded now can help investigators identify and prosecute criminals later.

But traceability and privacy are separate issues. A person should not have to publish every medical payment or transfer to the entire world in order to use a blockchain. The financial system already recognizes that ordinary people need privacy even though regulated institutions remain subject to sanctions and anti-money-laundering obligations.

The early debate over internet encryption offers a useful comparison. Strong encryption was once treated as a threat because criminals could use it. It was even classified under export rules alongside military technology. But encryption also made secure banking, ecommerce, and confidential communication possible.

Blockchain privacy raises the same distinction. Privacy protects legitimate activity. Secrecy designed to evade the law is still subject to enforcement.

How does CLARITY resolve the stablecoin rewards debate while allowing banks to keep building onchain?

Banks have argued that stablecoin issuers and wallet providers should not be allowed to pay interest on balances in a way that recreates a bank account outside the banking system. Their concern is that consumers could move deposits out of banks and into stablecoin products, reducing the funding banks use to make loans.

CLARITY addresses that concern by barring interest on stablecoin balances and products that are functionally or economically equivalent to an interest-bearing account.

The bill still leaves room for transaction-based rewards. A wallet provider or retailer could reward a customer for using a stablecoin to make purchases, much as a credit card offers points or a retailer operates a loyalty program. The distinction is between earning a benefit for activity and earning interest simply for holding a balance.

That compromise gives banks much of what they asked for without going so far that it would prohibit ordinary rewards programs, including programs resembling those already offered by card networks, payment apps, and retailers.

The debate is notable because banks are already adopting blockchain technology. Other banks, asset managers, and payment companies have their own initiatives. Goldman Sachs, Fidelity, BlackRock, Stripe, Wells Fargo, JPMorgan, and other large financial institutions have built or backed blockchain products.

Banks see the same opportunity the crypto industry sees: financial infrastructure is fragmented and difficult to modernize. Blockchains give those institutions a shared framework, allowing them to reduce layers of mediation, settle assets on common infrastructure, and modernize together rather than requiring every bank to rebuild an interconnected system on its own.

When are software developers liable under CLARITY?

CLARITY distinguishes between knowingly helping someone commit a crime and publishing general-purpose software. Developers who build tools for criminal use, market them to criminals, or directly assist illegal activity can still be held accountable.

What the bill rejects is unlimited liability for every downstream use a developer cannot foresee or control. Open source code can be copied, modified, and deployed by people the original developer has never met and for use cases they never anticipated. Making developers responsible for all of those uses would make open source nearly impossible to build or fund.

The consequences would reach far beyond crypto. Academic research, startups, venture investment, and open AI models all depend on open source software. The workable line is intent and participation: hold people liable when they knowingly facilitate crimes, not when someone later misuses a neutral tool.

What does CLARITY actually do to securities law?

CLARITY does not turn securities into non-securities simply because they are on a blockchain. A tokenized stock is still a stock. It is still a security and remains under SEC oversight. A company cannot avoid disclosure, registration, or investor-protection requirements by putting an asset onchain or calling it a token.

The bill addresses a different problem: how to regulate digital assets associated with blockchain networks that change over time.

A quick summary of CLARITY’s risk-based framework: A new blockchain generally begins with a central actor. A founder, company, or small group may control the network, possess information unavailable to the public, and make decisions that affect the token’s value. During that stage, the asset would be subject to SEC oversight and securities-style requirements. Those requirements include disclosures, restrictions on insiders, and lockup periods for founders and early investors.

As the network develops, control can become distributed. If it meets defined thresholds of decentralization, the asset may begin to resemble a commodity more than a corporate security. At that point, the CFTC would oversee it.

That does not mean the asset becomes unregulated. Commodity regulation addresses fraud, manipulation, market cornering, and other abusive conduct. The regulator changes because the nature of the asset has changed.

The bill would also impose restrictions that do not exist clearly today. Founders, venture investors, and other insiders more broadly could face longer lockups and stronger disclosure obligations while a network remains centrally controlled. Those restrictions are meant to prevent insiders from selling into the market before ordinary participants have comparable information or before the product has developed into a sufficiently decentralized network.

What happens if CLARITY doesn’t pass?

Crypto regulation would not disappear. The SEC, CFTC, Treasury, and other agencies have been and would likely continue issuing guidance and using the authority they already have to provide rules within their jurisdiction.

The problem is that these agency interpretations can change between administrations. A company can spend years building under one set of assumptions only to face a different interpretation after an election or a change in agency leadership.

The uncertainty affects consumer protection as much as investment. A lasting framework gives regulators clear authority and gives companies an obligation to register, disclose information, protect customer assets, and follow market rules. Without legislation, those responsibilities remain fragmented and contestable.

The industry has survived years of aggressive enforcement and political hostility. The more likely result is that companies continue building elsewhere. That would leave the United States with less oversight. Offshore companies are harder for American regulators to supervise, harder for law enforcement to reach, and less likely to build around U.S. standards.

Why is CLARITY part of the long tradition of American technological leadership?

Once a technology has been invented, it’s unlikely to be uninvented. The question is where it will be developed, which companies will lead it, and whose rules will shape it.

The United States has benefited for more than a century from being the country where major technologies are built. That leadership produces companies, jobs, tax revenue, technical expertise, and the economic capacity to fund national priorities. It also creates security advantages.

The history of encryption shows what is at stake. When the United States restricted strong encryption exports, foreign competitors did not stop building encryption. They built products outside the United States, and customers used those products instead. Once the restrictions changed, American companies helped build a secure internet economy.

Blockchain technologies present the same question. The financial systems, technical standards, and companies will develop somewhere. If they develop primarily offshore, the United States loses economic opportunity and regulatory power at the same time.

CLARITY would give responsible companies a reason to build under American law. That would support consumers, law enforcement, national security, and the country’s ability to set standards for the next generation of financial infrastructure.

Who else supports CLARITY?

CLARITY has drawn support from a mix of lawmakers, law enforcement organizations, financial institutions, and technology companies.

The legislation has been a bipartisan effort in Congress, with lawmakers from both parties working for years to establish a federal framework for digital asset markets. The Fraternal Order of Police, the country’s largest law enforcement organization, has also endorsed the bill, pushing back on claims that it would weaken sanctions or anti-money-laundering enforcement.

Support extends across the financial sector. Goldman Sachs CEO David Solomon has endorsed CLARITY, while firms and fintechs are already building blockchain products. Far-reaching support reflects growing agreement on the underlying need: the United States should have clear, enforceable rules for digital asset markets.

* * *

When the rules for markets are uncertain, consumers can’t know which protections apply, while responsible companies spend heavily on compliance and offshore competitors avoid it. CLARITY would replace that uncertainty with a defined system.

The relevant comparison is not between CLARITY and another law, but between CLARITY and the status quo. By giving responsible companies a clear path to build, CLARITY would strengthen consumer protections, support law enforcement, and help ensure the next generation of financial technology is developed in the U.S.

Highlights

00:00 Intro
05:31 From crypto subculture to financial infrastructure
08:37 Why crypto needs rules now
12:20 The regulatory war on crypto
15:41 How CLARITY could prevent another FTX
22:42 Why criminals using crypto may be easier to catch
26:20 Privacy, blockchains, and the invention of HTTPS
30:41 Government ethics and crypto
34:52 The banking lobby’s stablecoin fight
37:05 Why every major bank is building on blockchains
41:04 Developer liability as a kill shot
45:23 How CLARITY provides oversight
49:30 What happens if CLARITY fails?
50:50 Regulation vs. innovation
54:12 Why America should lead
55:36 What CLARITY could unlock

Trancript

Robert: Welcome to the a16z crypto show. I’m Robert Hackett, and I’m here with Marc Andreessen and Chris Dixon, who, if you watch the show, need no introduction.

Marc: Great to be here.

Robert: So today we’re talking about regulation. Congress is currently hashing through once-in-a-generation market structure legislation that could determine where the future of the financial system and the internet get built.

There have been lots of fights and obstacles, which we’ll get into, but before we do, I want to zoom out and go big picture and talk about why regulatory clarity matters, what the status quo costs us, and what’s at stake for the country and for anybody who may end up using this technology.

So, Marc, I’d like to start with you. Back in January 2014, you wrote an op-ed for The New York Times called “Why Bitcoin Matters.” Things were a little bit different then. This was a pretty contrarian position to stake out. The crypto industry, which, if you could even call it that back then, looked very different.

How have things changed from then to now?

Marc: Sure. 2014 was so long ago that The New York Times actually ran a positive piece on crypto.

Robert: From you.

Marc: From me, with my name on it. And so, yes, this was in the distant past. For the kids watching this, this is when dinosaurs roamed the Earth, thirteen years ago.

Might as well have been in the 1500s. But I’m actually really proud of that piece. And so if you read that piece today, I think it holds up really well. I think it holds up in a few respects. One is — and I should say this — it’s actually fairly amazing: this piece came five years after the Bitcoin white paper, already five years into the broader trend.

And it was still big news to believe that this technology mattered at all. So I think I was prescient, not that Bitcoin was going to matter, but in forecasting that we were still at the beginning of the adoption curve.

There were still a lot of people who were going to have to learn about this, and who would want to learn about it and be part of it. And so I think that holds up. I spent a lot of time in the piece and around that time trying to educate people on the nature of this technology, with Bitcoin and the blockchain, because that was new.

For people who do have computer science degrees, it’s a weird idea. So for people who don’t have computer science degrees, it’s a genuinely new idea. And I think the importance of that has held up really well. I think the thing that, in retrospect, needs to be changed about the piece is: every time I said Bitcoin, just swap in crypto.

At that time, Bitcoin was crypto. It was essentially the only game in town. Chris, you can correct me on the timing, but I believe at that point there was the proposal for what were called colored coins.

Chris: Around then.

Marc: Around then. For people who don’t remember, there was a proposal for this thing called colored coins, which is the single worst-branded proposal of all time.

Much worse than New Coke. But the concept was that bitcoins would be tagged with properties that would let them represent other things. The idea was that Bitcoin would be extended to enable both tokenization of real-world assets and virtual assets like NFTs.

Of course, that didn’t happen. What happened instead was the creation of new blockchains and new crypto platforms, and then ultimately Ethereum and everything else that followed. What started as a technology became an industry.

And so I think if you read that piece, you’ll see that I projected many of the use cases and the future development of the technology, but it took a different path. Bitcoin obviously has been very successful since then, but what we’ve really seen is this explosion of innovation.

This explosion of innovation has both happened and still wants to happen around all the other use cases we talked about in the piece.

Robert: It’s actually remarkable how well that piece holds up. But there is one forecast in there that I want to cite in particular. You wrote, and I quote, “The coming years will be a period of great drama and excitement revolving around this new technology.”

I think you were spot on there. Chris, you’ve been there since the start as well. You were an early advocate and acolyte of crypto.

Robert: How have things changed from the early days to now?

Chris: As Marc was saying, way back in 2014, in the early days, it was mostly a cultish subculture of hobbyists and super enthusiasts.

The technology — Bitcoin — was sophisticated, but a lot of the newer blockchains Marc alluded to were early and had performance issues, scaling issues, and other kinds of problems. Fast-forward to today: every day, you’ll see a new announcement from a major bank or fintech company about a platform involving stablecoins, tokenized stocks, or other kinds of digital assets.

Stablecoins, for those who don’t know, are basically what Marc described with Bitcoin, except they’re dollars on blockchains. Stablecoin volume now rivals the size of the Visa network. You can go into WhatsApp and send money anywhere in the world for almost free, in the same way you’d send a text message, using stablecoins.

It’s really the way money should work. I think Marc could speak to this as one of the pioneers of the internet, but it was one of the ideas early internet pioneers expected to happen much sooner. For a variety of reasons, it took a lot longer, but we’re now finally seeing that dream come alive, with money moving as easily as bits.

So now it’s a very mature industry. The technology, the underlying infrastructure, has gotten much more sophisticated. Even three years ago, what I just described might cost a couple of dollars, if not tens of dollars, per transaction. Now, on most of the popular blockchains — Solana, Ethereum, and so forth — transactions settle in under one second and cost less than a penny.

And that’s due to the equivalent of Moore’s law on blockchains: you’ve gotten more and more performant chains. You read about hacks and things in the newspaper about crypto. Bitcoin has never been hacked. Ethereum has never been hacked.

Organizations that use these things have had lax security and have been hacked, but the chains themselves are highly secure and highly performant now. You have big institutions and trillions of dollars moving around in these things, which brings us to our topic: when you have that going on, you need regulatory clarity.

You need a framework around that.

Robert: Let’s talk about that. You cited that stablecoin volumes are now rivaling the Visa network, which is insane. This is trillions of dollars per quarter moving through this system. Lots of gigantic financial institutions have rushed in: BlackRock, JPMorgan, Visa itself, Fidelity, Mastercard.

You could rattle off just any number of names, and they’re all building on this technology right now. But we’ve gotten this far without rules. So there’s been very little in the way of regulatory clarity or guidance.

Robert: Why do we need rules now?

Chris: Good question.

For political and policy reasons, regulation around crypto was broken into two components: stablecoins and the rest of the market. Stablecoins are roughly 15 percent of the market, let’s call it, and the rest is 85 percent.

Robert: And that’s by market cap.

Chris: That’s by market cap. Last year, a bill passed Congress and was signed by the president called the GENIUS Act, which provides a comprehensive regulatory framework for stablecoins. It’s not a coincidence that stablecoins are also the area that has taken off the most in the last year, because once you have regulatory clarity, builders have certainty and consumers know what to expect.

If you’re an American consumer and you use USDC or another GENIUS-compliant stablecoin, and you have a dollar of that stablecoin, there is a dollar sitting in the bank, because that’s the GENIUS regulation. That bank is audited, and the money is sitting in short-term Treasuries, which is basically the safest thing the financial system can offer. And so you can be assured that the dollar is there.

It’s not an FTX or Terra-Luna situation where it will disappear tomorrow. That gives you, as a consumer, confidence and protection. If you’re an institution, a bank, Stripe, or PayPal, and you want to enter this market, you want to know there are predictable rules and a framework around it. You want to know there’s not going to be an FTX and that what you build will be compliant not just next year, but for the next 10 years.

And then if you’re an entrepreneur, it’s similar logic. You want to know there’s certainty. You want to know there are rules. All our entrepreneurs are happy to hire lawyers and pay a lot of money to be compliant. They just need to know what they’re aiming for.

And so, as a result of that bill, we’ve seen a takeoff, particularly around stablecoins. A close cousin of stablecoins is the tokenization of other assets: once you put a dollar on a blockchain, you can put a stock or a Treasury bill on one. There’s also been a lot of adoption there, with people relying on agency guidance and other nonlegislative regulatory levers to guide them on how to do that.

But the big hanging problem right now is this other section of the market, including the blockchains themselves, which stablecoins are built on, that doesn’t have a comprehensive federal regulatory framework. That’s why the CLARITY Act, which is working its way through the Senate, is so important, because that’s the other 85 percent of the market.

You can imagine it as if we had a regulatory framework for cellphones, but not for cell towers. Half the technology is regulated; the other half has uncertainty around it. People are still building there and trying to do their best to understand what the rules are.

There’s another lever of policymaking: agencies such as the SEC and CFTC. They’ve provided some guidance. But ultimately, as you alluded to in the beginning with the internet, industries are built on legislation.

Legislation is solid ground to build on and has the stamp of approval of the U.S. Congress. Importantly, legislation is the result of a lot of smart people getting together, people from different political parties compromising and finding the right balance between industry and consumer protection.

And so we think ultimately that’s what is required and why the CLARITY Act is so important.

Robert: So as you mentioned, we do have some rules for stablecoins. That covers a sliver of the market, but there’s this gigantic swath of the market that remains largely unregulated. Marc, why do we need rules for all of this, and why do we need them now?

Marc: We’ve been in this bizarre twilight state for the last several years, where crypto was developing largely without rules, for better or worse, in what I would describe as experimental ways, between inception and 2020.

And then, between 2020 and 2024, we had an administration in the White House that absolutely decided to kill the industry and flat-out declared war. Specifically, it declined to regulate, refused to regulate, and instead decided to prosecute.

And this remarkable thing happened: the industry didn’t die through that process. Crypto is still standing. The most bizarre aspect of that for me is that I’ve had conversations with some lawmakers in D.C. where they say, ‘I’m going to support crypto now because if it didn’t die during this whole period when we tried to kill it, then it must be good.

We should probably support it.’ That is the most reverse-logic way I can think of to end up positive on something. But I guess I’ll take it. And then that also led to this bizarre situation where people said, ‘Crypto hasn’t materialized in the way that you guys all said.

You don’t have all these use cases, so where are all the use cases? Why isn’t everybody using crypto?’ And it’s like, well, because there was a five-year regulatory prosecutorial assault that tried to kill it. Most areas of the tech industry never go through that.

Occasionally, one does. When that happens, it operates under very different kinds of pressure. Along the way, when that assault was happening in America against American companies primarily, overseas companies by and large ran in whatever way their local jurisdiction allowed.

And that culminated in the catastrophe of FTX, with customer funds being stolen outright. That led to a catastrophic blowup. Then people look at that and they’re like, ‘Well, see, you can’t trust crypto.’ Of course, the real answer is you can’t trust fully unregulated crypto.

You actually have to have a framework. You have to have risk controls at these companies, compliance, auditing, and all the things that a grown-up financial services business has. That’s the long-winded way of getting to the point: we need a stable, permanent regulatory structure for crypto in the U.S. the same way we have for stocks, bonds, and everything else.

We need it both because we need this industry to succeed, which we and many people think is extremely important, and we can talk about that. We also need it so that it doesn’t lead to catastrophe, so that it doesn’t result in more FTXs, and so that people’s money doesn’t get stolen.

We need it so that the winners of the market aren’t fly-by-night operations operating on yachts off some coast somewhere. In a lot of ways, it’s the most obvious thing in the world. The U.S. financial system has been through this many times before. The most famous case is the Securities Acts, which created the SEC and regulated the stock market, after which U.S. capital markets became the envy of the world over the next 90 or 100 years.

We’re not looking for a free lunch. We’re not looking for subsidies, protectionism, or support in that way. We’re just looking for a permanent framework that lets people do business responsibly. I think it’s the most obvious thing in the world.

Robert: So you mentioned FTX.

I think everybody might remember that exchange and its big implosion and collapse and all the fallout from it. What is inside the CLARITY Act that would prevent something like that from happening again?

Chris: Right now, there is no federal regulator for crypto exchanges, as an example.

The New York Stock Exchange and Nasdaq have federal regulators. Right now, crypto exchanges like Coinbase, Kraken, and FTX do not. For every good company like Coinbase, there are many offshore ones that are not good. There is simply no regulator, and this is one of the most important things.

A big chunk of the CLARITY Act is to provide that, so that the SEC and CFTC have oversight. The act establishes disclosure regimes, anti-fraud regimes, insider-trading rules — all the things that we are used to with other financial markets.

It’s the same kinds of rules with the same agencies overseeing them. If you’re a federally registered crypto exchange, at an absolute minimum, you’re being audited. You have all sorts of controls and surveillance regimes. There’s a whole suite of things in the act that would empower federal regulators to do their job.

And if you’re not compliant with these things, you couldn’t operate in the U.S. It would have prevented that. Another big catastrophe was Terra-Luna. It was a stablecoin that wasn’t stable. It didn’t have dollars or any other kind of stable currency behind it.

It was essentially a self-referential thing in which the token’s value depended on the token itself. It should never have been presented as a stablecoin and would be illegal. It is illegal under the GENIUS framework.

Robert: Works well when things are going up.

Chris: A lot of things work well when things are going up and don’t work otherwise.

The details are hard, which is why some of us have been working on the CLARITY Act for seven years or so. It’s now been over a year since the House passed its version, and the Senate has been working on it for a year.

There are a lot of hard details, and it takes a lot of work across the aisle. This has been a very bipartisan effort. Financial industries have been regulated before, and the principles used in other industries are being applied here. It’s frustrating to read some of the news coverage of the released CLARITY Act text. As far as I can tell, the authors apparently don’t read the bill, because they say there aren’t safeguards such as AML and sanctions regimes.

This is simply not true. It has strict definitions of who is subject to these requirements. It applies the same money-laundering and Treasury rules to crypto intermediaries that are applied in other markets.

The Fraternal Order of Police just came out with its endorsement of the CLARITY Act, the largest law-enforcement organization in the country. Opponents are saying it doesn’t have proper law-enforcement clauses and rules, but this is simply not true.

Robert: Let’s talk about that. We’ll go over the fights going on with this bill. It’s 600-something pages at this point.

Chris: Drag it into ChatGPT and ask, ‘Does it address money laundering?’ You can drop it into an LLM.

It’s not that hard. If people want to actually learn about this, you can do it.

Chris: It’s no excuse.

Robert: Let’s go over each of these fights, each of these debates. You talked about illicit finance. That’s one of the big ones.

Robert: In fact, one of the biggest opponents to the bill is Senator Elizabeth Warren. She and her camp have called it a ticket to sanctions evasion.

That’s a quote from her. If you hear her describe it, she says North Korea, terrorists, and ransomware hackers are going to run wild if this bill passes.

What do you say to that?

Chris: Right now, it’s extremely unclear, for example, which entities are subject to what rules. What I’ve learned is that when you have gray areas in regulation, you essentially have a race to the bottom.

I was on the board of Coinbase for a long time. Marc is on the board. Coinbase is a U.S.-based company. The company takes regulation and compliance extremely seriously. But compliance costs a lot of money and slows down product development.

Every year, a new offshore competitor pops up that doesn’t do those things and becomes popular because it can offer lower fees or iterate its product faster. Over and over again, we’ve seen this phenomenon: we try to bet on the good, compliant actor, but for each one there’s an evil twin — an offshore doppelganger that copies all the features, skips the compliance because it’s cheaper, and gains popularity.

The ambiguity ends up favoring the bad actors. The important thing to do with good regulation, and I think the CLARITY Act does this, is clearly define the regulatory perimeter: who’s subject to this and who isn’t. And if you’re not subject to it, you can’t do business in the U.S.

If you’re in the U.S. and doing business, you have to be subject to it. You give clear definitions and clear rules. As an example, in the CLARITY Act, if you’re a financial intermediary, somebody like Coinbase or a company that’s holding customer funds, you’re subject to the same rules as fintech companies like Stripe or PayPal.

It’s black and white in the bill. I don’t know how to argue with people who are saying things that are factually untrue. As I mentioned, there’s the endorsement of various law-enforcement agencies. There are a lot of Republican senators who are very strong on national security, sanctions, and North Korea who support the bill, and a number of Democrats who we think will end up supporting it and are similarly strong on those issues.

So I don’t think the criticism is true, nor does the range of support for the bill suggest otherwise.

Robert: Marc, what do you think about this criticism that the bill enables sanctions evasion?

Marc: I’ve spent quite a lot of time in the national-security world.

The national-security people I talk to typically don’t agree with that. In fact, they typically say something quite different: ironically, they’re hoping more criminals and terrorists use crypto and blockchains because there’s a trail.

There’s a blockchain, and there’s a trail. That is contrasted specifically with how a lot of terror financing happens today. I’ll give you an example from the world of Middle Eastern terrorism.

There is an existing legacy payment system that is not the banks or any piece of technology. It’s called the hawala system. It’s an ancient, centuries-old system of informal peer-to-peer payments, and it operates fluidly across borders.

The way it works is you’ll have two cousins, one in one country and one in another country. You give money to the cousin in one country, and the money never actually moves. That cousin tells the other cousin that the family now has that amount of money, and as a result, the money effectively becomes unlocked in the other location.

You literally have peer-to-peer payments happening around the world with no actual transfer of physical cash, so you can’t catch it at the border. There’s no digital trail because nothing has happened digitally. No paper trail, nothing.

That, for example, is how a lot of terror plots are financed. There was a term national-security people were kicking around a while ago. They called crypto ‘prosecution futures’: if we could get the bad guys to use crypto instead of other methods, we would be able to mine the blockchain to prosecute down the road.

This is one where the whole thing has been incredibly confusing because it’s an upside-down, backwards thing. Not talking about anybody in office, but even people in the industry have said things like, ‘Crypto is inherently anonymous, and you can’t trace anything.’

Those are upside-down, bizarro-world claims, which I think stem from people not understanding the technology. The reality is, if more bad guys were running more money through crypto, I think more bad guys would get caught, not fewer.

Robert: There’s this thing that happens on TV shows — crime procedurals, cop procedurals — where they’re like, ‘I’m the criminal. Wire me the money in Bitcoin. Nobody’s going to know about it.’

Marc: That’s totally crazy. There are a couple of things.

The traditional thing they do in TV shows is to put a gun to somebody’s head and say, ‘Do a wire transfer. Wire transfer the money.’ Then they’ll show a digital display that goes from zero to 100 percent as the money moves, and the whole thing happens within two minutes.

In the real world, try to move money through the banking system across borders. Good luck. It takes days and days because the systems are shut down, they don’t work on weekends, and there’s the paperwork and everything else. It’s completely, flat-out implausible.

So you’re exactly right. What’s happened is that they’ve extrapolated that straight to crypto. Contrary to what a lot of people may think, television does not reflect reality.

Neither the existing payment system nor crypto works that way in real life.

Chris: It’s actually a nontrivial research and product-development problem to build blockchains that have privacy, because they are so public.

We think that’s important, not for nefarious reasons. If you want to pay someone for a service — maybe it’s a private medical or financial matter — you don’t want everyone to know about it, just like in the regular financial system.

You’d want some privacy when sending someone a digital dollar. It seems natural. For a long time, everyone said, ‘Why would you want to build privacy? It’s only for criminals.’ Since GENIUS passed and dollar-denominated stablecoins have gotten more popular, we’ve heard people in D.C. say, ‘What about privacy? You need to have privacy.’

We agree. It’s reminiscent of the early internet. Marc, wasn’t it Netscape that invented SSL?

Marc: That’s right. We did.

Chris: And so, for the kids who don’t know this, HTTPS — that was Netscape.

That was an invention. At the time, it was like, ‘Who would need that? Why would you need to send secret stuff?’ Probably some bad guys did use it, but over time we all realized it was important, and 99.9 percent of it is good behavior.

It’s just people interacting with their bank or something. You guys had to go in front of Congress.

Marc: It was a four-year fight. It was actually quite analogous to this. Netscape was the first widely used piece of consumer software that incorporated modern public-key encryption.

At the time, we had to ship two different versions. We could ship a strong-encryption version inside the U.S., but encryption was classified as a munition under what were called ITAR export-control rules at the time.

Encryption was classified in the same category as a Tomahawk missile, which meant the Netscape browser was classified in the same category as a Tomahawk missile. We had to ship deliberately insecure versions of the product overseas, and you can imagine how much our overseas customers enjoyed that.

We clearly labeled it on the box: weak encryption, basically, do not trust this. You can imagine how that went for sales. Then foreign competitors immediately stepped in, started cloning it, and started making strong encryption.

People outside the U.S. just didn’t use our product. They used products made outside the U.S. We went to Washington, and as you do when you’re a kid, you explain things logically. They tend to stare at you like you’ve grown a third eye, so you start to get into the specifics and details.

But then you get into this fundamental dichotomy: is encryption bad because bad guys are going to do bad things with it? Or is it good because it is the key to establishing trust, enabling business, and allowing law-abiding citizens, both here and overseas, to cooperate and conduct business?

Do you want to completely eliminate the risk of any cybercriminality, or do you want to make Amazon.com possible? Those questions are deeply intertwined. It was a very long education process.

It took four years. It took a very long time to unwind. They ultimately made that change. The world did not end, for several reasons. One is, as you said, the use cases were overwhelmingly positive.

The intelligence services adapted in many ways. The other incredibly positive thing that happened was that American industry won. Not just Netscape, but many others. The global internet economy is dominated by American companies.

And if you’re in the national-security state and you ask whether you’d rather have this giant industry dominated by American or non-American companies, 100 percent of the time they’ll tell you they want it dominated by American companies because they can deal with American companies.

They can’t deal with some company running in another jurisdiction. A lot of times, they can’t even talk to them. It was a classic case study. That one took four years.

This one, Chris, we’re now in what, year seven? For people who think the modern world is speeding up, at least in this respect, it’s slowing down. This is why there’s so much pressure right now to get CLARITY passed: seven years to work through these issues is clearly enough.

Robert: Let’s talk about some of the other objections to the bill. Another big one is ethics reform. The president and his family members have interests in crypto businesses. This will come as a surprise to nobody who’s been reading the news. Critics say this bill will enrich the people who are most vocally supportive of it. How do you address that?

Chris: First of all, it’s not my area of expertise, but as a citizen, I think there should be ethics rules for government officials. I don’t think those should be exclusive to crypto. I think there should be ethics rules around trading stocks, crypto, and other financial assets.

That seems reasonable to me. That’s number one. Number two, the CLARITY Act, even without specific ethics provisions, will add significant restrictions to anyone in crypto, including disclosure requirements around the risks and holdings of a crypto asset. It will add lockup requirements.

That is an increase in the restrictions applied to anybody, including government officials. Point number three: what has been frustrating from a political point of view is that the specific ethics provision around crypto has been coupled with the topic we’ve been discussing, which is regulating this industry.

This bill, as proposed right now, would be the first bill in U.S. history, I’m told, that regulates an industry and also adds specific ethics provisions for government officials. This is highly unusual. Normally, government ethics rules are done separately, and an industry is regulated separately.

I think this industry is being held to a completely different standard than other industries. That said, I think it’s reasonable for people, politicians, and policymakers to debate and figure out what the ethics rules should be for all financial assets, including crypto.

Right now, there is an active negotiation. I hope there’s a resolution. It’s between the various politicians, not the industry. I hope they can figure something out so that we can move forward, because we want to regulate this industry.

Marc: This bill puts controls on government officials’ use of crypto that are stronger than those that apply to government officials’ stock trading. If you don’t like how government officials trade stocks, this bill puts much stronger controls in place for crypto.

Chris: One other point I would add is that most people think of crypto as a thing you can trade. What they’re missing is that increasingly, like the stablecoins we’ve been discussing, it’s a thing you can use. If the industry is successful, you’ll have more and more use cases and more ways in which it’s embedded into our financial life.

It’ll be normal for a kid to buy tokenized stocks instead of going to a traditional stockbroker. It’s important with these rules to separate trading behavior from use behavior, because it’s important for government officials to have access to modern technology, and I think this is the future of finance.

Robert: Okay, so there are government ethics rules in the bill, but it’s also a separate issue from creating market-structure rules for crypto. Maybe these are separable issues.

Chris: The reality of where the politics are right now is that there are ethics provisions in the current bill. They’re still being negotiated.

I hope they come to a resolution. But from our perspective, after this many years of regulatory gray area and an environment that favors a race to the bottom, the important thing is to get a regulatory framework passed. We think the CLARITY Act, while not perfect, is significantly better than the status quo.

And so I hope those issues are resolved.

Robert: You mentioned stablecoins again. One of the big fights has been over the treatment of stablecoins in this bill. In particular, the banking lobby, most notably JPMorgan Chase, has been very vocal about this.

They are unhappy with the idea that you could earn interest on stablecoin holdings. The fear is that if consumers can earn interest on stablecoins, they’re going to pull their money out of banks and keep it in these other accounts. That’s going to cause deposit flight.

Robert: And they don’t want that to happen. So what’s the state of that debate right now?

Chris: In the bill, I would say the banks got mostly what they wanted: you cannot pay interest on balances. I think the exact language covers anything functionally or economically similar to a bank account, and there are further legal clauses.

What you can do is something more complicated. For example, if somebody completes multiple transactions and a formula is used that isn’t functionally or economically similar to paying interest on a balance, you can give them rewards for using their stablecoin wallet twice a month at Walmart.

Walmart could give them some kind of reward back.

Robert: A credit-card kind of program.

Chris: That’s what the banks had asked for. There was a prolonged discussion — it’s all been in the press — and they seemed to get what they had asked for, but then seemed to want more.

I think going any farther, you start to ban things like Starbucks rewards points. It’s gotten about as far as it can go, and it was a compromise that was somewhat painful for the crypto industry.

But in the context of the broader bill, it still nets out very positively, and we’re very supportive of the overall bill.

Robert: The thing I find so interesting is that this debate has been embodied by Brian Armstrong versus Jamie Dimon. They’re the two leading voices on this. But JPMorgan has a big blockchain unit. They’re working on this stuff. They have tokenized deposits that are live onchain. So it’s interesting that they’re trying to slow-roll CLARITY while also building on this technology.

Chris: These are massive organizations. They’re almost like countries, so there are many different factions and things within JPMorgan.

But as far as I know, every big bank, including JPMorgan, has significant blockchain efforts. If CLARITY passes, a lot of those efforts will roll out in a significant way. We’re way past the point where enterprises dip their toe into new technology and do experimental stuff.

We’re way past that now. We speak to a lot of these organizations regularly. They have real and significant deployments that are going to go live. Some are live, and some are going to go live with greater regulatory clarity. I think they see significant opportunity. You have to remember, the banking system is the way it is for a bunch of reasons, and it’s not necessarily their fault.

These systems are very complex and intertwined with each other, but a lot of the technology is antiquated. It’s a joke in Silicon Valley that banks still have COBOL programmers, but it’s actually true because a lot of them have very old codebases.

In their defense, the systems are so intertwined that you can’t just upgrade your software by yourself. You have to do it all together. One of the things blockchains have done for the financial industry is give them a unified framework where they can all say, ‘Together, let’s move into the 21st century.’

That has solved not just a technology problem, but a coordination problem. If you speak to them, as we have, they’re all very excited about that. The other thing I’ll say is that the simple way to think about blockchains is that they remove layers of mediation. They’re disintermediation machines.

Blockchains remove many of the layers involved in sending money through a wire. We had a case where we funded an international portfolio company. Two weeks later, we couldn’t find the money. This was a significant financing. And it turns out, if you dig into it, there is no international wire system.

You wire to a bank, the bank has some piece of paper that gets moved somewhere else, and they wire another bank. There is no global financial network the way there is a global internet or WhatsApp network. It’s a patchwork of systems. One of the beautiful things about stablecoins is that you’re building something the way you would build it on the internet: one big over-the-top network.

The banks see that opportunity. They say, ‘We can save a lot of money. We can provide better products. We can modernize a whole bunch of things.’ We see across the board that a lot of these organizations are very excited about this, including JPMorgan.

Robert: So this is an opportunity for a big technological upgrade. All these systems you mentioned have parts still running on COBOL, many decades old, and this is a chance to move off them.

Chris: And don’t take it from me. Goldman Sachs CEO David Solomon just came out endorsing the CLARITY Act, along with Fidelity, BlackRock, and a bunch of other major financial organizations.

They all have major efforts. Much of this is public. The most innovative fintech companies, like Stripe, have gone in a big way. You don’t have to listen to me; just look at the news. At the beginning, we were talking about the early Bitcoin days. We’re very far away from that now.

These are very big organizations, and the technology is on the verge of going mainstream with the right regulation.

Robert: There are a lot of endorsements and a lot of support. The one that sticks out for me, which you mentioned earlier, is that a lot of the biggest law-enforcement organizations are for it. I think that’s a major tell. I want to shift to another debate over the CLARITY Act: liability for developers. Former White House cybersecurity official Carole House has raised the concern that developers should be held more liable for the software they write.

She argues that if you don’t hold them liable, it could set a dangerous precedent for other areas of technology, like AI. Marc, I know you spend a lot of your time in the AI world working on this stuff. What do you think about this argument?

Marc: It’s a kill shot to the industry.

It’s impossible. It makes software development impossible, because how can any software developer anticipate the use of the software down the road? You don’t even need to think about this only in terms of software; think about any product. If I run a hotel and a criminal stays at the hotel and plans an operation, does that make me part of the conspiracy?

If I’m an engineer who designed a car and the car is used in a bank robbery, does that make me an accessory to bank robbery? It’s completely insane. If you saddle people who build products used by many people for many reasons with downstream liability for those uses, it kills the industry, which is clearly the goal.

Chris: And just to add to that, it’s important not to conflate things. In some of these court cases, those things have been conflated. To use Marc’s analogy, if you build a car and then knowingly help a criminal, then you are in trouble.

No one is debating that. If you build a piece of software and say, ‘Hey, criminals, come use it,’ and they send you an email asking, ‘How do I do a crime?’ and you say, ‘Here’s how you do a crime,’ that is a crime. There have been cases like that in the software world, and no one is debating it.

What we’re talking about here is making open-source software, whether it’s AI or crypto, for a constructive use case. You’re making a car, a hammer, an AI model, or a blockchain, and then, for example, making it open source.

If you add unlimited downstream liability, whether criminal, civil, or otherwise, how could you ever decide to do that as an open-source developer with that kind of risk? You couldn’t. In the AI world, I think there have been a few bills, including in California, that almost passed and would have imposed basically unlimited downstream liability on AI models.

These people aren’t big companies that can afford to defend themselves. Are five guys in a garage going to take unlimited liability? It’s going to kill it. It’s going to kill the whole thing. It’s going to kill open source.

Marc: This is the thing.

This is a very live issue in AI policy, and people are trying to do this in AI, which I think clearly shows what they’re trying to do. First, open source dies. Right out of the gate, open source dies, because open-source developers are not getting paid to start with.

They certainly can’t take on enormous liability exposure. Even if insurance were available to cover unlimited downstream liability for a product used by millions of people, open-source developers couldn’t afford it.

So open source dies. Because open source dies, academic research dies, because without open source, there is no computer-science research in any of these fields. It’s entirely dependent on open source. It kills computer science as a field, and that’s just the start. Then it kills venture investing.

We obviously can’t invest in a company if it has a million times the level of exposure because somebody does something the company didn’t even know about. Then it kills all the companies, all the startups, and then the big companies, because they can’t take that on either.

It rips right through. The understanding of what it takes to run a business or do anything productive is either completely absent when people make that argument, or they are deliberately trying to kill the industry.

I think they’re deliberately trying to kill the industry.

Robert: Maybe. Let’s move on to one of the final objections, and one I’ve heard most often: that this bill would punch a hole in securities laws. Marc, you mentioned that securities laws have allowed U.S. capital markets to thrive over the past 90-some-odd years and have made the U.S. the envy of the world.

Some people say that if this bill gets through, a company will say, ‘I’m going to tokenize some asset, put it on a blockchain, and that is going to exempt me from SEC oversight and securities law in general.’ What’s your response to that view?

Chris: No. This is in the bill. If you take a stock and tokenize it, that’s a security, and it’s regulated by the SEC like a security. It’s very black and white and clear. The only thing that is different with respect to blockchain regulation is that the CLARITY Act delineates certain situations in which a token like Bitcoin or Ethereum will be regulated by the SEC.

In other cases, it will be regulated by the CFTC. In all cases, it has a federal regulator. Even the last administration implicitly agreed with this framework. Let me walk you through it. Bitcoin, when it started, was presumably created by one person or a group of people — whoever Satoshi was. By definition, when something starts, it’s centralized.

There’s a person behind it with inside information and control. The way the CLARITY Act works, at that point, the token is regulated by the SEC. If someone creates a new blockchain, they have control over it, and there are rules such as lockups.

There are disclosure rules and all the things you expect from securities laws. Over time, when you hit certain thresholds of decentralization, it starts to look more like Bitcoin or Ethereum today, where there is no central actor that controls it or has inside information.

The CLARITY Act designates that the CFTC oversees that token, so it’s regulated as a commodity and not a security. At that point, you still need regulation to make sure people aren’t dumping on the market, cornering the market, or doing all the other things commodities laws regulate.

But it’s a different regulatory regime depending on the nature of the asset.

Robert: So it becomes more like gold or precious metals.

Chris: That’s right. It’s like Bitcoin. Even the last administration said at some point that Bitcoin and Ethereum were sufficiently decentralized to be regulated as commodities.

This has been implicit in court cases and agency decisions from both parties for the last 10 years. All CLARITY does is enshrine it in law and make the definitions specific, so people know exactly where they are and don’t have to go to court to figure it out. It’s taking what has been the consensus across courts and regulators, hard-coding it, and making it specific.

First, existing assets are still securities when you put them on the blockchain. Second, all digital assets have a federal regulator and a framework. Today, if I create a token, the biggest issue is that there are no rules for disclosure.

There are no rules for insider trading or lockups. If the CLARITY Act passes, our effective lockup period — the period in which we can sell — will increase significantly. We’ll invest in someone, they’ll launch a product with a token, and until the criteria for sufficient decentralization are met, there are lockups on venture capitalists, founders, and so forth.

That is how it should be. It’s a smart, risk-based regulatory framework that we think will allow people to build long-term products and build trust in the market. Consumers, investors, and all market participants will have trust in the market because it’s regulated and there’s a framework around it.

And that’s how you build a real industry for the long term.

Robert: We’ve walked through a number of the live fights. All these things are being debated right now. Any one of them could potentially tank this bill. What happens if CLARITY does not pass?

Chris: Number one, we’ll keep working on it, and someday it will pass.

Number two, a lot of these things can be and are being done at the agency level, meaning the SEC, CFTC, Treasury, and so forth. The big difference is that legislation is de facto permanent in a way that agency rulemaking is not.

That’s important both for ensuring that the rules are in place over the long term to protect consumers and for giving the industry confidence. You want to make an investment or build something, and those things can take many years. If you have to build on shifting sands, you’re much less likely to make that investment of time and money.

It’s much harder. You’ve got enough to worry about as an entrepreneur without a shifting maze of regulatory changes. That would be the downside: it would prolong the uncertainty. But I’m optimistic it will get passed in the near future. If it doesn’t, we will keep working on it.

Robert: Some people have this idea that regulation and innovation are opposites. They’re at odds with one another. Marc, you spend a lot of your time on what you call the Little Tech agenda, advocating for startups, builders, and developers. How does all of this fit into that regime, and why is regulation good in this case?

Marc: Political theorists talk about all the different forms of government and debate how well they work. The one I think everybody agrees on is a specific degenerate form of government called anarcho-tyranny.

The term combines anarchy and tyranny. Anarcho-tyranny is essentially a government in which you regulate law-abiding people to death. The anarcho part is that you let the rule breakers go absolutely nuts. The tyranny part is that you regulate or prosecute the good actors to death.

You encourage the criminals and condemn the law-abiding. That was the governance regime under the prior administration. That was the system, and startups felt it very directly.

It happened in two ways. Startups outside the U.S. went wild and did whatever they wanted, which, as we discussed, led to FTX. Startups inside the U.S. got brutally punished, with no route to safety and no way to conduct business safely.

If you were designing from scratch and asking, ‘What’s the worst possible policy? What’s the worst possible approach?’ it would be anarcho-tyranny, which is what we had. What do you actually want? You want predictability, stability, prudence, reasonableness, and protections at the right level.

You want consumer protections and investor protections. You want things to happen in a fair way. But you also want freedom and the ability to innovate. You want the ability to deploy high-quality products that work the way they say they do.

You want customers to be able to trust you, and you want to know you’re going to be in business in a year without being wantonly attacked and destroyed by a system that gives you no way to do it legally. Like a lot of things in life, it turns out there is a middle ground.

There is a centrist position. The centrist position involves the correct level of regulation. Either side of that is devastating: not enough is devastating; too much is devastating. You want to Goldilocks it right in the middle, which is very much what we think CLARITY does.

Robert: Let’s zoom out and talk about what’s at stake for the country and for American leadership. Why is passing the CLARITY Act so important for the country? What will it lead to?

Marc: We would argue this is in the long tradition of American technological leadership, and I would say American technological supremacy. Do you want to be the country that leads the world in technology or not?

There are two parts to that. One is whether you want the technology to exist; some people argue you’re better off without certain technologies. But technologies tend to exist once they get invented. Then, do you want that to happen inside your own country, or somewhere else?

We always make a very strong argument on this: every American citizen, regardless of political position, should want America to be the technology leader in the world. That has many benefits. It has direct economic benefits in terms of the wealth of the country and the ability to pay for all the things we want to pay for.

It has enormous benefits for our security. For the crypto industry to be based in the U.S. is overwhelmingly good for U.S. law enforcement and national security.

If you’re the FBI or any agency of government, you’re dealing with American companies, which is much more straightforward when you need something. It’s good from a security perspective.

I think it’s also good in terms of opportunities for kids. Are your kids able to get educated in state-of-the-art technologies, have lots of career opportunities, and work in those fields? American kids have had that edge on kids all over the planet for a hundred years, and hopefully that continues.

It’s one of these amazing things where we’re so used to America being the technology leader that it almost feels embarrassing to make the argument because it’s so obvious. America has clearly benefited from this for the last hundred years.

In my view, we should all want that to continue for hundreds of years to come.

Robert: Chris, you literally wrote the book on this stuff, on blockchains, Read Write Own. Tell us, what does CLARITY unlock if it passes in one year, two years, or farther into the future?

Chris: As I mentioned before, a simple way to think of a blockchain is as a disintermediation machine.

It lets you build internet services where, instead of having a bunch of organizations in the middle taking a cut of the money along the way, you can build advanced, peer-to-peer internet services. I can send you a dollar. A stablecoin is a simple example.

I can send you a dollar with no intermediary taking a cut. If you send money to Mexico right now using, they charge 8 to 10 percent. With a stablecoin, it’s almost zero because there’s no one sitting in the middle. That’s the simplest case. The use cases that have worked so far have been financial.

We’re seeing stablecoins, stocks, bonds, Treasury bills, lending, and other kinds of financial assets. I think that will be the dominant use case in the next couple of years, so I think we’ll see what I hope is a significant upgrade of the financial system. A lot of the benefits are global.

But I think we take for granted here that we have a high-functioning currency and financial system. This technology makes those financial benefits accessible globally.

Longer-term, as I talk about in my book, you could take this technology and apply it beyond that. You can imagine a world where, if AI agents proliferate the way we hope they will, and the way a lot of people in the AI industry think they will, there could be an internet of billions and trillions of AI agents conducting economic transactions. The natural way they would do that would be through crypto assets.

You can also imagine all sorts of other future use cases, which I allude to at the end of my book: services for creative people, social networks, games, and more.

But for the next couple of years, I think we’ll mostly be focused on this finance 2.0 upgrade that we talked about.

Robert: All right. Time will tell whether policymakers listen to you, heed your advice, and realize that rules are better than no rules, that the time is urgent, and that the benefits will accrue to whoever sets the standards first.

Thank you both for coming on the show.

Chris: Thank you.

Marc: Thanks, Robert.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

For centuries, the central challenge of commerce has been the same: How do you get people with different roles, different information, and different incentives to work together toward a common purpose? The answer has almost always involved some form of organizational innovation: new structures that allocate risk, reward, and responsibility in ways the previous generation couldn’t. The history of commerce is also a history of coordination.

The corporation was the last great organizational leap, built for the industrial age and the coordination problems (and opportunities) it created. But software and internet-native protocols can now reduce the need for much of the previously unavoidable overhead — the many layers of centralized management, bureaucratic bloat, and intermediation — of the traditional firm.

Existing legal structures were not designed for this new world. Only one relatively new entrant is emerging as a serious contender for the next organizational leap: The DUNA, the sole legal entity explicitly recognized by once-in-a-generation market structure legislation now advancing through Congress. It’s arguably the only structure credibly built for internet-native organizations.

To understand why new organizational forms are emerging today, it helps to remember what problem the corporation solved in the first place — and where we’re going.

How merchants managed risk

Before corporations, business was personal: Imagine Marco Polo undertaking long-distance sales calls with his father and uncle. Family businesses like these quite literally took their lives in their hands. If a contract went south, their personal fortunes could be wiped out — or worse. Their very lives could be at stake.

Merchant ventures largely relied on two kinds of protection, but neither was guaranteed. The first was geopolitical: the relative peace of the Mongol Empire’s Pax Mongolica. If you attacked someone the Mongols liked, then you would be in trouble with the notoriously brutal Mongols. The second was social: If you cheated someone or defaulted on a contract and so violated the Lex Mercatoria — a self-enforced merchant honor code (ca. 1100–1600) — your reputation would be sullied, and you could be blackballed from the trade from Quanzhou to Timbuktu.

A merchant’s word really was, in the absence of strong institutions, worth more than gold. The Polo family had it somewhat easy, too, since they were bound by blood. Many other business partnerships have been less sanguine.

One of the most persistent issues businesses face is tensions between principals and agents; or, in this case, investors and merchants. One innovation was the medieval commenda, which provided some limited liability: Investors could only lose what they put in, and the same (theoretically) went for the merchants. The partners split the profits according to their original contributions. Commendas arose organically, long before there were any formal statutes. Yet every concern was still one small shock away from going under. And the model didn’t scale: Commendas dissolved at the end of a voyage, or in the event of bankruptcy. Or death.

A further innovation was the Florentine compagnia. Think: Medici bank. This form was a more durable and operationally sophisticated legal entity than the commenda. A compagnia could sustain ongoing multi-partner commercial relationships, but it continued to rest on the personal liability of all partners. This was the most advanced pre-corporate instrument yet devised — the high-water mark of medieval partnership — yet it still left its partners exposed. Unlike the Church and universities, which had long enjoyed legal personhood derived from Roman concepts of the universitas (a collective entity treated as a single legal person), commercial enterprises still lacked a fully separate legal identity.

These shortcomings wouldn’t be remedied until the 17th century, when early modern Europe came up with something new. This innovation, and its legal protections, allowed ventures to more easily raise capital, distribute ownership through stock offerings, and shield owners from liability; it was the corporation. Corporate powers, famously granted to the Dutch East India Company (or VOC: Vereenigde Oostindische Compagnie), arrived as a revelation and spread to the rest of Europe once it became clear what a good idea corporations were. (While the English East India Company preceded the VOC by a couple of years, it was far less sophisticated, raising capital only for specific voyages, and lacking a mechanism to raise public equity.)

By lowering the risk of doing business, and reducing coordination costs, corporations made large-scale, capital-intensive ventures — and much of the modern world — possible.

The price of scale

While solving a host of real problems, the corporation introduced new ones. Its first achievement was making participants care about one another’s outcomes: By tying shareholders, directors, and captains to the same legal entity and the same bottom line, the corporation forced each party to internalize costs it would otherwise impose on the others with impunity. But shared stakes are not the same as perfectly aligned incentives.

Let’s consider the VOC as an example. Its legal form is familiar but complicated: Shareholders included many Dutch citizens eager for a return on their investments, but who were too busy with their lives to be concerned with the VOC’s day-to-day operations or its grand strategy. The board of directors, the “Heeren XVII,” made plans for how everyone was going to make money. Ship captains and traders on the ground in Southeast Asia had to make the best choices for the company in the moment with limited information and resources.

In theory. In reality, these three parties’ interests were not perfectly aligned; they diverged slightly insofar as one party could make more money for themselves at the expense of others.

How do you make sure the captains, far removed from the oversight and control of the Heeren XVII, don’t ransack another ship and abscond with the loot? Prevent the traders from taking bribes, or cutting larger deals for themselves? Ensure the board is making the right calls? What if you’re a group of shareholders who also cleave to Protestant values and have concerns about the VOC’s sometimes rapacious practices? Questions like these led to all sorts of innovation in incentive design — options like bonuses, profit-sharing, audits, surveillance, and some might even say efficiency wages — and to new legal protections where the state would ensure fair play. It also led to untold abuses.

And yet! The corporation as it evolved over time remains the best thing we’ve got for aligning incentives, reducing coordination costs, generating profits, and protecting everyone involved in a venture.

Shortly after the United States was formed, it recognized corporations through special legislative charter, but these were rare. The First Bank of the United States, chartered by Congress in 1791, was both the first and most famous example. New York introduced the first general incorporation statutes in 1811. By the mid-19th century, more states allowed incorporation without special acts and the idea of “limited liability” became more standardized across states. An explosion of corporations followed during the phase of late 19th century industrialization, culminating in the Delaware General Corporation Law of 1899.

Cooperatives were another option that gained momentum in the 19th century. They experimented with a different solution to the problem of coordination: member ownership and democratic governance. Farmers, consumers, workers, and credit unions used cooperatives to align the interests of participants more directly with the organization itself. While cooperatives succeeded in some settings, like agriculture (see for example Land O’Lakes), they generally remained more specialized. The corporation, meanwhile, grew more and more popular.¹

Another solution was the limited liability company, or LLC. While the LLC had its precursors in older forms like the German GmbH or the United Kingdom’s limited company (Ltd.), the LLC itself is surprisingly modern: Wyoming introduced it into law only in 1977. Prior to this, corporations did offer limited liability but were rigid and subject to double taxation, while partnerships were flexible but exposed participants to personal risk. The LLC combined elements of each — limited liability with pass-through taxation — making it a better fit for a wide range of smaller ventures. Today, they’ve become a default form for many startups, small businesses, and investment vehicles.²

A succession of minor variations have followed: the LLP, or limited liability partnership (1991); the L3C, or low-profit limited liability company (2008); the benefit corporation (2010); and so on. These are certainly useful, refining the corporation for particular uses. But every now and then, technology changes what’s possible and you get a form that is, in comparison, revolutionary.

DAOs and their discontents

Decentralization is one such revolutionary idea: the ability for large groups of people to coordinate without centralized management or trusted intermediaries.

Before crypto — and especially before Satoshi Nakamoto’s invention of the blockchain — that possibility was more philosophical than practical. One of crypto’s first great innovations was the DAO, the decentralized autonomous organization. A DAO is an organization governed by rules encoded in software and collectively managed by its participants rather than a central authority. There is no centralized management team or board of directors, no Heeren XVII.

But decentralized governance is hard. Getting token-holders to vote on important issues has proven even harder than getting individual shareholders to vote on board members, a process that already has abysmally low voting rates, on par with U.S. municipal elections. Ensuring that power isn’t concentrated in the hands of only a few token holders can also be challenging.

The recent legal environment only exacerbated these challenges. Regrettably, the previous administration’s Securities and Exchange Commission declined to provide clear rules for crypto projects and also weaponized that ambiguity through aggressive enforcement actions against the industry. Entrepreneurship doesn’t thrive under uncertainty; it’s hard enough even when you do know the rules.

At the heart of this question of legality is one of the three prongs of the so-called Howey test, which the SEC uses to evaluate whether an instrument is a security: (1) An investment of money; (2) in a common enterprise; (3) with profits to come solely from the efforts of others. Asking whether the value of the asset depends on the “efforts of others” in the case of a public company would include, for example, the managers who run the firm. In the case of crypto projects and their DAOs, the SEC noted that the ongoing development of the protocol, even if it were done by a bunch of unaffiliated people who may or may not hold the token, subjected a related token to securities laws — making widespread participation and onchain transactions impossible.

Equally importantly, since DAOs weren’t officially recognized by the state, the owners of a project wouldn’t have any of the protections mentioned earlier, like limited liability. In other words, members of a DAO could face unlimited personal liability, putting crypto governance on a footing that is, legally speaking, almost medieval.

So crypto projects did what their lawyers told them. They set up offshore foundations as separate entities that oversaw the ongoing development of protocols, severing the connection between those efforts and the U.S. businesses. Or they established themselves outside the United States. Either “solution” was to the detriment of innovation in the U.S., and to U.S. jobs and tax revenues.

Offshore crypto foundations were, to put it kindly, convoluted. These lawyer-concocted workarounds shifted authority and ongoing development efforts to an “independent” entity in hopes of avoiding securities regulation. The strategy was understandable at a time of regulatory hostility, but it also exposed deep flaws: Foundations have weak incentive alignment, limited ability to drive growth, and an inevitable tendency to entrench centralized control.

But what choice did projects have when stuck between the rock of “get sued by the SEC” and the hard place of “set up a weird organizational structure that’s going to create incentive alignment issues”?

This is why the DUNA — the decentralized unincorporated nonprofit association — is so significant. It draws on the long history of structuring ventures and governance to achieve the goals of every business ever: coordinating people efficiently around a common purpose. But it does so without centralized managerial control, reducing the principal-agent problems and information asymmetries that arise in traditional firms. In doing so, the DUNA departs from one of the core assumptions underlying the Howey test: that participants are relying on the managerial efforts of others to generate value.³

The crowd gets its own legal form

Before the DUNA, there were three options for organizing around and governing a crypto project: a DAO, which lacked legal recognition and exposed members to potentially ruinous liability; a traditional corporate entity, which forced projects into ill-fitting hierarchical structures and exposed them to action from the SEC; or an offshore foundation, which was legally and practically convoluted and pushed much of the industry abroad.

Until recently, there was no clean way for a group of users to govern a decentralized network — a form of organization that blockchain technologies have only recently made possible — while maintaining some of the protections enjoyed by companies. Now there is.

Put simply, the DUNA turns a crowd into a legal entity. Three states so far — Alabama, West Virginia, and Wyoming — have passed laws authorizing this new business structure, which combines the legal superpowers of existing forms with the ability to decentralize control, something very different from the traditional corporation and which no previous entities have ever fully achieved.

What protections does the DUNA provide more specifically? Its powers include legal personhood, limited liability, persistence over time, and state recognition — many of the same elements that make modern corporations possible. Recognition of the group’s “personhood” allows the entity to enter into contracts on behalf of its participants, while limited liability ensures that members aren’t held personally liable for organizational obligations. Together, these features make it possible for large, loosely knit groups of people to coordinate — raise capital, hold assets, hire administrators, pay taxes, make deals — without putting members at undue risk and exposing them to potentially ruinous liability.

Organizational forms don’t take hold all at once; they spread as jurisdictions compete, as lawyers grow comfortable with them, and as entrepreneurs begin to trust them. Before Delaware became the venue of choice for incorporation, New Jersey held sway⁴; now, Texas and Nevada are gaining ground. The LLC first received approval in Wyoming and then migrated, once its tax treatment was clarified, to all fifty states by 1997. As for the DUNA, Wyoming was once again the pioneer, enacting it as law in March 2024. Already, crypto protocols and communities including Uniswap Governance and Nouns DAO have implemented it.

Much as the corporation gave large-scale enterprise its first native form, the DUNA is giving open, internet-scale, decentralized networks a legal form of their own.

A new era of organizational design

Think of the DUNA as a legal wrapper that allows the governance mechanism of a decentralized network to do business, but without introducing traditional centralized management. It’s built on the unincorporated nonprofit association (or UNA), a body of law adopted by 17 states and Washington, D.C., that helps groups like homeowners and civic associations, recreational sports leagues, religious congregations, and hobby clubs to organize under the law. UNAs provide lightweight governance without the overhead of a corporation or LLC, allowing these groups to own property, enter contracts, and sue (or be sued) as an entity.⁵

The DUNA is similar: It allows a group of tokenholders or contributors to govern through onchain rules or token-based voting rather than a board or management team. Members get the protections of limited liability, separating the entity’s obligations from their personal assets; and the organization can be understood by and interact with courts, regulators, and counterparties.

But the DUNA doesn’t solve every problem. It doesn’t eliminate governance challenges. It doesn’t guarantee decentralization (although to qualify as a DUNA, a DAO must have at least 100 active members). And it doesn’t magically bypass securities law. What it does do is close a specific gap, making it possible for a decentralized organization to be a legally recognized one.

• • •

From informal merchant networks, to partnerships, to corporations, to LLCs, to now, DAOs, each new organizational technology emerges when people need new modes of coordination. The DUNA may mark the beginning of a new era in this evolution of organizational design. But just as the corporation didn’t replace all partnerships, the DUNA doesn’t supplant everything that came before. It merely expands the menu. And for the first time, it makes possible decentralized networks that are represented by fully legible legal entities.

For most of human history, organizing at scale — even small scale — meant taking on enormous personal risk. Daring entrepreneurs, like the Polos, relied on family, reputation, and fragile norms to hold things together, always one shipwreck away from ruin.⁶ The corporation changed the calculus by separating the fortunes of a venture from the fortunes of the people behind it. The DUNA extends that separation into a new domain: communities governing decentralized networks based on blockchains.

Now, even a loosely organized network of strangers on the internet can act as a single entity — make deals, hold assets, take risks — without any one participant having to bet their livelihood on the outcome. In this way, it’s a new answer to one of commerce’s oldest problems.

• • •

Acknowledgments: We’d like to thank Aiden Slavin, Alejandro Flores, Miles Jennings, Scott Kominers, Sonal Chokshi, and Steph Zinn for their helpful comments and emendations. Any errors remain our own.

Notes

1. While cooperatives may seem spiritually aligned with internet-native organizations like DAOs, co-ops presuppose a relatively stable set of identifiable members and hierarchical leadership structures that many decentralized networks lack. ↩
2. Curiously, one of the United States’ other big contributions — corporate bankruptcy law — was for a long period not widely copied around the globe. This codifies the idea that someone can take a risk, fail, reorganize, and try again, and it is an engine of American dynamism. ↩
3. Wyoming had attempted to solve the problem in 2021, letting DAOs organize as LLCs. But an LLC still assumes a defined membership list, K-1 tax filings, and a profit-making purpose. While it was a good fit for some small investment clubs, it was an awkward fit for a permissionless, pseudonymous network governed by a nonprofit mission, and did little to resolve the Howey question of whether membership interests themselves counted as securities. ↩
4. That is, until Woodrow Wilson, then governor of New Jersey, did Delaware a massive favor by cracking down on corporate-friendly incorporation laws in his home state. ↩
5. While trusts might seem like a natural vehicle for decentralized groups, they are a poor fit. Trusts are designed around the relationships between identifiable fiduciaries and beneficiaries, which makes them an awkward choice for organizations whose governance is designed to be intentionally diffuse. ↩
6. Incidentally, after commanding a Venetian galley in a war between rival trading powers, Marco Polo was captured and thrown into a Genoese prison, where he would dictate his famous travelogue. ↩

In the 1990s, it was the internet. Policymakers worried that it would disrupt existing industries and create new risks. But instead of cramming the technology into outdated regulatory boxes, lawmakers established rules that let innovation flourish while protecting consumers. The result was one of the greatest periods of economic growth in American history.

Today, we face a similar opportunity with blockchain networks. Stablecoins are making payments faster and cheaper. Tokenization is modernizing capital markets. Both depend on blockchain infrastructure. 

These technologies will advance with or without Congress. It’s up to the United States whether it will set the standards, or cede that ground to others.

The GENIUS Act proved that clarity works

The GENIUS Act showed what well-considered policymaking can achieve. Stablecoins are digital dollars that move across the internet as easily as bits. Sending $200 from the U.S. to Colombia via traditional rails can cost more than $12 and take days. With stablecoins, the same transfer costs pennies and settles in seconds.

Regulatory uncertainty had for years held back stablecoin adoption. The GENIUS Act changed this. By establishing reserve requirements and a framework for issuers, it got rid of any confusion and unlocked growth.

The results speak for themselves. The stablecoin market amounts to roughly $315 billion, up more than 50% from a year ago, with dollar-backed tokens becoming one of the fastest-growing channels for U.S. currency abroad. By Visa’s count, stablecoins processed $100 trillion in transaction volume over the last twelve months. Major institutions — including JPMorgan, Citi, Visa, Mastercard, and BlackRock — are drawing deeper into blockchain infrastructure.

Crucially also, the dollar is winning. Millions of people worldwide who lack easy access to traditional dollar accounts can now hold and transact in digital dollars through software wallets. Clear rules have helped anchor this new system in U.S. standards and institutions rather than in competing alternatives.

GENIUS demonstrated that smart regulation expands markets rather than constricting them. But it addressed only part of the picture.

Blockchain networks still lack rules

Stablecoins account for less than 15% of the crypto market by market cap, yet they depend on the remaining 85% of foundational blockchain networks that still operate without a coherent federal framework. This is equivalent to regulating smartphones while ignoring cellular networks. 

The CLARITY Act rectifies this. 

At its core, CLARITY provides clear rules for blockchain networks and defines regulatory responsibilities across digital asset markets. It creates incentives for transparency, risk reduction, and competition under a common set of standards. It helps prevent the next FTX disaster by empowering regulators to supervise intermediaries with proven principles from traditional finance: proper custody, segregation of customer assets, and full disclosure. And it opens the next wave of institutional adoption. 

Major firms are already moving. BlackRock has launched tokenized funds. JPMorgan is building blockchain-based payment systems. And DTCC, which holds $114 trillion in assets under custody, is preparing to scale tokenized securities with the Canton Network. CLARITY removes the barriers to entry, giving traditional finance a clear pathway to participate compliantly. This benefits not just the crypto industry, but consumers, investors, and the long-term competitiveness of U.S. capital markets.

The choice before us

History shows that open, neutral platforms governed by clear rules create the greatest value. The internet succeeded because entrepreneurs knew the rules of the road. They could build, attract capital, and compete on merit rather than on regulatory guesswork. Blockchain networks deserve the same chance.

Regulation versus innovation is a false choice. GENIUS already disproved that: During the second half of 2025 — after GENIUS was signed into law — more than $13 billion flowed into crypto startup investment, nearly double the $6.9B invested pre-GENIUS in the first half of that year. Meanwhile, forecasts project the market for tokenized assets to grow 100x growth over the next few years. The choice is between certainty and uncertainty, America can assert its leadership and build the future here at home or watch it be built elsewhere.

No law is perfect, and the CLARITY Act is no different. The reality is that there are no consumer protections on crypto currently. CLARITY puts those in place. As with any bill, no one will get everything they want, but the version released today will move the industry forward. It reflects months of bipartisan negotiations and significant compromises from industry. Whatever its imperfections, the CLARITY Act is significantly better than continuing with no safeguards at all.

The decisions Congress makes this year will determine where the next era of financial infrastructure will flourish and who writes its rules. If the CLARITY Act passes, it will let the U.S. lead once again, just as it did with the commercial internet. Failing to act means innovation will migrate elsewhere, under frameworks devised by others.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Unlike traditional equities, tokenized stocks can be self-custodied in digital wallets, moved permissionlessly, traded at any time, and used seamlessly as collateral across onchain finance. In just the past couple months, Coinbase, DTCC, NYSE, Robinhood, and others have made moves here, ranging from processing trades onchain to forming new joint ventures to launching new chains.

The momentum shows up in the data. The market cap of tokenized stocks reached about $1.7 billion at the end of June, up from $329 million a year earlier — more than 5x growth. This makes tokenized stocks one of the fastest-growing categories of tokenized assets (which some refer to as real-world assets).

How much of that growth comes from new issuance versus price appreciation of the underlying stocks? Unlike stablecoins, whose circulating supply is a direct proxy for demand — one token, one dollar — tokenized stocks move with their underlying equities, so market cap does not cleanly separate the effects of new tokens minted and existing tokens repricing. 

The evidence points to issuance though. More than half of today’s market cap sits in assets that weren’t onchain a year ago. And most of the remaining balances arrived mid-year — after much of the period’s price movement in the underlying stocks had already occurred.

While the market for tokenized stocks is new, its composition has already changed greatly in the past year:

Crypto-linked products once dominated the market, but their share by market cap has dropped to 21% (as of June) from 79% a year earlier. They have lost the top spot to the “other” category — a long tail of hundreds of smaller listings — that now makes up 35% of the market, up from 15% a year earlier. 

The rest of the market is climbing too. Megacap tech — tech companies with a market cap around $100 billion or more — now make up 10.6% of the market by market cap as of June, up from 0.6% a year ago. ETFs and indices grew to 17.3% of the market in that same period, up from 4.5% a year earlier. 

The fastest riser has been, unsurprisingly, the AI and chips category. The category vaulted from less than $1 million in June 2025 — 0.3% of the market by market cap at the time — to 15.5% of the market as of a year later. 

Transfer activity is on the rise too. Monthly transfer volume for tokenized stocks reached $9.22 billion in June, up from $53 million last June — a more than 170x increase. This metric measures any onchain movements, such as trading, transfers between wallets, or collateral deposits into DeFi protocols. 

***

The institutional groundwork is now falling into place. In just the past month, DTCC processed its first live production trades of tokenized Treasuries and equities on Digital Asset’s Canton Network. (A full launch of the broader tokenization service is planned for October, giving Wall Street a direct on-ramp to the ~$114T in assets custodied at DTC.) At the start of the month, Robinhood launched its own chain on mainnet, bringing traditional markets, crypto, and real-world assets together on an open network. On June 22, the parent company of the NYSE announced a joint venture with OKX to make tokenized NYSE-listed stocks available to users, pending regulatory approval. And a week earlier, on June 16, Coinbase said it will offer 1:1-backed tokenized U.S. stocks to non-U.S. users, with dividends, full shareholder rights, and 24/7 trading. Binance shipped its own version days before. 

The market for tokenized stocks is still a small one compared to traditional equities, which trade in the double-digit trillions per month. But the trend is clear: More issuers and platforms are bringing tokenized stocks online, and the category is growing.

***

Robert Hackett is features editor and head of special projects at a16z crypto.

Ryan Holloway is a data consultant.

***

This article was translated into Korean by Token Post, available here.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

The crypto market structure legislation currently moving through Congress, the Clarity Act, is unlike anything we’ve seen since the Telecommunications Act of 1996: a large-scale, forward-looking effort to embrace technological change and introduce consumer safeguards for a nascent technology. Like that earlier landmark law, this effort enjoys broad bipartisan support and requires coordination across multiple committees. And that only happens when Congress recognizes the high stakes.

For too long, American financial policy has been in a reactive crouch. Since the 2008 financial crisis, virtually every major piece of legislation – from Dodd-Frank to subsequent banking reforms – has been backward-looking, designed to address past crises rather than prepare for the next frontier. While these measures serve important purposes, they’ve left us with a regulatory framework built for yesterday’s risks, not tomorrow’s opportunities.

Crypto market structure legislation can break this pattern. For the first time since the Gramm-Leach-Bliley Act of nearly three decades ago, we have the chance to enact comprehensive financial policy measures that aren’t precipitated by crisis.

Some argue that nearly centuries-old securities laws already cover what crypto needs, or even claim that the industry would prefer to have no regulation at all. As someone who has spent his career cutting red tape and championing the free market, I can understand where those arguments come from. In this case, though, they are simply wrong. 

Those who claim the legal status quo for crypto is just fine have not listened to what businesses large and small are actually saying. They are asking for rules to operate. When entrepreneurs know boundaries, they build confidently within them. The internet and the thousands of startups that made America the world’s tech leader flourished not despite regulation, but because smart regulation created the infrastructure and protections users and investors required.

Over time, decentralized innovations can also prove their worth within a regulatory framework. But first, we need the framework itself. 

The framework in the Clarity Act can ensure entrepreneurs feel confident in creating jobs and building businesses without fear of arbitrary or unexplained crackdowns. Upon becoming law, it will establish protections for consumers and investors, and give law enforcement agencies the tools they need to identify and stop criminals and bad actors seeking to hurt the American people. I’m proud of many bills passed during my tenure in Congress, including as chair of the House Committee on Financial Services, that made important strides forward, but this effort dwarfs them in scope and significance.

This legislation is governance at its best: Congress identifying an emerging sector with enormous potential, and responding with the regulatory clarity needed for that sector to flourish within appropriate guardrails. While there is disagreement on precise details, almost everyone – from entrepreneurs and investors to academics and policymakers – agrees that this regulatory clarity is crucial for blockchain technologies and the innovation and economic growth they enable to thrive.

What gives me confidence that we are closer than ever to a breakthrough is the remarkably broad support this has garnered in Congress. The GENIUS Act, which regulates stablecoins, passed both chambers of Congress with strong bipartisan backing. And various crypto market structure bills have attracted sponsors across parties and ideological spectrums. Members understand digital assets aren’t disappearing and recognize it’s time to act.

Other countries are moving aggressively, and global capital and innovation will flow to markets with clear rules. The Clarity Act is our best chance to maintain America’s role as home for global capital markets deployed in the world’s safest, most resilient regulatory environment, with rule of law and property rights clearly defined. This is about digital assets, but it’s also about deciding if America will lead the 21st-century economy and set the pace for the world or only react to crises.

I hope that decision is obvious.

***

This piece originally appeared in Fortune.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

It’s a comforting story. It’s also mostly wrong.

Here’s the more honest version: where TradFi can use a blockchain to make its existing business better, it will. Not because it has embraced decentralization, but because it’s a compelling COGS story — the technology happens to cut costs, improve settlement, expand distribution, and tighten its grip on customer relationships.

What this means is that institutions aren’t somehow merging with DeFi. Instead, they’re selectively using the parts of DeFi that fit within their operating constraints and discarding the parts that do not; they’re reconfiguring DeFi around institutional requirements. The result is unlikely to look like either traditional finance or today’s DeFi. Instead, we’re beginning to see the emergence of a new category built on blockchain rails but optimized for institutional constraints: programmable financial infrastructure.

That dynamic may evolve as regulatory frameworks mature. Legislation such as the CLARITY Act could eventually make it easier for institutions to engage directly with permissionless systems. But regardless of what becomes legally possible, TradFi’s risk posture won’t reset overnight. Institutions still adopt technology through the lens of cost, risk, control, and operational fit — which is why this presents the industry with two opportunities, not one.

The first is helping institutions adopt the infrastructure they are ready for today. Every primitive an institution adopts – from atomic settlement to programmable money to tokenized collateral – validates the technology, builds out shared rails, and pulls real volume and capital onchain.

The second is continuing to build the open, crypto-native financial system that institutions are not yet prepared to use.

These are not competing bets. They can and should exist in parallel, and done well, each reinforces the other. Open networks and ecosystems keep producing the primitives, markets, and innovations that institutions eventually adopt. If both succeed, convergence happens naturally – not because one system fully replaces the other, but because both increasingly come to rely on the same underlying infrastructure.

What TradFi is actually doing

TradFi adopts a primitive when it does two things at once: improves cost, risk, or distribution and stays compatible with control and accountability. The primitives that institutions discard — open access, pseudonymity, immutable execution — pass the first test but fail the second. That’s why the adoption pattern is predictable rather than arbitrary, and why builders can use it as a design test. That is, if a feature delivers value only by removing institutional control, it will almost certainly be reshaped or rejected, however elegant it may be.

Let’s put some primitives through the test. Atomic settlement collapses the gap between trade and finality, erasing counterparty risk and freeing the collateral institutions park against unsettled trades. A shared ledger turns the largest hidden cost to the back office, reconciliation, into a non-event. Programmable money lets coupon payments, margin calls, and corporate actions run as code instead of a chain of manual instructions. AMM curve math, stripped of its permissionless wrapper, reappears as the pricing engine for onchain FX and tokenized money-market NAVs.

Each improves a number on a P&L or removes a line of operational risk and its associated cost, but none requires an institution to believe in decentralization. So let’s be precise about what’s happening with initiatives like JPM’s permissioned blockchain for institutional deposits, or the tokenized money market funds at BlackRock, and Franklin Templeton: These aren’t enterprises dipping a toe into DeFi. They’re using blockchain to do things they already do — settling interbank payments, managing fund subscriptions, and distributing yield-bearing instruments — but with better plumbing. These are deployments that use blockchains’ technical properties (programmability, transparency, atomic settlement) and deliberately discard the properties that make native DeFi work (open access, pseudonymity, and trustless execution).

That’s not a failure or a compromise. It’s a deliberate architectural choice, and one that tells us a great deal about where this is heading.

Different buyers, different rules

It would be a mistake to assume that institutional adoption is simply a larger distribution channel for existing DeFi infrastructure. Institutions are not evaluating protocols the way crypto-native users do. When institutions consider software vendors, infrastructure partners, operational risk, compliance controls, and long-term ownership of critical systems, they follow their standard operating procedure. The result is that success in DeFi does not automatically translate into success with institutions.

Enterprises rarely buy the “best” technology. They buy the technology that best fits existing workflows, risk models, and  procurement processes, among other things. 

Any technology that enters a heavily regulated, risk-managed, liability-averse institutional environment gets shaped by that environment. It happened with the internet (enterprise firewalls, private intranets). It happened with cloud computing (private cloud, VPCs, FedRAMP). It’s happening with AI (internal deployments, data residency requirements, model governance). Blockchain is no different.

The reconfiguration happens along two axes: 

1. Compliance: KYC, AML, sanctions screening, investor accreditation, and regulatory reporting requirements aren’t negotiable for most institutions. Permissionless systems don’t accommodate these requirements natively. Institutions need the ability to freeze assets, reverse transactions, and identify counterparties. DeFi was not originally designed around those requirements, and accommodating them often requires meaningful architectural changes. This may evolve. For example, CLARITY could make it easier for institutions to access permissionless systems while meeting regulatory requirements. But today, most institutions must evaluate blockchain infrastructure through the lens of control, accountability, and operational risk.
2. Enterprise value delivery. This axis is often underappreciated. Institutions aren’t adopting blockchain because they believe in permissionlessness as a principle. They’re adopting it because it can compress costs, reduce reconciliation friction, create new distribution channels, or embed them more deeply in a customer relationship. The value proposition has to be expressed in those terms, or it doesn’t survive procurement.

Stablecoins may be the clearest example. Banks, payment providers, and fintechs increasingly view them as useful settlement infrastructure because they enable faster movement of dollars across networks and geographies. Yet few are embracing the broader philosophy of permissionless finance. They’re adopting programmable dollars because they’re useful, not because they’re trying to recreate the financial system around DeFi’s principles.

Circle’s evolution is an apt illustration. Arc reflects how blockchain infrastructure is increasingly being packaged for institutional buyers: emphasizing compliance, operational controls, trusted counterparties, and integration into existing workflows rather than permissionless access and composability. The value proposition isn’t permissionlessness for its own sake. It’s faster settlement, global reach, and improved capital efficiency delivered in a form institutions can actually adopt.

Even organizations like SWIFT increasingly frame blockchain through this lens. Their efforts around tokenized asset interoperability are not attempts to replace existing financial institutions. They’re attempts to improve how existing institutions coordinate with one another using the SWIFT network. The pattern appears repeatedly: Blockchain adoption that strengthens established financial networks rather than displacing them.

This is simply how powerful technologies evolve when they encounter large, established markets.

Two opportunities for builders

At the industry level, it would be a mistake for everyone to abandon one opportunity for the other. At the company level, it would be a mistake to try to pursue both at once.

Institutional adoption and open networks can reinforce one another at the ecosystem level. But for most teams, they remain fundamentally different businesses. Building for institutions requires understanding procurement, compliance, controls, channel partners, and long sales cycles. Building for open networks requires optimizing for developers, liquidity, composability, and network effects. The customer, distribution model, product requirements, and success metrics are often entirely different. 

This doesn’t mean one opportunity is better than the other. It simply means founders should be clear-eyed about which market they’re serving, and be aware that what unites them are the rails underneath: public blockchains as neutral settlement. 

Partnering with institutions and building an adjacent financial system aren’t in tension. When done right, each makes the other more valuable. The permissioned layer brings volume, legitimacy, and capital; the open layer keeps producing primitives the permissioned layer adopts next. Convergence, when it comes, happens at the rails — not by one system surrendering to the other.

Public blockchains may become increasingly important settlement rails, even as the applications built on top of them become progressively more permissioned.

Building for programmable financial infrastructure

There are two approaches to consider when it comes to building for this new programmable financial infrastructure: building something from scratch or adapting existing products. 

Consider networks like Canton. Rather than adapting existing DeFi infrastructure, they are designed specifically around institutional requirements for privacy, compliance, and controlled interoperability. The goal isn’t to bring banks into DeFi. It’s to use blockchain-based coordination while preserving the governance, confidentiality, and operational controls that institutions require.

Not every successful institutional strategy requires rebuilding from scratch. Morpho, for example, is taking the opposite approach. Rather than abandoning its DeFi primitives, Morpho has focused on making them easier for institutions and asset issuers to consume. Apollo’s ACRED fund, for example, uses Morpho as part of its onchain lending strategy, pairing a DeFi- native lending primitive with institutional-grade distribution, compliance, and fund structures. The result is neither pure DeFi nor a fully isolated institutional stack. It is a model where institutions selectively adopt existing crypto infrastructure while packaging it in a way that aligns with their own requirements for control, compliance, and distribution. 

This new category is purpose-built for institutional constraints. It draws from DeFi, but operates in a more permissioned, compliant manner and is therefore necessarily different from what exists today. 

Some teams, like Morpho, have successfully adapted crypto-native infrastructure for institutional use cases. But builders shouldn’t mistake this for the default playbook. Institutions are a distinct customer segment with distinct requirements. In many cases, designing for those requirements from the outset will prove more effective than adapting products originally built for open networks.

The opportunity to keep building in DeFi

The innovations institutions are adopting today did not originate inside banks, asset managers, or existing financial infrastructure. They emerged from open networks where builders were free to experiment with new market structures, coordination mechanisms, and financial primitives. 

That distinction matters. Institutions are not the industry’s primary source of innovation: The permissioned layer is often downstream from the open one.  

This brings us to the more important strategic point: If our industry becomes too focused on selling to banks and asset managers, we risk mistaking a large buyer category for the entire opportunity. TradFi is an important customer. But it’s not the only one. 

Designing for institutional requirements is a legitimate and valuable pursuit, but it is only one lane, not the whole road. The companies that endure will be the ones that remain clear-eyed about who they are building for. Institutional adoption may be a large opportunity, but it is not simply an extension of DeFi. Success in one market does not guarantee success in the other.  

***

If you’re building for institutions, embrace it fully. Don’t assume crypto-native traction automatically translates into enterprise adoption. Learn the customer, understand the buying process, and build intentionally around institutional requirements.

If you’re building for open networks, keep doing so. Don’t abandon your vision simply because institutions are the loudest buyers in the market today. 

Remember: These are complementary, not competitive. One adapts, commercializes and scales proven innovations. The other discovers them. A version of this technology will almost certainly become part of the financial plumbing of the existing TradFi system. But that is not the only future being built. Open networks remain the industry’s most important source of experimentation and innovation, and many of the primitives that shape tomorrow’s institutional infrastructure will likely emerge there first.

TradFi isn’t adopting DeFi. It’s selectively adopting parts that fit its model. The opportunity for builders is not to chase every market at once, it’s to understand which one they’re building for. And to execute accordingly. The future may indeed run on institutional infrastructure, but many of its most important innovations will continue to emerge from open networks.

***

This article was translated into Korean by Token Post, available here.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Long before crypto made coordination programmable, Nobel Prize winner Alvin Roth was designing markets where coordination could save lives.

In this episode of First Principles, Roth tells the story of how he helped build systems for some of the hardest matching problems in the world, from where doctors train and where students go to school to how kidney donors can reach the patients who need them.

He joins Tim Roughgarden, Head of Research at a16z crypto, and Scott Kominers — Harvard Business School professor, a16z crypto research partner, and one of Roth’s former students — for a conversation about how market design moves from theory into the real world.

They explore how economic theory becomes practical engineering, whether that’s matching riders to Ubers, doctors to medical residencies, students to New York City high schools, or organ donors to people whose lives depend on it. They also cover how these same problems show up in today’s crypto networks. Roth explains why markets are not just natural forces, but engineered systems; why the details of timing, congestion, incentives, and trust can make or break a marketplace; and why some of the most important markets are the ones where simply exchanging money can’t do the work.

This is a conversation about economics at its most practical and profound: how to design systems that coordinate people, solve real problems, and sometimes save lives.

Highlights

00:00 Intro: Why market design matters
04:18 The economist as engineer
08:09 When theory meets the real world
07:02 Fixing the medical residency match
15:32 Why markets unravel
18:22 Redesigning NYC high school admissions
28:05 The hidden problem of congestion
34:47 How kidney exchange saves lives
45:26 How the internet changed market design
48:25 Airbnb, Uber and smarter marketplaces
51:28 Repugnant transactions and moral economics
53:32 When markets need social support
54:32 The unexpected effects of criminalizing surrogacy
01:04:58 Preference signals and the job market
01:18:53 A broken market: resettling refugees and other migrants

About First Principles

First Principles is a special limited series from a16z crypto about the scientific roots of modern computing — especially blockchains — told through rare conversations with the pioneers who helped shape the foundational ideas behind distributed systems, consensus protocols, economics, mechanism design, cryptography, zero knowledge, and more.

People often tell the story of the Bitcoin whitepaper as if it appeared out of nowhere. But the ideas behind Bitcoin — and blockchains more broadly — come from decades of computer science, economics, mathematics, and cryptography. First Principles is a guide to that lineage, as told by the people who helped build it.

Subscribe to follow along

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Here are a few of the books the a16z crypto team is reading and recommending this summer, from mountaineering memoirs and market histories to sci-fi, children’s books, and epistolary novels. Our picks also cover the seismic shifts in energy, manufacturing, biology, and technology that forced reinvention throughout history.

See more of our recommendations in reading lists from seasons past, including: summers 2025, 2024, 2023, 2022; and winters 2025, 2024, 2023, 2022.

Dan Boneh, research

Apple in China by Patrick McGee

How did China become so good at tech manufacturing? This book tells the story of how Apple spent over a decade and many resources to train firms in China to build Apple products, with stringent requirements and high precision. Along the way it tells the history of many Apple products and how they came to be. It is an interesting story of how we ended up where we are.

Conner Lyons Brown, tech ops

We Few: U.S. Special Forces in Vietnam by Nick Brokhausen

Brokhausen’s memoir of MACV-SOG, the unit that ran the most dangerous missions of the Vietnam War behind enemy lines in Laos and Cambodia, is the rawest portrait of high-performing operators I’ve read. What stays with you is the psychology of doing something unprecedented and lethal, the macabre humor and resilience it takes to keep walking into the jungle knowing each step might be the last. For anyone who works in adversarial, high-stakes environments, it’s a bracing look at what real edge cases demand.

Leisure: The Basis of Culture by Josef Pieper

Pieper wrote this in 1948, but it reads like it was written for the age of AI. He diagnoses the two roads most traveled by any professional: the total world of work, where a person’s worth collapses into output, and a hollow pursuit of pleasure with no center. Pieper refuses both. Leisure, he argues, is not idleness or time off but a contemplative, receptive stance toward reality, the stillness to actually perceive the world rather than just operate on it. As AI absorbs more of the doing, the open question becomes what the doing was ever for, and Pieper’s answer, that the highest human activities are the ones done for their own sake rather than for compensation, lands harder now than when he wrote it.

Drew Coffman, editorial

Antkind by Charlie Kaufman

Narcissistic, recursive, and deeply in love with its own absurdity (complimentary). The whole book is a 700(!!)-page bit that never breaks character. Big Charlie Kaufman energy.

Apple: The First 50 Years by David Pogue

An exhaustive march through fifty years of Apple’s history. Pogue has been chronicling Apple for so long (starting at Macworld in the 80s) that he’s one of maybe three people alive who’d get this kind of access, which is what makes it great. Dropped weeks before Tim Cook announced his exit, this text works as a bookend, catching Apple right as one chapter closes and another cracks open.

Augustus by John Williams

Everyone always talks about Stoner, but not enough people talk about John Williams’s other incredible work: Augustus. Told in the form of an epistolary, this portrait of Augustus as an idealist and pragmatist feels deeply relevant.

Robert Hackett, editorial

Fahrenheit-182: A Memoir by Mark Hoppus

Punk rejects mass culture. Pop is mass culture. So how does a pop-punk band like Blink happen?

In his memoir, bassist and only constant member Mark Hoppus recounts playing a show in Bologna where the crowd hurled rocks and bottles at them. A year earlier, Blink had released a viral video mocking boy bands, and had become so successful that some audiences — a contingent of hostile Italian metalheads, in this case — could not tell the parody from the thing being parodied. (Or didn’t care to.) As Hoppus puts it: “They didn’t want some bullshit American pop act fucking up their heavy show.”

I love this anecdote — not just because it’s darkly comic (guitarist Tom DeLonge reportedly swore off pasta bolognese afterward) and boy bands are dumb, but because it illuminates one of modernity’s supreme ironies: What happens when countercultures go mainstream. A similar arc can be traced across jazz, abstract art, hip-hop, skateboarding, streetwear, open source software, and, relevantly for this audience, crypto.

Fahrenheit-182 is also a founder story. Hoppus and DeLonge are a volatile creative duo — like Lennon and McCartney with fart jokes. The same forces that allowed them to pump out hits eventually tore them apart. Fortunately, unlike most founder breakups, this one had a reunion tour. I caught them at Madison Square Garden a few years ago and it was a glorious, stupid, perfect blast of nostalgia.

So is the book.

Andy Hall, research

The Education of Henry Adams by Henry Adams

An extraordinary reflection of living through a period of intense technological change. Adams reflects on how the educational system failed to anticipate the radical transformation he experienced, and describes where he learned life’s true lessons. A classic and timeless book that’s perfect for today’s moment.

A World Beneath the Sands: Adventurers and Archaeologists in the Golden Age of Egyptology by Toby Wilkinson

A sweeping exploration of the history of Egyptology and how humanity deciphered hieroglyphics, developed the science of archaeology, and competed for prestige through scientific discovery. You’ll find interesting and strange connections to today’s tech race.

Liz Harkavy, deal

The Will of the Many by James Islington

I loved how The Will of the Many blends an elite academy setting, Roman-inspired politics, and a genuinely propulsive mystery. The book is smart and immersive without feeling slow and it has novel world-building.

Maggie Hsu, go-to-market

Nyla the NFT by Maggie Hsu and Nyla Hayes

Most of the crypto books for kids are basic glossaries. I wanted to write something educational that also had an engaging storyline. Nyla the NFT teaches basic crypto concepts inspired by the story of The Wizard of Oz. A group of NFTs venture down the “yellow blockchain road” and learn about self-acceptance along the way. My co-author Nyla Hayes is a talented multidisciplinary teen artist who also did all the illustrations.

Stablecoins for Babies by Edward Woodford

Stablecoins for Babies is exactly what it sounds like — a board book about stablecoins. It makes for the perfect baby shower gift (along with some USDC).

Scott Duke Kominers, research

Moral Economics: From Prostitution to Organ Sales, What Controversial Transactions Reveal About How Markets Work by Alvin E. Roth

My doctoral advisor Al Roth has been thinking for decades about “repugnance” – why some people prefer that some markets should not exist. In Moral Economics, he tackles the motivation for these prohibitions, and the trade-offs they force, head-on. And he explores in particular how such (non-)market norms emerge and sometimes later collapse (limitations on alcohol, drugs, and – in a completely different category – same-sex marriage have all been relaxed in recent years), and what this means for making markets in the future.

The Proof in the Code: How a Truth Machine Is Transforming Math and AI by Kevin Hartnett

I’ve recently become fascinated by the potential of formal verification to refine our understanding of math and economic theory.

The basic design paradigm is simple: proofs written in a machine-checkable language, where “compiling” means that every logical step is completely sound. Lean – first launched by Leonardo de Moura in 2013 and now chronicled in Hartnett’s book – is the leading such “proof assistant” these days. It spawned a massive project to formalize mathematics from the ground up, and has crossed over to cryptography and software security (and some of my collaborators and I have been bringing it to economics).

This tech has taken on particular importance in recent years, as it is a perfect complement to generative AI. Lean at once turns proof logic into a machine language that AI models can reason in more easily, and provides a trusted system to verify (or reject) conclusions when AI models try to “prove” in ways that are difficult for humans to understand.

The Man Who Saw Seconds by Alexander Boldizar

A fascinating speculative-fiction framework: what happens if one can see just a tiny bit into the future, and pick the preferred passage through time? While you might hope that you could just spend those precious seconds winning at casinos, such a metaphysical gift – at least seen through a certain lens – might put pressure on the very foundations of civil society.

Scaling People: Tactics for Management and Company Building by Claire Hughes Johnson

My friend and co-author Jesse Shapiro recently recommended this book to me and it’s fantastic: easily the best book I’ve ever read on both (1) managing people and (2) managing people’s interactions with you. Hughes Johnson, who served as COO at Stripe through their period of massive growth, gives a vivid and very, very practicable guide to understanding individuals’ and teams’ motivations and operating modes. The book comes with detailed principles and strategies for everything from hiring and onboarding to day-to-day management, communication, evaluation, and feedback. But the biggest payoff is the way the book trains awareness and recognition of people as they are, and how to build on that to help them maximize themselves.

Introduction to Quantitative Economics by Jesse M. Shapiro

And speaking of Jesse Shapiro, he has a new book out this spring that teaches his special brand of theory-grounded reasoning about data in economics. Shapiro is one of the most brilliant pedagogues in the field, and the book translates his crystal-clear thinking into a series of illustrations and exercises that take you from classic first-principles price theory to the empirical frontier in just 64 pages. It’s perfect for a self-study course, and appropriate for both newcomers and experts alike.

The Art and Craft of Problem Solving by Paul Zeitz

While I’ve been “Lean”-ing into formal mathematics this spring, I’ve also been revisiting classic works to fine-tune my mathematical thinking and problem-solving. And there’s no better book for that than Paul Zeitz’s magnum opus, now in its third(!) edition.

The Murder of Roger Ackroyd by Agatha Christie

This year marks the 100th anniversary of one of the most brilliant and category-defining mystery novels of all time.

Christie’s detective Poirot has retired to the countryside (to grow vegetable marrows – I’ve always liked this detail). But he quickly gets wrapped up in investigating the stabbing of the eponymous Ackroyd, as narrated by village physician Dr. Sheppard – who plays a sort of “Watson” throughout. To hear Sheppard tell it, with all the murderer’s misdirection and so many false leads, this might almost have been one of Poirot’s greatest failures. But in fact, it became perhaps Christie’s greatest success.

She Walks at Night by Seishi Yokomizo & The Clock House Murders by Yukito Ayatsuji

While you’re in the mystery mode, you absolutely must read this summer’s newest honkaku translations from Pushkin Vertigo.

She Walks at Night is the perfect follow-up to Ackroyd: a tale of family dysfunction, jealous relations, and not one but two hunchbacks (or possibly three?), who find themselves tied up in a crisscrossing web of their own weaving. The story as told is like a cross between a stage play and a waking dream – you’re never quite sure who’s who, or who’s playing which role. Despite literally every clue being handed to the reader, there’s a sense of almost sleepwalking into the final resolution.

The Clock House Murders, meanwhile, unfolds in parallel both inside and outside the mysterious Clock House, where 108 timepieces of all shapes and sizes pace the action day and night. Timing distortion is a classic mystery trope – the stopped clock, the wristwatch set back an hour, and the like – and here you naturally expect that similar games are at play. Yet the author demolishes each such idea almost as soon as it occurs, leaving the reader to wonder whether all the clocks are somehow a meta-ruse just to redirect attention away from what’s really going on. I couldn’t put this one down – time seemed to speed up as the story raced towards its conclusion, and I found myself turning the pages faster and faster.

Kim Milosevich, marketing

Hoang Nguyen, marketing

Think You’ll Be Happy: Moving Through Grief with Grit, Grace, and Gratitude by Nicole Avant

A book about grief, purpose, and the legacy a mother leaves behind — woven with stories from music, culture, and politics.

SungMo Park, go-to-market

The Super Gap (초격차, or Cho-gyeok-cha) by Oh-Hyun Kwon

This book is written by the former chairman of Samsung Electronics, who led the company to become the number one semiconductor company globally. “The Super Gap” means a gap so wide that competitors cannot even think about challenging you. Based on his 33 years at Samsung Electronics, he distills his leadership experience into four pillars: leader, organization, strategy, and talent. The author explains that this “Super Gap” applies not only to technology but also to systems, culture, people, and organization. It’s especially compelling because these insights come from someone who experienced both success and failure firsthand, and who ultimately walked away at the peak to practice the “sustainability” he preaches.

Robbie Petersen, deal

I See Satan Fall Like Lightning by René Girard

One of the most thought-provoking frameworks that reveals why we desire the things we desire and what this implies for how society is organized. If you love philosophy, anthropology, history and theology – great read.

Jason Rosenthal, operations

Aiden Slavin, policy

Energy by Richard Rhodes

Rhodes is the author of one of my favorite series of all time, the four-volume nuclear history that begins with the Pulitzer-winning The Making of the Atomic Bomb, plus several novels that are very good, if less heralded. Energy is his five-century study of how humanity moved from wood to coal to oil to electricity and on toward nuclear and renewables, told the way Rhodes tells everything: through people, and through false starts, bad bets, and inventors who accidentally solved the right problem. Reading his work, it’s hard not to notice how often progress came down to scaling capacity as much as fundamental breakthroughs, exactly the question the AI buildout is now forcing, as compute demand revives interest in nuclear and reshapes grids. The constraint on this next computing era may be joules rather than algorithms, a challenge we’ve navigated before.

Kira Song, finance

Tim Sullivan, editorial

An Engine, Not a Camera: How Financial Models Shape Markets by Donald MacKenzie

Sociologist Donald MacKenzie’s An Engine, Not a Camera is nominally a book about financial economics, but its real subject is more provocative: What happens when theories stop describing the world and start changing it? MacKenzie traces how academic models of markets escaped the university and became embedded in the markets themselves. First published in 2006, it remains relevant today in a number of contexts — not least because of the recent interest from TradFi in crypto. But I wanted to read it again because of the book I’m working on with colleague Robert Hackett, about how computer science theories acted on the world.

Thrown by Kerry Howley

In this work of literary non-fiction (maybe?), the narrator, Kerry Howley (who may be the same person as the author, but may not be) follows a pair of mixed martial artists around Iowa (where author Howley had done her MFA). Both Howleys explore obsession, uncertainty, and the search for meaning with unusual intelligence and empathy — and with a prose style that both surprises and delights. It’s nominally a book about fighting, but it’s really a book about commitment (and maybe a book about a book about fighting).

Ertem Nusret Tas, research

Sátántangó by László Krasznahorkai

This is one of the unique books I liked mainly for its narrative style, characterized by lengthy sentences and complex paragraphs. Very Kafkaesque undertones as well.

Benjamin Wu, talent

A Man for All Markets: From Las Vegas to Wall Street, How I Beat the Dealer and the Market by Edward O. Thorp

Growing up, I was obsessed with the movie 21 and the idea that you could actually use math to beat the house. This book is the autobiography of the man who essentially started it all. Edward Thorp went on to apply those same principles to Wall Street and became the father of quantitative investing. It’s an incredible story about using pure logic to disrupt rigged, legacy systems, which completely captures the exact same builder mindset we see in crypto today.

Michael Zhu, engineering

The Carrier Bag Theory of Fiction by Ursula K. Le Guin

The Carrier Bag Theory (coined by Elizabeth Fisher) centers the container, or bag, as the original human technology — a tool for gatherers to carry food and other items — in contrast to the prevailing narrative that centers the weapon (rock used as bludgeon). In Le Guin’s essay, she applies this lens to narrative fiction, framing the novel as a container of characters and stories. Shape rotator that I am, I instead found parallels in math and software: The foundational object underlying all mathematics is the set, and what is a set if not a container, a bag of elements? And in software — beyond the obvious syntactic parallels (see, e.g., Docker), it is the nesting-doll containment (also known as “abstraction”) of complexity that makes this whole enterprise possible. There are bags everywhere for those with the eyes to see.

Stephanie Zinn, editorial

Catherine the Great & Potemkin: The Imperial Love Affair by Simon Sebag Montefiore

Just your standard romance but with more urban planning and foreign policy memos. Catherine the Great & Potemkin is part imperial history, part power-couple biography, and part extremely high-stakes situationship. The story of this Antony-and-Cleopatra-level romantic and political partnership is told with all the drama and detail you’d hope.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Every launch is momentous. Every setback is the end of the industry. Every rival is a fraud, every partnership is historic, and every week is the most important week.

The volume knob has been snapped off at eleven. 

The problem is that attention is a fixed budget, and too many players in the industry insist on spending it all at once.

When every message is maxed out, your audience loses the ability to rank them. The press can’t tell what’s important and what’s just a routine Tuesday update. Partners can’t separate a real crisis from reflexive outrage. 

If everything is loud, the audience becomes deaf. Yell about everything, and the people who need to hear you will discount you by default. The day you have something to say, no one will care. You will have already set your baseline to “ignore.”

Contrast is a more effective way to get — and hold — attention. The loudest take may win the short game — X engagement, virality, a good day on the timeline — but winning the long game means winning over people — that is, specific people. They may be regulators, reporters, institutional partners, long-term customers. People who can help you achieve outcomes. They’re hungry for more than just the hype and “vibes.” And they have long memories for who cried wolf.

Don’t be silent. Be selective. Decide, intentionally, which moments are worth the fullest expression of your voice:

• Ration the superlatives. Save “historic” and “first” for things that actually qualify. Spend them like they’re scarce. To your audience they are.
• Let proof do the work adjectives can’t. In the show-me era, numbers beat press releases. Traction speaks louder than words.
• Pick your outrage. Not every bad headline deserves a response. Reacting to all of them will get you nowhere. Crybabies don’t win fights.
• Build a quiet baseline. Your loud moments need somewhere to stand out from. Silence gives signal contrast.

Being loud doesn’t mean being heard. In a market where everyone is shouting, the most powerful positions belong to those still capable of discerning when it’s time to speak up. Raise your voice rarely, and people will hear you when you do.

***

Paul Cafiero is a communications partner on the a16z crypto marketing team.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Before blockchains could reach consensus, Leslie Lamport had to define what agreement even meant when computers fail, lie, or disappear. In this episode of First Principles: The Scientific Roots of Blockchain Technology, Turing Award–winning computer scientist Leslie Lamport joins Tim Roughgarden, Head of Research at a16z crypto and Professor of Computer Science at Columbia University, and a16z crypto Research Partner Ittai Abraham to trace the ideas that helped define modern distributed computing.

Lamport’s work formalized some of the field’s deepest questions: how to reason about concurrent systems, how distributed systems can agree despite failures, and how to prove that protocols do what they are supposed to do. His work on logical clocks, state machine replication, the Byzantine Generals problem, and Paxos has shaped everything from cloud infrastructure to the consensus protocols underlying modern blockchains.

The conversation begins with Lamport’s early work on concurrency and the origins of the Byzantine Generals Problem, and then turns to fault tolerance: what happens when machines crash, behave unpredictably, or even act maliciously? We also cover the feedback loop between theory and practice, the long arc of fundamental research, and how blockchains are inheriting and extending decades of distributed systems work.

Highlights

• 00:00 – Intro: The problem every blockchain is built to solve
• 02:52 – Why concurrent systems are surprisingly tricky
• 04:40 – The origins of the bakery algorithm
• 07:37 – What does it mean for a protocol to be “correct”?
• 12:03 – The origins of the Byzantine Generals problem — and what happens when some computers fail
• 17:49 – How Paxos emerged from an attempted impossibility proof
• 23:47 – Why theory and practice need each other
• 33:48 – Government funding, DARPA, and the long arc of foundational research

About First Principles

First Principles is a special limited series from a16z crypto about the scientific roots of modern computing — especially blockchains — told through rare conversations with the pioneers who helped shape the foundational ideas behind distributed systems, consensus protocols, economics, mechanism design, cryptography, zero knowledge, and more.

People often tell the story of the Bitcoin whitepaper as if it appeared out of nowhere. But the ideas behind Bitcoin — and blockchains more broadly — come from decades of computer science, economics, mathematics, and cryptography. First Principles is a guide to that lineage, as told by the people who helped build it.

Subscribe to follow along.

Hear more from

• Tim Roughgarden: @Tim_Roughgarden
• Ittai Abraham: @ittaia

Follow a16z crypto

• X: @a16zcrypto
• LinkedIn: a16z crypto
• YouTube: @a16zcrypto
• Substack: a16zcrypto.substack.com

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

In the 2017 and 2021 cycles, the country became one of crypto’s most important retail markets. Korean exchanges could move meaningful volume; local investors had an appetite for new tokens; and listings on its largest exchanges, Upbit and Bithumb, became major milestones for projects seeking liquidity, holders, and market awareness.

That retail market is quieter now, but a more structural shift is underway. Some of Korea’s largest financial institutions and internet platforms are using this window to build on the country’s retail foundation — a push for institutional infrastructure driven by two trends, in particular: stablecoins and real-world asset (RWA) tokenization.

This guide is written for companies outside Korea who want a high-level map of that landscape, for partnerships or other collaborations. Many of these pilots and partnerships are already known locally, and many are still developing in real time. My goal is to highlight a few emerging patterns as a starting point for teams not based in Korea: which institutions are moving, what they are building, and where crypto-native projects may be able to help them.

Stablecoins are moving into the mainstream

Stablecoins are now a mainstream policy and business priority in Korea. The question currently being hashed out among policymakers, banks, and other industry players is whether Korea will create a regulated market for won-backed stablecoins.

At the center of the debate is who will be allowed to issue them: Banks want a bank-led model. Fintech and internet platforms want room to compete. And regulators are weighing the benefits of innovation against the tradeoffs of monetary control.

The urgency here comes from dollar-based stablecoin adoption. South Korea is watching an estimated $115 billion flow in capital move offshore to overseas exchanges, much of it into dollar-denominated stablecoins like USDC — meaning more payment, trading, and remittance activity happens outside domestic financial rails.

So a regulated KRW stablecoin is more than a product opportunity. For Korean banks and policymakers, it’s seen as a defensive move: a way to keep future digital money activity tied to the won, local institutions, and Korean regulation.

For now, much of the activity is still in pilot stages. Institutions aren’t sitting still, but they are working under real constraints. They can test, partner, and prepare — but they cannot fully commit until the legal framework is clearer. So what are they doing now?

Banking

KB Financial Group

As one of Korea’s most established institutions, KB Financial Group has delivered a credible proof point that gives other banks the confidence and blueprint to move forward. When an industry leader successfully demonstrates a case study like this, it removes much of the uncertainty that typically makes large enterprises hesitate, and that changes the calculus for the entire industry.

Earlier this year, the group’s banking arm tested real payments and remittance flows, including QR-code payments at major Korean coffee chain Hollys Coffee and cross-border remittances to Vietnam completed in under three minutes, reportedly at 87% lower cost than SWIFT. The significance isn’t the scale of this pilot itself; it’s that one of the country’s largest banks has actively completed end-to-end blockchain-based financial services.

Hana Financial Group

Among major financial groups, Hana Financial Group has moved quickly — and its recent investment in Dunamu, which operates Korea’s largest crypto exchange, helps explain why. Hana is now Dunamu’s fourth-largest shareholder, giving the financial group a more material strategic link to Korea’s crypto ecosystem than other traditional institutions.

This connection is showing up in the work. Hana is part of a multi-institution stablecoin consortium running remittance tests on GIWA, Dunamu’s own blockchain network, while Hana Card is already piloting live USDC payments for foreign visitors.

NH Bank

Other banks are moving more selectively, often based on their existing advantages. For instance, NH Bank, one of Korea’s major commercial banks, is working with NHN KCP, a large payment gateway, on stablecoin-based merchant settlement — a practical use case for retailers and payment processors.

KBank

KBank, Korea’s first internet-only bank, has a different edge as Upbit’s exclusive banking partner, which gives it direct exposure to crypto users and exchange-linked payment flows. It’s now exploring a stablecoin wallet and remittance services, including work with Ripple.

Payments and Card Networks

Shinhan Card

On the card side, one of the country’s largest payment companies Shinhan Card partnered with the Solana Foundation to build real-world stablecoin payment infrastructure.

With 28 million cardholders and 200 trillion won (approximately $145 billion) in annual transaction volume, Shinhan Card brings the kind of distribution that makes this partnership genuinely meaningful, not just symbolic.

BC Card

BC Card, one of Korea’s largest payment networks, is less visible to consumers than individual card brands like Hyundai Card or Samsung Card, because it operates underlying processing and settlement infrastructure that banks and other issuers rely on. In late 2025, BC Card completed a two-month pilot that let foreign visitors spend stablecoins with Korean merchants. By routing stablecoins through the established card authorization-and-settlement layer, the pilot addressed the volatility and real-time reconciliation problems that had previously kept stablecoins out of domestic card payments.

Danal

Outside traditional finance, Danal is one of the more interesting companies to watch, because they’re starting from a very different position than many crypto-native payment projects.

The company has spent decades building payment infrastructure in Korea and operates Paycoin, a digital payments network that reaches 3.2 million users and 150,000 merchants. It launched its won-pegged Korean Stable Coin (KSC) at Korea Blockchain Week 2025 and became the first Korean company to join Circle’s Alliance Program. The significance of this move is less about the stablecoin itself, and more about distribution. Through Paycoin, Danal already has consumer and merchant relationships that many stablecoin projects are still trying to build.

Consumer platforms

KakaoPay and NAVER Pay

In Korea, internet platforms often have financial distribution advantages, as consumer hubs with enormous reach. These platforms sit within daily consumer behavior through a variety of services, from search and messaging to banking and shopping.

This position gives internet platforms something banks usually have to work harder for: high-frequency user relationships, and wallet-like interfaces already embedded in everyday life.

A quick lay of the land for those outside Korea:

• KakaoPay sits inside the Kakao ecosystem, which includes Korea’s dominant messaging app, KakaoTalk.
• NAVER is the country’s dominant internet platform, occupying the role that Google holds elsewhere, running Korea’s largest search engine, news aggregator, and digital payment service through Naver Financial.
• Naver Financial has also agreed to acquire Dunamu, the company behind Upbit, Korea’s largest crypto exchange, further underscoring the expansion of Korea’s internet platforms across payments, financial services, and digital assets.

Both Kakao and NAVER are embedded in services where hundreds of millions of transactions flow through every year. Both have signaled interest in stablecoins as part of their long-term payments strategy. KakaoPay, for its part, has declared the won stablecoin its second growth axis for international expansion.

Implications for crypto companies

The bottom line for crypto-native projects is that the market is early, but relationships are already being built. By the time regulation in Korea clears, many of the important choices may already have been made: which chains institutions trust, which wallets they integrate, which custody providers they use, and which teams they already know.

That means the window for relationship building is open now. Korean institutions are still learning what blockchain infrastructure can do in practice. The projects that show up, make themselves useful, and work with institutions at their own pace will have a much better chance of becoming part of Korea’s stablecoin infrastructure when the market opens more fully.

Some good examples of this are Solana’s partnership with Shinhan Card; Avalanche hosting KRW1; LayerZero partnering with Korea Gold Exchange and Nexpace, which is Nexon’s blockchain subsidiary; and Kaia powering KB’s end-to-end stablecoin pilot. None of these happened because those networks waited to be invited — these relationships were built proactively by crypto-native companies.

The same dynamic plays out at the infrastructure layer. Fireblocks is embedded inside NH Bank’s tokenization stack. BitGo established a dedicated Korean entity with Hana Financial (25% stake) and SK Telecom (10%) as co-investors. These are structural integrations that were built before the regulatory framework was clear.

Traditional financial institutions move deliberately, build consensus carefully, and take time to trust. The crypto-native projects that will win in Korea are the ones willing to match that rhythm without losing their edge: the very thing that makes them compelling to these companies.

RWAs: From pilot to pipeline

Korea’s tokenized asset market tells a very similar story to stablecoins. Regulation is still catching up, but institutional commitment has translated into real infrastructure faster than most outside observers have noticed.

Most global discussion around RWAs focused on U.S. Treasuries or private credit. Korea’s approach casts a much wider net, while being more locally specific. Securities firms are already running sandbox-stage issuances across real estate, bonds, gold, carbon credits, and short-term debt.

But the more interesting activity is actually happening around assets tied to Korea’s industrial and cultural strengths: Mirae Asset Securities (one of Korea’s largest securities firms) working with Korea Land Trust and HJ Heavy Industries (a major shipbuilder) to explore tokenized ship financing. Hanwha Investment & Securities (part of the Hanwha conglomerate) announced plans to tokenize assets from Hanwha Group’s defense supply chain. Story Protocol is partnering with Seoul Exchange, a fractional investment platform, to bring K-pop royalties and creative IP onchain.

Both Mirae and Hanwha are owner-driven organizations, which matters in a market where long-term conviction and speed of decision-making can be an advantage. Mirae Asset has been aggressive across blockchain generally, building its own products instead of waiting for the ecosystem to mature around it. Hanwha has staked out one of the more specific roadmaps in the market, including a retail-facing tokenized security platform planned for mid-2026 and tokenization work tied to Hanwha Group’s defense supply chain.

Taken together, these partnerships suggest that Korea’s RWA market is developing around sectors where Korea already has global strengths, like shipping, industrial supply chains, and entertainment — instead of coalescing around some global playbook.

The legal foundation is also coming into place. In early 2026, the National Assembly passed amendments to the Capital Markets Act and Electronic Securities Act, creating a clearer path for security token offerings. The law does not take effect until early 2027, so meaningful volume is still limited, but the framework is already forming.

Two competing platforms have received preliminary approval to operate the trading systems that would let tokenized securities be issued, bought, sold, and settled under Korea’s new framework: NXT, backed by Shinhan Investment Securities, Hana Securities, and Eugene Investment; and KDX. The Korea Exchange is also building a complementary secondary market — addressing liquidity after issuance — which is one of the biggest gaps in tokenized assets globally.

Shinhan Investment Securities is the early frontrunner. It completed 10 investment contract issuances in 2025 and has built a 50-plus-firm alliance around the NXT platform. NH Investment Securities and Mirae Asset Securities are close behind, with pipelines extending into maritime assets and IP. None of these partnerships are large-scale yet, but the positions being staked now will be hard to dislodge once the market opens to them.

Implications for crypto-native projects

Korea’s securities firms have strong fundamentals. They also have the regulatory licenses, institutional relationships, and asset pipelines. What they don’t have is the infrastructure layer around them, and that’s where crypto-native projects can find the biggest opportunities.

Three gaps matter most:

1. Global distribution. Korean institutions are tokenizing assets that global investors might want, but they still need ways to reach those investors. Crypto-native projects that can help bridge that gap that most domestic institutions can’t fill on their own, connecting Korean issuers to liquidity pools and buyers outside Korea.
2. Liquidity and cross-chain interoperability. A tokenized asset trapped on one chain is only marginally better than a traditional security. LayerZero’s work with Korea Gold Exchange Digital Asset shows the role interoperability providers can play. Its partnerships around tokenized gold are broadly intended to let the asset move across multiple chains without fragmenting liquidity across separate wrapped tokens. The bigger opportunity here is helping tokenized assets reach more markets without becoming trapped inside a single venue or chain.
3. Infrastructure that supports rather than competes. These institutions know their business and have their own core assets. They’re not looking to be disrupted. The projects that win here are the ones that make Korean institutions better at what they already do, which is structuring, distributing, and managing tokenized assets at scale. The picks-and-shovels opportunity in Korean RWA is still largely unclaimed.

So where are the users?

The most consequential crypto move in Korea right now may actually be happening at the consumer platform layer.

In late 2025, NAVER confirmed its acquisition of Dunamu, the operator of Upbit. The deal still requires regulatory approval, but the outlook is optimistic. If it closes, NAVER Financial would combine Korea’s largest crypto exchange with Naver Pay’s 34 million users.

This combination matters because it brings retail crypto trading infrastructure and everyday consumer payments, which usually sit apart. NAVER has also been developing its own wallet infrastructure, while Dunamu built Giwa, an OP Stack-based L2, signaling that Upbit’s ambitions extend well beyond exchange operations and into onchain business.

Kakao is taking a different route, with work happening at the subsidiary level. KakaoBank has been exploring stablecoin development, hiring blockchain engineers and building toward a KRW-pegged coin.

Coming off its first annual profit and positioning blockchain as its next growth axis, the company is building toward a super wallet. This could be a unified interface spanning KakaoPay, KakaoBank, and KakaoTalk that can hold fiat currency and stablecoins.

Together, NAVER and the Kakao ecosystem represent two of the most powerful consumer distribution channels in the country, both moving toward blockchain-as-infrastructure rather than feature.

Finally, there’s Toss, which holds banking, securities, and insurance licenses — and its app already behaves like a financial wallet. In mid-2025, Toss filed trademarks for 24 won stablecoin names and stood up a dedicated blockchain unit. It has also indicated that a proprietary L1 mainnet with an integrated wallet is something it may explore down the road.

Toss’s payments infrastructure already runs deep. Beyond digital transfers, Toss has deployed physical point-of-sale terminals in offline merchants across the country, giving it a footprint that extends from app to physical retail. Payments and remittance are two of the clearest near-term use cases for stablecoins, and Toss is already embedded in both.

For crypto native projects, this is the distribution layer to watch. Banks may shape issuance and compliance, and securities firms may shape tokenized assets. Consumer platforms, however, can decide how blockchain products actually reach users.

Korea is at an inflection point. The regulatory framework is coming together, institutional infrastructure is being built, and the question is no longer whether Korea’s establishment will engage with blockchains — they already are — it’s which projects and protocols will become part of that infrastructure?

One external factor worth watching is the U.S. CLARITY Act. Korean institutions don’t move in a vacuum. As the U.S. works toward clearer crypto market structure domestically, it also gives international regulators including Korean agencies and institutions a very helpful reference point. Clearer U.S. rules would make it easier for Korean institutions to make the case for deeper blockchain integration internally.

Korea already has what it needs when a regulatory framework arrives — including deep retail engagement, sophisticated financial institutions that have already done the groundwork, and a tech-native population that understands and uses digital assets at scale. The crypto companies that have spent this period building genuine relationships and real use cases alongside Korean institutions will help define what’s next.

I’ll share more on other markets across Asia next.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Weekly prediction market volume hit $14.4 billion for the first time last week — up from roughly $5-6B at the start of the year. The previous all-time high, roughly $10B, was set only a week prior.

Total volume across all platforms is now more than 10x what it was a year ago.

Open interest hit $1.6B last week — the third consecutive weekly all-time high.

Unlike volume, open interest can only grow when new positions are opened faster than existing ones resolve, so the growth seen here reflects how more money is being put at stake in these markets.

The chart has grown steadily since last fall. Over that stretch, open interest has increased roughly 8x, from under $200M to $1.6B.

The World Cup is attracting lots of interest and activity, and other areas are growing alongside. Non-sports volume — across categories like politics, economics, geopolitics, and current events — hit $3.6B last week on Kalshi and Polymarket combined. That’s larger than total prediction market volume (sports included) was just last year.

In July 2025, non-sports weekly volume was around $200M. It has grown roughly 18x since then, with the trajectory steepening sharply this month.

***

Acknowledgments: Data and charts by Ryan Holloway and Robert Hackett (a16z crypto features editor and head of special projects).

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

At the center is compute, which has quickly become one of the most important commodities in the world. Yet the market around it remains shockingly immature.

Every other capital-intensive commodity — oil, real estate, energy — has developed the tools needed to price, hedge, finance, and trade risk. These markets have trusted price benchmarks, forward curves to plan and hedge against, frameworks for underwriting asset prices down the road, and instruments that turn future cash flows into investable assets. 

Compute has almost none of this. Prices are opaque and negotiated deal by deal. Operators raising capital to build new capacity can’t easily hedge the revenue projections they’re underwriting against. Investors who want exposure to the growth of AI infrastructure have few direct ways to get it. The result is a multi-trillion dollar asset class financed largely through private spreadsheets and handshakes. Ornn is building the financial infrastructure that addresses this. 

That’s why we’re so excited to lead their $33 million seed round as they help transform compute from a collection of backroom deals into a functioning marketplace.

Every market starts with a price. Ornn’s first contribution here is the Ornn Compute Price Index (OCPI), an institutional-grade benchmark for GPU compute built from cleared trades rather than scraped list prices, and spanning hardware types, regions, and contract durations. The market has already started to adopt the index as a trusted benchmark. OCPI is live on the Bloomberg Terminal, and already being used by some of the most trusted names in finance. 

The team has since extended the same idea to AI itself with the Ornn Token Price Indices (OTPI): benchmarks for the realized cost of inference tokens from the leading model providers.

On top of that data layer, Ornn is building the rest of the commodity market stack that compute has long lacked. 

• Futures let operators and enterprises hedge forward pricing and convert uncertain rental income into more predictable cash flows. 
• Residual value protection replaces crude straight-line depreciation schedules with market-based estimates of what GPUs will be worth years out. 
• Capital markets products broaden access to investors, giving them exposure to data center revenues without requiring them to own or operate the underlying infrastructure. 

Ornn is building the infrastructure that allows compute markets to work, regardless of where prices ultimately go.

Kush Bavaria and Wayne Nelms met at MIT and together bring an unusual fluency across capital markets, quantitative trading, and technical systems. Their core idea is simple: compute is becoming one of the most important commodities in the world, but the markets around it are failing to keep pace.

Ornn is building a modern marketplace that delivers clear price signals, the tools to hedge, and the capital rails to make AI infrastructure as financeable as oil or real estate. 

We’re excited to back Kush, Wayne, and the entire Ornn team as they build the financial future of compute, one of the world’s most important commodities.

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investment-list/.

The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures/ for additional important information.

Prediction markets hit a record $10.8 billion in weekly trading volume in the week ending June 15 — their biggest week ever. Several big events coincided that contributed to the all-time high: the SpaceX IPO, a U.S.-Iran peace deal, the NBA Finals, the Stanley Cup, and the opening of the World Cup, among them. 

A year ago, a typical week on prediction markets ran around half a billion dollars, and even the busiest weeks stayed under $1 billion. That floor has climbed steadily since. It climbed past $1 billion last fall, past $4 billion by winter, and into the $6–7 billion range this spring. 

Even a quiet week today dwarfs the biggest week of a year ago.

Prediction market open interest — the sum of outstanding contracts that remain “open” in the market — has also reached a record weekly high. It hit $1.48 billion in the week ending June 15, the second record-setting week in a row. 

Unlike trading volume, which reflects daily activity, open interest tracks the capital that participants have deployed and which remains at risk. The recent surge represents a roughly sixfold increase over the past year, highlighting how users are maintaining longer-term financial positions and how prediction markets are maturing into durable market infrastructure.

Big events bring in the crowds, and more traders are staying to participate in a wide range of markets across politics, economics, culture, crypto, and more.

***

Acknowledgments: Data, charts, and editorial by Ryan Holloway and Robert Hackett (a16z crypto features editor and head of special projects).

***

This article was translated into Korean by Token Post, available here.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.

Every blockchain today leans on replication ideas worked out in the 1980s, by a Turing Award winner who wasn’t thinking about how it might apply to money at all.

In this episode of First Principles, a16z crypto Head of Research and Columbia professor Tim Roughgarden speaks with Barbara Liskov, MIT professor, Turing Award winner, and one of the most influential computer scientists in programming languages, data abstraction, fault tolerance, and distributed computing. a16z crypto research partner Ittai Abraham joins the conversation.

The discussion traces Liskov’s path from programming languages and modularity to distributed systems research; from CLU and Argus to viewstamped replication; and from benign failures to Practical Byzantine Fault Tolerance, or PBFT — a protocol family whose ideas now shape many modern blockchain systems.

Liskov explains why modularity matters, how systems researchers thought about replication in the 1980s, why view changes were such a key idea, and how PBFT extended earlier work to handle malicious behavior on the internet. The conversation also explores the bridge between theory and practice, the importance of proofs and specifications, and why the next generation of systems research may be reshaped by AI.

About the series

First Principles is a special, limited series from a16z crypto about the scientific roots of modern computing — especially blockchains — told through rare conversations with the pioneers who helped shape the foundational ideas behind distributed systems, consensus protocols, economics, mechanism design, cryptography, zero-knowledge, and more. People often tell the story of the Bitcoin whitepaper as if it appeared out of nowhere. But the ideas behind Bitcoin — and behind blockchains more broadly — come from decades of computer science, economics, mathematics, and cryptography. First Principles is a guide to that lineage, as told by the people who helped build it.

Subscribe here to follow along

Highlights

00:00 Intro: How do systems stay reliable when parts fail?
01:18 Barbara Liskov’s path from programming languages to distributed systems
05:45 Why modularity is “everything”
07:22 The replication problem: keeping data available across many machines
09:58 Viewstamped replication and the “ledger” before blockchains
16:32 Why good research starts with what you don’t understand
18:10 Leslie Lamport, Paxos, and the inevitability of ideas in the right time, in the right place
21:48 Practical Byzantine Fault Tolerance: what changes when replicas can lie
19:35 How PBFT bridged theory and practical systems
22:38 Why you should never trust an individual replica
28:39 Why blockchains are state machine replication in the wild
31:27 AI, verification, and the future of computer science

Follow:

• Tim Roughgarden: https://x.com/Tim_Roughgarden
• Ittai Abraham: https://x.com/ittaia Follow

Follow a16z crypto:

• X: https://x.com/a16zcrypto
• LinkedIn: https://www.linkedin.com/showcase/a16zcrypto/posts/
• YouTube: https://www.youtube.com/@a16zcrypto
• Substack: https://a16zcrypto.substack.com/subscribe/

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investment-list/.

The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures/ for additional important information.

Bitcoin often gets credited with inventing trustless consensus. It didn’t.

The problem was named decades earlier — in the world of distributed computing — and researchers spent years studying how machines could reach agreement even when some participants were faulty, adversarial, or corrupt.

What Bitcoin did was something different: It solved a classic Byzantine agreement problem in a radically new, permissionless setting. And it took the research world years to fully recognize what Satoshi had done.

In this episode of First Principles, a16z crypto Head of Research and Columbia professor Tim Roughgarden ( @timroughgardenlectures1861 ) is joined by a16z crypto research partner Ittai Abraham — one of the world’s leading researchers in Byzantine agreement and consensus protocols, a founding member of VMware’s blockchain project, and founder of the technical blog Decentralized Thoughts — to unpack the scientific roots of blockchain consensus.

Together, Tim and Ittai trace the line from classic distributed systems research to Bitcoin, proof-of-stake, Tendermint, Casper, DAG-based protocols, Solana’s Alpenglow, and the modern race for higher throughput and lower latency. Along the way, they explain why concepts like Byzantine fault tolerance, state machine replication, safety, liveness, and partial synchrony are not just academic abstractions — they are the language and design principles behind today’s blockchain protocols.

About the series

This conversation kicks off First Principles: The Scientific Roots of Blockchain Technology — a special, limited series from a16z crypto on the scientific ideas behind modern computing — especially blockchains — told through conversations with the pioneers who helped create them, including Barbara Liskov, Leslie Lamport, and more. Hosted by Tim Roughgarden, the series explores the foundational concepts behind distributed systems and consensus protocols; economics, mechanism and market design; and cryptography, from digital signatures to zero knowledge.

People often tell the story of the Bitcoin whitepaper as if it appeared out of nowhere. But the ideas behind Bitcoin — and behind blockchains more broadly — come from decades of computer science, economics, mathematics, and cryptography. First Principles is a guide to that lineage, told by the people who helped build it.

Subscribe here to follow along

Highlights

00:00 Introduction to First Principles: The Scientific Roots of Blockchain Technology
00:56 Why consensus matters for blockchains
02:30 Byzantine agreement: The old computer science problem Bitcoin made practical
04:34 Blockchains as a shared system of record: State machine replication and blockchain state
06:41 How two research worlds — distributed computing and crypto — began to converge
07:49 Proof of work vs. proof of stake
09:27 Why Ethereum’s move to proof-of-stake took years
11:08 When crypto rediscovered decades of distributed systems research
11:50 Why BFT became practical
12:49 Throughput, latency, and modern consensus design
14:05 DAG-based protocols and faster blockchains
15:25 Peace time vs. war time: why modern blockchains need two modes
16:47 Theory, practice, and the future of blockchain research

Hear more from:
Tim Roughgarden: https://x.com/Tim_Roughgarden
Ittai Abraham: https://x.com/ittaia

Follow a16z crypto:
X: https://x.com/a16zcrypto
LinkedIn: https://www.linkedin.com/showcase/a16zcrypto/posts/
YouTube: https://www.youtube.com/@a16zcrypto
Substack: https://a16zcrypto.substack.com/subscribe/

***

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. Certain information contained in here has been obtained from third-party sources, including from portfolio companies of funds managed by a16z. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the current or enduring accuracy of the information or its appropriateness for a given situation. In addition, this content may include third-party advertisements; a16z has not reviewed such advertisements and does not endorse any advertising content contained therein.

This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z, and there can be no assurance that the investments will be profitable or that other investments made in the future will have similar characteristics or results. A list of investments made by funds managed by Andreessen Horowitz (excluding investments for which the issuer has not provided permission for a16z to disclose publicly as well as unannounced investments in publicly traded digital assets) is available at https://a16z.com/investments/.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Past performance is not indicative of future results. The content speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects, and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see https://a16z.com/disclosures for additional important information.