Tokenized Stocks Without the Issuer's Permission

What Bloomberg actually reported

The leak landed on May 18. The framework is being prepared under SEC Chair Paul Atkins' "Project Crypto" initiative, with Commissioner Hester Peirce reportedly central to drafting. The headline features being circulated are these. Crypto-native platforms will be able to offer tokenized versions of publicly traded US equities without securing full broker-dealer or exchange registration. Third parties can mint these tokens without direct approval from the underlying companies. The tokens themselves are expected to be required to carry the same economic rights as common stock, including dividend pass-through and voting access, with delisting as the consequence of failing to deliver on either. The framework is time-limited and experimental, with exposure caps and disclosure requirements baked in. DTCC has confirmed it plans limited production trades of tokenized assets in July with a broader launch in October. Banks and traditional exchanges have already filed objections on custody, anti-money-laundering, and market-fragmentation grounds.

The framing in most secondary coverage is some version of "the SEC is about to let anyone tokenize anything." That is not what is happening. The exemption opens a very specific lane, and the lane has been carefully walled off from two other categories the SEC took an explicit position on in January. Distinguishing the three is where most of the analysis has to start.

The three things "tokenized stock" can actually mean

Equity-linked tokens have been floating around crypto venues for the last five years, and the label has been used loosely enough to cover structurally different products. The new exemption only touches one of them. The other two stay where they were.

Category one: the synthetic. A token that tracks a stock's price via an oracle, collateralised by something other than the stock itself. The canonical version was Mirror Protocol on Terra, where a basket of UST and other crypto collateral backed tokens that referenced Apple, Tesla, and a long list of other US names. FTX's "tokenized stocks" worked the same way, with the exchange's balance sheet as the implicit backstop. No real share sits behind the token. The wrapper is a price feed plus a promise. From a securities-law standpoint the cleanest analogue is a security-based swap, which has its own registration regime that nobody in DeFi has ever complied with.

Category two: the issuer-authorised native token. The company itself integrates its blockchain record into its official shareholder register. The token is, legally, the share. This is what Nasdaq's tokenized-securities plan, approved by the SEC in March, is building towards. NYSE's parent ICE got equivalent rule changes through in April. Coinbase has filed for something in the same direction. Conceptually it's the same as moving the stock from a paper certificate to DTC's electronic book-entry system in the 1970s, with a different database underneath. The issuer cooperates by definition.

Category three: the third-party fully-collateralised wrapper. A custodian, broker, or platform holds the real share in a regulated custody account, mints a token against it one-for-one, pays dividends through to the token holder as they're received from the issuer, and arranges voting via the custodian acting as proxy. The issuer doesn't have to authorise any of this. The token holder ends up with the same economic exposure as a direct shareholder, just routed through one more layer.

The January 2026 staff statement from Corporation Finance, which is the document everything since has been built on top of, said category one tokens are a problem and don't change their securities-law treatment by moving to a blockchain. It said category two tokens are fine because they are just shares on a different ledger. And it explicitly flagged category three as a separate question that needed its own answer. The May exemption is that answer.

What the exemption actually does

Stripping out the press-release language, the exemption does three specific things.

First, it says category three tokens, with full pass-through of dividends and voting, can be issued by third parties without the underlying issuer's consent. This is the headline change. It is the part OpenAI and Anthropic have publicly opposed in their context, where the shares being wrapped are theirs and they don't get a veto.

Second, it says the platforms minting and trading those tokens don't need to register as broker-dealers or exchanges in full. They sit under a lighter regime with exposure caps, disclosure requirements, and the sandbox-style oversight that goes with an experimental authorisation. This is the load-bearing part for the venues. Full broker-dealer registration is the wall that has kept crypto-native firms out of the US equity market for a decade.

Third, it does this on a time-limited basis, which is how the SEC manages exemption authority within Section 28 of the Exchange Act. The SEC can't simply repeal registration requirements by fiat. It can grant conditional, time-bounded exemptions when it judges the public interest is served, with the implicit understanding that the framework will be reviewed and either codified or revoked at the end of the period. This isn't a permanent rule change. It's a sandbox with teeth.

The things the exemption does not do are equally important.

It does not legalise the synthetic. Mirror-style price-tracking tokens remain security-based swaps offered without registration to retail, which is still illegal in the US. Anyone reading the news as a green light for pure synthetics is reading it wrong.

It does not let anyone wrap anything. The pass-through requirement is binding. If a platform mints a token, advertises it as a wrapper of AAPL, and then fails to actually credit dividends to holders on the record date, the token gets delisted. The delisting consequence is what keeps the wrapper honest. Without it, category three drifts back into category one.

It does not solve the private-company problem. There is no public share to custody for OpenAI, Anthropic, SpaceX, or Stripe. The exemption only works for already-public equities, because the wrapper has to be backed by a real, transferable share. The prestocks already trading on Jupiter and elsewhere are unaffected by this rule.

The mechanism the SEC is borrowing from

The clearest way to understand what the exemption is doing is to look at where the same trick has been run before. The answer is "in the foreign equity market, since 1927, by the largest US banks."

Here is the setup, simplified. A US investor in 1930 wants to own shares in a German industrial company. The shares are denominated in reichsmarks, trade on the Berlin exchange, pay dividends in reichsmarks on a German calendar, vote at German shareholder meetings, and settle through the German clearing system. None of that is accessible to a US brokerage account. The mechanical friction alone makes the holding impractical, even before legal questions about cross-border ownership.

The solution that JPMorgan invented for Selfridges in 1927 and that Citibank and BNY then scaled across the foreign equity universe is the depository receipt. A US bank buys the foreign share, holds it in custody at a local sub-custodian in the foreign market, and issues a USD-denominated paper receipt against it. The receipt is registered with the SEC under a specific form, currently Form F-6. The receipt holder gets the dividends, collected by the depository in the foreign currency and passed through in USD, net of FX and a small handling fee. The receipt holder gets voting rights through the depository, which acts as proxy on a best-efforts basis. The receipt can in principle be redeemed back for the underlying share. From the receipt holder's standpoint, the product is functionally equivalent to owning the foreign share, with a US ticker and US settlement.

Two flavours emerged. The "sponsored" version is set up in cooperation with the foreign issuer, usually because the issuer wants to raise capital in US markets. The "unsponsored" version is set up by the depository bank without the issuer's involvement, because there's US investor demand and the bank decides there's a viable market. The foreign issuer doesn't have to authorise an unsponsored program. There have been periods where issuers actively opposed unsponsored programs on their shares and were unable to prevent them, because the underlying legal question (do investors have a right to wrap shares they own and issue receipts against them?) has been answered yes for nearly a century.

That is the precedent the May exemption is built on. The economic argument is identical. As long as the wrapping is faithful, meaning the share is actually held, dividends actually flow, and voting actually works, the issuer's interest is preserved. The issuer can object on any number of business or reputational grounds. None of those grounds is a securities-law objection. The wrapper is just a routing layer.

What changes in 2026 is the ledger. Replace "Citibank in 1927" with "Coinbase in 2026," "Form F-6" with whatever new filing template the exemption ends up specifying, "paper receipt cleared through DTC" with "ERC-20 token on Ethereum or SPL token on Solana," and the mechanism is the same. The on-chain token is, structurally, an unsponsored depository receipt with a different settlement layer. The novelty is the plumbing, not the legal theory.

Why the rights-passthrough condition is doing all the work

The piece of the exemption most commentary will skim past is the requirement that tokens pass through dividends and voting. That condition is what makes the rest of the framework legally coherent.

For an unsponsored depository receipt, the bank passes dividends through by collecting them at the foreign custodian, converting from local currency, and crediting USD to the receipt-holder accounts on the record date. The bank arranges voting by polling receipt holders and submitting a proxy at the foreign shareholder meeting. The pass-through is enforceable in practice because the depository is a US-regulated bank with assets, audit obligations, and a public reputation to lose.

For a tokenized stock under the new exemption, the same pass-through has to work technically and legally. Technically, the platform needs an on-chain mechanism to credit USDC or fiat to every wallet holding the token at the dividend record date, and a way to surface voting decisions to token holders that the custodian then aggregates and submits. Several existing token standards already support snapshot-based distribution and on-chain voting infrastructure. The technical layer is mostly solved. Legally, the platform has to actually do these things, with delisting and exemption revocation as the enforcement mechanism.

Without that condition, the exemption would have legalised category one tokens by another name. With it, the line between category three and category one is sharp and observable. A platform either pays the dividend or it doesn't. The dividend either matches the issuer's declared payment or it doesn't. The vote either gets cast or it doesn't. The pass-through is binary and auditable, which is exactly the kind of condition exemption regimes need to be administrable.

This is also where the exemption parts company with crypto-native synthetic protocols. A protocol that mints a token referencing AAPL against USDC collateral, without holding the share, cannot satisfy the pass-through condition. There is no share to collect dividends from. The protocol could simulate dividends by paying USDC to holders out of fee revenue, but that turns the token into a yield-bearing synthetic with no enforceable claim to the actual dividend stream, which puts it back in category one. The exemption isn't being unfriendly to those protocols out of bias. It just doesn't recognise them as the same product as a faithfully-backed wrapper.

What this changes in practice

If the exemption ships as leaked, several things change over the next year that have not been possible under any prior framework.

The first is meaningful 24/7 on-chain price discovery for US equities, operating under US oversight, in the US. Nasdaq and NYSE have rules for tokenized equities trading alongside traditional shares, but those products inherit the underlying venue's hours and the issuer's cooperation. A category three wrapper minted by Coinbase or Robinhood Crypto trades whenever the wallet is online. For names with thin after-hours liquidity on regulated venues, the on-chain order book becomes the live price between 4pm and the next morning's open. The first time a major macroeconomic event prints at 3am London time and the only live US-equity reference price is on Solana, the reading of what tokenized equity markets are for will change overnight.

The second is collateral utility in DeFi. Aave, Morpho, and Compound currently accept a handful of major crypto assets and a growing slice of tokenized Treasuries as collateral. A tokenized AAPL that the protocols can verify is genuinely backed by AAPL shares is a different kind of collateral entirely. It carries equity beta, fits portfolio construction logic that the tokenized-Treasury rails do not, and is familiar to every retail and institutional user in a way that no DeFi-native asset is. The order of magnitude on collateral value is large. Even a single-digit-percent capture of the US public equity market's float would dwarf the current crypto-collateral universe.

The third is the loss of the regulated-exchange first-mover advantage. Nasdaq and NYSE got their tokenized-securities frameworks approved earlier in the year, which positioned them as the default venues for the issuer-cooperative model. Once the third-party model is live, any platform with custody and a token contract can mint the wrapper. The competition for tokenized-equity volume becomes a contest between cooperative issuer-led products on regulated venues and third-party wrappers on crypto-native venues. The crypto-native side doesn't need issuer cooperation, which is the slower piece of any Nasdaq-led timeline, and can list any S&P 500 name on day one.

The fourth is the natural extension into perpetual futures and options on equities, traded on-chain. Hyperliquid's HIP-3 permissionless perp framework, dYdX, and a handful of other venues already have the infrastructure to list a perp on anything with a reliable reference price. The friction has been the reference. A category three wrapper with on-chain spot liquidity is a clean reference, much cleaner than the oracles used for prestocks today. The first wave of equity perps on-chain has been waiting for exactly this layer. It will not wait long.

What it does not change

The exemption is narrower than the coverage will suggest, and the narrow parts are worth being clear about.

Private-company synthetics are unaffected. The prestock products on Jupiter, Aevo, and Whales Market that reference Anthropic, OpenAI, SpaceX, and Stripe wrap shares that aren't publicly transferable. Even if a third party wanted to mint a faithfully-backed wrapper on Anthropic stock, the company's transfer restrictions prevent the underlying share from being held in a custody account the wrapper could collateralise against. Anthropic and OpenAI have publicly opposed unauthorised tokenisation of their shares, and the May exemption does nothing to override that position because the exemption only contemplates publicly traded equities. The prestock market remains where it is, which is the synthetic-with-no-spot category I wrote about in Pre-Launch Perps Are Float Markets, Not Token Markets, with all the funding-rate and missing-arb dynamics that come with it.

Pure on-chain synthetics remain illegal. Mirror-style protocols cannot relaunch in the US under the new exemption, because they don't hold real shares and can't satisfy the pass-through condition. Whatever optionality the synthetic-equity protocols still have lives offshore or in non-US-targeted form.

KYC and AML survive at the platform layer. The exemption is widely expected to condition lighter registration on robust customer- identification and transaction-monitoring controls, which a fully permissionless DEX cannot run. The likely effect is that the tokens themselves can be held in any wallet, but the mint and redeem function, and the primary distribution layer, sit at a regulated platform with identified accounts. This is the same shape as how stablecoin issuance works in practice today. Mints and burns are regulated; secondary transfers are permissionless.

Custodial risk does not go away. The wrapper is only as safe as the custodian holding the underlying shares. A failure at the custodian makes the tokens unbacked, the same way an FTX-style failure made the synthetic stocks worthless. The risk is reduced because the custody framework will be regulated, but it is not eliminated, and the chain of enforcement is theoretical for users whose tokens sit in wallets the platform has no oversight of.

Rights enforcement reach is the subtlest constraint. The token's promise that dividends will be paid and votes routed is only worth as much as the holder's recourse if it isn't. For a US user holding a token minted by a US platform against shares held by a US custodian, recourse is straightforward. For a non-US user holding the same token through an offshore wallet, the path to remedy is murky even if the legal entitlement is identical. The exemption will likely focus enforcement on the mint-and-redeem layer rather than on every holder, which is a sensible scope choice but leaves a real gap for the secondary holders.

The geopolitics

Every other major market has shipped a tokenisation framework before the US. The EU has MiCA and the DLT Pilot Regime, which is what allowed Backed Finance, Dinari, and the rest of the EU-based tokenized equity ecosystem to operate openly while the US debated. The UK has the Digital Securities Sandbox running under the Bank of England and FCA. Hong Kong has its STO framework. Singapore has Project Guardian. Even Switzerland has the DLT Act. The US has been the laggard for a decade, not because it lacked institutional capacity but because the SEC's working position under Gensler was that existing law covered everything and nothing new needed to be written.

The May exemption is partly catch-up and partly venue protection. Category three wrapping is going to happen. Backed and Dinari prove that demand exists. The remaining question is whether the venues that capture US-domiciled flow are EU-licensed offshore platforms or US-regulated onshore platforms. The exemption tilts the answer toward onshore. From the SEC's standpoint, that's worth quite a lot. An offshore venue with US users is the worst of all worlds for a regulator. An onshore venue with US users is an entity that can be examined, fined, and pulled back if something goes wrong.

The opposition from banks and traditional exchanges is exactly what you'd expect. The fragmentation argument has merit on its face. Adding a parallel rail for the same underlying instrument splits liquidity, and the lighter regulatory regime on the new rail does create a competitive asymmetry against the registered venues. The custody and AML arguments are partly genuine and partly self-interested. There is a real point about whether retail users will understand which lane they're trading on, and which protections come with which lane. There is also a real point about whether existing exchanges should have to compete with a regime they're not eligible for.

The SEC's response, judging from the leaked structure, is to make the exemption time-limited and conditional rather than permanent. If fragmentation gets bad, the framework can be tightened. If custody failures happen, the framework can be revoked. The sandbox can be taken down faster than a registered exchange can be created. That asymmetry of policy reversibility is, I'd guess, the unstated part of the case Atkins is making internally.

The wrapper question, partly answered

The recurring theme across these essays is that the wrapper matters more than the underlying. Pre-launch perps look like project derivatives but mechanically price first-day float clearing. Prestocks look like equity exposure but mechanically price the inability to arbitrage a synthetic. Tokenized stocks under the new exemption look like deregulation but mechanically formalise the depository-receipt model on a different ledger.

The wrapper question I posed in The Race to Put Wall Street On-Chain has been, until now, an open one in US markets. Synthetic wrapping was tried in DeFi and failed for retail. Native issuance was the long path that required cooperation from every issuer one at a time. Faithful third-party wrapping was the obvious middle option and had no legal home. The May exemption gives it one.

None of that means the next twelve months will be tidy. The exemption is conditional, the pass-through condition is demanding, the custody stack is not battle-tested at scale, and the platforms most likely to build first are exactly the ones with the highest reputational risk tolerance. There will be a custody failure. There will be a wrapper that drops a dividend. There will be a platform that finds out the hard way what the difference between an exemption and a permanent rule is. What's interesting is not that the path will be smooth but that the path now exists. For ten years the question was "could this product legally exist in the US." For the next ten the question is "who builds it well." That is a much better question to be asking, and the people who watched the prestock market mature on Solana, the on-chain options AMMs settle into RFQ models, and the EU venues run live with category three wrappers for two years before the US joined them, already have the head start.

The boring reading of the news is that the SEC has authorised the depository receipt for the on-chain era. The interesting reading is that the SEC has chosen to authorise the depository receipt for the on-chain era, on a time-limited basis, with rights pass-through as the load-bearing condition, while keeping the synthetic banned and the private-company wrapper unaddressed. The first reading is the answer to a yes/no question. The second is a sequence of design decisions that will shape which products get built, which platforms capture the flow, and which corner of the on-chain stack ends up looking the most like the equity market five years from now. The yes/no question is what the headlines will cover. The design decisions are what the rest of the work will turn on.