Same Stock, Three Prices

Where the two rails actually are

Start with the scoreboard, because it has moved in an instructive way since May.

The issuer-cooperative rail — what I called category two, the token that legally is the share — is shipping on schedule. Nasdaq's rule change for tokenized trading of Russell 1000 stocks and index ETFs was approved in March; NYSE's parent ICE got equivalent changes through in April. The piece that makes it real is settlement: DTCC announced in early May that its tokenization service, built on the ComposerX platform, will support limited production trades in July and launch fully in October, covering Russell 1000 constituents, major-index ETFs, and US Treasuries. More than fifty firms are in the working group, including BlackRock, Goldman Sachs, JPMorgan, Circle, and Ondo. The legal foundation is a December 2025 no-action letter giving DTC a three-year window to tokenize precisely that list of highly liquid assets. None of this needed new law. It is the 1970s dematerialisation of paper certificates running again, with a different database underneath, which is exactly why it could move this fast.

The third-party wrapper rail — category three, the unsponsored depository receipt on a new ledger — went the other way. Bloomberg reported on May 22 that the exemption had been pulled back days before publication, after Nasdaq, NYSE, and Cboe leadership pushed back hard in closed-door meetings. The reported sticking point is the provision that would let third parties issue tokens against company shares without the issuer's knowledge or approval, and specifically the question of how public companies are supposed to administer dividends and count shareholder votes as wrappers proliferate across networks. As of writing, there is no new timeline.

I want to flag what stalled it, because it is not a random detail. In the May essay I argued the rights-passthrough condition was the load-bearing part of the framework — the thing that separates a faithful wrapper from a synthetic with better marketing. The January staff statement had already hedged on exactly this point, saying tokenized assets "may or may not confer" the rights of the underlying security. The delay is the SEC discovering, under pressure from the incumbents, that the load-bearing condition is genuinely hard to specify: mandatory pass-through is operationally demanding for issuers who never consented to the wrapper, and optional pass-through collapses category three into category one. The exemption is stuck on the one clause that makes it coherent. That is not a reason to think it's dead. It is a reason to think the final text, whenever it lands, will be defined by how it resolves that clause, and nothing else in it will matter as much.

The third market nobody authorised

While the two US rails were being negotiated, a third market shipped. Hyperliquid's HIP-3 standard, live since November, lets approved builders deploy perpetual futures on anything with a reliable reference price — and the dominant builder, TradeXYZ, chose US equities. There are now 24/7 perpetuals on the S&P 500 (the first officially licensed one, under an agreement with S&P Dow Jones Indices), the Nasdaq-100, and the major single names — NVDA, TSLA, AAPL, MSFT. Reported open interest in these builder-deployed markets grew from roughly $800 million in January to over $1.4 billion by March and past $2.5 billion by late spring, and on the strongest days the equity and commodity perps have accounted for close to half of Hyperliquid's total volume. For scale, the entire offshore tokenized-stock spot market — Backed, Dinari, xStocks and the rest — is around $1.4 billion outstanding, having 40x'd from under $30 million at the start of 2025.

The perp market shipped first for a structural reason, not a regulatory-arbitrage one. A spot wrapper needs custody, a mint-and-burn pipe, dividend plumbing, and a legal theory. A perpetual needs a reference price and a margin system. It imports none of the share's rights and therefore none of the share's administrative problems — no dividends to pass through, no votes to count, no custodian to trust. It is the purest possible product: price exposure and nothing else. That purity is exactly why it could go live years before the products that carry the full bundle, and it is also why it can never replace them. A perp cannot be redeemed into a share. It has no anchor except the willingness of arbitrageurs to trade the funding rate against some hedgeable reference, which — as with the pre-launch perps I wrote about in May — is where all the interesting information hides.

Three prices, one claim

So by October the same underlying corporate cash-flow stream — pick Apple — will trade in three forms. It is worth being precise about what each form actually contains, because the price differences follow mechanically from the contents.

A conventional share is a bundle: the cash-flow claim, the vote, the legal recourse of a registered shareholder, and an access constraint — you trade it during exchange hours, through a broker, with T+1 settlement. The DTC token is the same bundle with the same CUSIP on a faster database; by construction its price cannot detach from the share's, because they are interchangeable at the clearinghouse. It is not a second market. It is the first market with better plumbing, and the open question is only whether the plumbing runs on weekends.

The category three wrapper, when it arrives, is the bundle minus recourse plus hours: the holder gets the economics (if pass-through is mandatory), a diluted proxy for the vote, weaker legal standing than a registered shareholder, and a market that never closes. The perp is neither of these. It is the price alone, rented period by period at the funding rate.

The law of one price says all three should trade together. But the law of one price is not a law of nature — it is a statement about arbitrage capacity, and each pair of these markets is connected by a different pipe. The share and the DTC token share a clearinghouse: basis pinned at zero. The share and the wrapper are connected by a mint-and-redeem pipe that will be KYC-gated at a licensed platform, which means the arbitrage is not free entry — it is a quota. Write the wrapper basis as B = P_wrapper − P_share. Arbitrage bounds |B| only up to the cost of a round trip through the pipe plus the shadow price of the pipe's capacity. When the pipe is wide and cheap, B is a few basis points. When the pipe is narrow, licensed, or one-directional, B can be anything. We have run this experiment many times: unsponsored ADRs trade at persistent premiums and discounts when conversion is restricted; ETFs held their pennies-wide basis for years and then gapped to multi-percent discounts in March 2020 when authorised participants stepped back; GBTC traded 40% above and then 40% below its NAV purely on the direction its one-way pipe pointed. The wrapper's basis will be a live gauge of exactly how much redemption capacity the final exemption text actually grants — a number the SEC will publish without meaning to.

The perp has no pipe at all, only funding. The funding rate is the recurring fee longs pay shorts (or vice versa) to hold the price to its reference, f ≈ (P_perp − P_ref)/P_ref per funding interval, and it is best read as the shadow price of the missing arbitrage leg. During US market hours, an equity perp is trivially hedgeable against the underlying and funding should sit near the cost of carry. From Friday close to Monday open, the hedge leg does not exist. Whatever the perp does over the weekend, no arbitrageur can hold it to a spot price that isn't printing. Weekend funding on equity perps is therefore the market's price for bearing gap risk to Monday's open — a number that has simply never been continuously observable in US equities before, because there was never a venue open to print it.

Who owns the weekend

This is the part of the structure I think matters most and gets discussed least. A US exchange session is six and a half hours, five days a week — call it 32.5 of the week's 168 hours, or 19%. CME equity index futures extend that substantially, but even they go dark from Friday evening to Sunday 6pm Eastern: roughly 49 hours in which, today, the only live, executable prices on US equity risk are on-chain perps. That is not a hypothetical about some future 3am crisis. It is already the case, every weekend, that the marginal price-setter for the S&P 500 is a perpetual on Hyperliquid. When news breaks on a Saturday, the perp gaps, funding spikes, and Monday's official open "discovers" a price the on-chain market has been trading against for thirty-six hours.

The regulated rails get to decide how much of the clock they want. If DTC's October launch keeps standard clearing hours — and nothing in the announced scope suggests otherwise — then the issuer-cooperative rail owns the 19% of the week it has always owned, with faster settlement inside it. The other 81% belongs to whoever is open: today the perps, tomorrow the perps plus whatever wrappers the exemption eventually permits. Price discovery migrates to the venue that is switched on. That is not an argument that the on-chain price is better — thin weekend books produce noisy prices, and the oracle-latency problems I wrote about in April get worse when the reference market is closed. It is an argument that the on-chain price is increasingly the only price, for most of the hours in a week, for the most-watched equity market in the world. Institutions that mark risk on Friday's close are already carrying an unmarked two-day option against everyone who doesn't.

What to watch

Three observables will tell you how this resolves, and none of them is a press release.

First, the pass-through clause in the final exemption text, whenever it emerges from the current delay. If dividend and voting pass-through is mandatory with delisting as the penalty, category three is a real product and the wrapper basis will trade tight. If pass-through is optional or "disclosed", the SEC will have authorised category one with better paperwork, the wrapper will trade like the synthetic it is, and the offshore venues will keep the flow they already have.

Second, the scope of DTC's October launch — specifically whether tokenized shares can leave the walled garden. A DTC token that moves only between DTC participants is a settlement upgrade, valuable and invisible. A DTC token that can sit in external wallets and post as collateral on-chain is a different animal entirely, and would collapse most of the wrapper rail's reason to exist before it launches.

Third, weekend funding on the HIP-3 equity perps. It is the one number that aggregates everything above: how much gap risk costs, how much hedging capacity exists off-hours, and how credible the 24/7 spot alternatives are. If a licensed wrapper with genuine redemption goes live and weekend funding on the S&P perp compresses toward carry, the pipes are connecting. If funding stays wild, the three markets are still three islands. Either way the number is public, updates hourly, and doesn't care about anyone's roadmap.

The recurring theme of these essays is that the wrapper matters more than the underlying. This is the year that stops being a metaphor. For a century the equity bundle — cash flows, votes, recourse, hours — was priced as a single indivisible package, and you could believe the package price was about the company because there was nothing to compare it against. By the end of 2026 the components trade separately: the full bundle at DTC, the economics-only wrapper offshore or under the exemption, the naked price on a perp that never closes. Three prices for the same stock sounds like fragmentation, and the incumbent exchanges will use that word in every comment letter. To a market microstructure person it is something better: an unbundling, with a basis you can finally measure. The company is the same in all three. Everything the prices disagree about is the wrapper — which was always the honest way to read them.