Six factors that sort forty initiatives — and the two questions that place the next announcement in a minute
In the last week of October 2026 four things called "tokenised shares" happened in the same market, and they were four different legal objects: a depository's entitlement on a permissioned ledger, a third-party token that more than sixty issuers refused to have traded under their names, a perpetual future on a single stock brought onshore, and a transfer agent's on-chain register for an issuer that asked for it. The word is the problem. Reported as progress on one project, the announcements cannot be told apart by a reader who needs to know whether an initiative funds an issuer, re-plumbs an existing security, or sells somebody else's product on top of an issuer who was never asked.
This essay sets out six factors that separate them — economic function, offering and distribution, holding model, claim type, cash leg and collateral eligibility — and applies them to roughly forty initiatives across the United States, Europe, Switzerland, Singapore, Hong Kong, Korea and Brazil. The sort produces four clusters and three empty zones, a twelve-to-thirty-six-month prognosis, and a falsifier. It is a sort, not a league table.
In the last week of October 2026, four things called "tokenised shares" happened in the same market. DTCC's Tokenization Service went live with 34 securities, each a one-for-one entitlement against a position held at DTC, with DTC's own books remaining the official record (DTCC, 28 October 2026). Sixty-seven listed companies — Microsoft, Alphabet, Apple, JPMorgan and Goldman Sachs among them — had filed objections by 6 November to having their shares traded as tokens on the OKX–ICE venue, 56 of the 62 names the venue had proposed, as the first thirty-day windows closed (objection notices, October–November 2026). ICE announced a venture with trade[XYZ] to bring Hyperliquid-based perpetual futures onshore (ICE, 29 October 2026). And Securitize bought a full page in the Wall Street Journal addressed not to investors but to boards: "Wall Street has rules. Tokenization done right plays by them."
Then, in the first three days of November, the sort completed itself. The NYSE added four transfer agents to its tokenisation network, so that the six firms on it include the three that serve nearly all of the S&P 500 (NYSE, 2 November 2026). One Nasdaq-listed company agreed to put its own shares on a tokenised venue without asking anyone's permission, because it was the issuer (Canton Strategic Holdings, Form 8-K, 29 October 2026). And the CFTC said that perpetual futures on individual stocks may now trade on its registered exchanges (CFTC, 3 November 2026).
A sovereign bond on HSBC Orion, a DTC entitlement on Canton, a Jersey-law note on Solana and a perpetual on Hyperliquid share one word and nothing else. They are four different legal objects, held four different ways, offered to four different investor populations, and they touch the issuer in four different ways, including not at all.
The word is the problem. Every announcement is reported as progress on the same project, so a reader cannot tell whether a new initiative is a funding instrument, a change of plumbing, or somebody else's product built on top of an issuer who was never asked. This essay sets out the factors that separate them, in the order that discriminates most, and then applies the factors to roughly forty initiatives across the United States, Europe, Switzerland, Singapore, Hong Kong, Korea and Brazil. Earlier essays in this library took the pieces one at a time — whose record makes the holder the owner, when a transfer can no longer be unwound, where servicing stalls, and what a wrapped claim actually is. This one puts the pieces on one map.
The aim is a sort, not a league table. At the end, a reader should be able to place any new announcement in under a minute by asking two questions: who holds the register, and did the issuer agree?
The first question is whether the token raises money for the issuer. It sorts the field into three groups, and most of what follows is downstream of this cut. The classification attaches to the transaction, not to the firm: one platform can run a funding product, a re-plumbing service and a third-party wrapper side by side, and several do.
Funding. The token is the instrument through which capital is raised. The Hong Kong government's digital green bonds on HSBC Orion and GS DAP, the European Investment Bank's notes on GS DAP and on Clearstream's D7, the UK's DIGIT gilt pilot, tokenised commercial paper, and every tokenised fund from BlackRock's BUIDL to the Franklin Templeton and DBS retail money market fund in Singapore. A fund is continuous funding: each subscription is new money. Hong Kong took close to half of global digital bond issuance between 2025 and the first half of 2026. The centre of gravity in Asia and continental Europe sits in this group.
Re-plumbing. The security already exists and is already funded. Tokenisation changes how it is held, settled or traded, and the issuer's balance sheet is untouched. DTCC's Tokenization Service, Nasdaq's approved rule change to trade Russell 1000 stocks in tokenised form, NYSE's planned platform and its network of transfer agents, Computershare's issuer-sponsored tokens with Securitize, and Securitize's own listed shares. The centre of gravity in the United States sits here. The issuer gains, at most, a new holder record and a new channel. Nobody buys a house for its plumbing.
Third-party. The holder's claim does not run to the issuer at all. Someone else buys the shares, holds them with a custodian, and issues a token, a note or a derivative referencing them; the holder owns that instrument, and the issuer of the underlying is an input to it. Kraken's xStocks, Robinhood's European tokens, Ondo Global Markets, the OKX–ICE venue's original plan, and every perpetual future on a single stock.
The cut explains the veto at once. The objecting companies did not object to tokenisation as such, and no one asked them about DTC, because nothing about their register moved. They objected to the third group, because it is the only one in which they are not a party.
It also explains the full pages. The Securitize advertisement was a reply. The weekend before, a page headlined "Your stock. Their tokens. Your problem." had told chief executives to use their thirty-day right and opt out; it was paid for by a 501(c)(4) advocacy group, and its QR code led to a Nasdaq page explaining Nasdaq's own decision to opt out — a use Nasdaq said it had not authorised (CoinDesk, 28 October 2026). The objection count went from sixteen to sixty-seven inside the window. The library's standing question, who benefits from the gap, has an unusually clean answer here: the incumbent exchange that already holds the issuer relationship, and whose own tokenised trading runs on the same CUSIP through DTC.
It also explains why Asia looks orderly and the United States does not. Where tokenisation is a funding tool, the issuer is the client and the rest of the structure follows from what the client wants. Where it is re-plumbing, the issuer has to be persuaded. Where it is third-party, the issuer has to be ignored, or, since September 2026, notified and then overruled by its own objection.
Two cases sit on the boundary. One is the tokenised secondary offering by an already-listed company: new funding of an existing security, the transaction that would turn re-plumbing into money. It has not happened. The other has, or is about to. On 29 October, Canton Strategic Holdings, a Nasdaq-listed company with a Canton treasury, agreed to put its own common stock on TSV LLC's permissioned venue through Fairmint, an SEC-registered on-chain transfer agent, with trading intended from 2 November. Because the issuer sponsors the token, the order's thirty-day notice does not apply. It is the first listed US equity tokenised under the Innovation Exemption, it is issuer-controlled, and it is thin by design. Sixty-seven companies said no to the third group in the same week that one company said yes to the second.
The second question is who the law says is offering, to whom, and through what channel. This factor explains the geography of the map better than any other. Three attributes belong together, and they are three, not one:
Four things fall out of this factor.
It explains why the third-party group exists at all. A wrapper is the one object that can carry its own prospectus across borders to retail without the issuer: the note travels, the share does not. It does not repeal local marketing, custody or distribution law; it separates the instrument that is offered from the security that is held.
It explains why one firm runs two products with different legal claims. Ondo and Dinari each offer an onshore product built on a transfer agent and an offshore note. The split is not a preference about holding models. It follows from offering regime: the onshore product can only be sold to US investors through US-registered entities, and the offshore product can only be sold to non-US investors as something other than the share.
It explains Hong Kong's distinctiveness. The SFC authorised tokenised funds for retail and then permitted their secondary trading on licensed virtual-asset trading platforms. Hong Kong is, on this dataset, the only jurisdiction where a sponsored, publicly offered retail product trades on a crypto exchange with the regulator's authorisation.
It explains why the Innovation Exemption bit, and where it does not. A Tokenized Securities Venue is a new channel for an existing object, and the issuer's right to object is a condition of that exemption, attached to the channel, not a general veto over tokenisation. The right only exists where the token was made by an unaffiliated party; an issuer-sponsored token on the same venue needs no notice at all. The two routes into a TSV are therefore not two variants of one product but two different legal objects, and the second is the one the objectors would have to be persuaded, not notified, into.
Offering type overlaps with consent, and the boundary needs stating. Offering type is who the law says is offering; consent is whether the issuer of the underlying agreed. They coincide in the funding group and diverge below it.
The third question is who keeps the register, and how far the token holder sits from it. Direct versus intermediated, CSD versus depositary versus transfer agent, are the same question asked of different institutions. The answer places every initiative on a ladder. The ladder orders one thing only — where the authoritative record sits; what the holder legally owns is the next factor, and is read off separately.
Two properties of the ladder matter more than the rungs.
First, rights tend to decay downward and reach to improve downward. L1 and L2-native are, as a rule, the rungs with a constitutive transfer, holder-protective insolvency treatment, and the chain as the controlling record — the governing law decides the detail in each case. L4 is where the round-the-clock, stablecoin-settled, cross-border retail trading actually happens. No rung is strong on every axis.
Second, L4 is a different object, not a lower rung. A note referencing a share is not a weaker way of holding the share. It is a claim on someone else. Diagrams that stack DTCC, a broker entitlement and a Robinhood note on one vertical make them look like points on a scale. They are not.
Switzerland ran this experiment first. The 2021 DLT law created the Registerwertrecht, a security that exists natively on a ledger with no custodian chain. Five years on, native volume is marginal and almost everything still settles as Bucheffekten through SIX SIS. The more elegant legal form did not win; the custody network did. DTCC's mirror is a bet that the United States repeats the Swiss outcome. The issuer veto is the one variable Switzerland did not have.
The fourth question is what the holder actually owns: the security, an entitlement to it, a note referencing it, or a derivative on its price. The asset class does not matter. A tokenised bond, a tokenised share and a tokenised fund unit can each be any of the four.
The claim type is where the Inheritance Test does its work, and the inheritance differs by claim. A share, a bond and a fund unit are each a direct claim on the issuer, but of different kinds — residual, contractual, pro rata — and nothing else stands between the holder and the issuer's fate. An entitlement adds the intermediary: custody, segregation, the shortfall that Dole made visible. A note adds the note issuer's credit, with the underlying held as collateral that is only as good as its ring-fencing, perfection and enforcement. A derivative adds the counterparty or the clearing house. Each step adds a party whose failure the holder is exposed to; what the holder should ask is whether the claim survives the failure of the entities that operate the structure.
The live example of a claim-type switch is ICE. Its venture with OKX planned to trade third-party tokens, which are notes or entitlements, and ran into the issuer objection. In the same week, ICE agreed to bring trade[XYZ]'s perpetual markets onshore, with Hyperliquid as the infrastructure, ICE as the regulated market and clearing operator, and the OKX venture named as one of the distribution channels. Five days later the CFTC's chairman said that perpetual futures on individual equities may trade on CFTC-registered exchanges that are also notice-registered with the SEC for security futures. He named no venue, no stock and no date; the product class is open, and the ICE venue still needs its own approvals (CFTC, 3 November 2026).
A perpetual future is not a tokenised share, and the point is sharper for that. There are two competitions here, not one: a competition to represent, hold and distribute the security, and a competition to replicate exposure to its price. The issuer's veto runs in the first. It has no standing in the second — neither the issuer nor the register is asked when a derivative is listed. So where issuers resist third-party tokenisation of their shares, the market has an alternative that confers no shareholder rights and needs no one's permission, and the second competition grows exactly where the first is constrained. The holder's exposure is then to a clearing house, which is arguably a better claim for the holder and a worse outcome for the issuer, since price discovery on weekends happens anyway, with no shareholder rights attached.
The claim type also separates the two Asian clusters from the American one. Hong Kong and Singapore tokenise securities. The United States, outside funds, mostly tokenises entitlements and notes.
Two factors do not describe an initiative. They decide whether it can scale: what settles the cash leg, and whether the token counts as collateral. Three propositions are easily run together and should be kept apart — that the cash leg can technically settle, that it discharges the obligation with legal finality, and that a particular central bank, clearing house or lender will take the token as collateral. No initiative on the map is strong on all three, and the regional split is the reverse of what the headlines suggest.
Europe has the rails and not the asset. The Eurosystem accepts tokenised securities as collateral from the end of March 2026, provided they are issued through a CSD and settle via TARGET; the EIB's commercial paper on Clearstream's D7 was posted as ECB collateral in June. Eurex Clearing has BaFin approval for DLT-based margin collateral. Pontes, the Eurosystem's settlement of DLT transactions in central bank money, launched on 21 September 2026 with an initial group of participants and full implementation expected by 2028, again through CSDs (ECB, 21 September 2026); in late October the European Stability Mechanism priced a €1 billion digital benchmark on D7 with the cash leg settling in T2 over Pontes, the first supranational to do so at benchmark size. The Swiss National Bank's Project Helvetia gives SDX wholesale central bank money through at least mid-2027, and has been extended to synchronised settlement over the existing SIC system (SNB, 30 June 2025). All of it is keyed to "issued via a CSD", and European CSDs have tokenised bonds and commercial paper. There is no tokenised listed European equity at L1 or L2 to put on these rails yet; Clearstream's e-shares are a second-half 2026 plan.
The United States has the asset and neither rail. DTC launched with 34 securities and the cash leg unspecified. There is no Fed-level eligibility for tokenised collateral, and no clearing house yet accepts DTC tokens as eligible margin under its rulebook, though the Citadel Securities production trade in July posted a DTC-tokenised Treasury to CME. The cash column is a competition between bank rails that do not interoperate: JPMorgan's Kinexys deposit token, Citi Token Services, HSBC's tokenised deposit service, BNY's tokenised deposits since January 2026, and a bank-consortium network targeted for 2027. Stablecoins settle the third-party group offshore. A deposit token is a liability of the issuing bank; it is not central bank money.
Hong Kong is the only place with both live in one pilot. Project Ensemble moved tokenised deposits to real-value transactions in November 2025, and one digital bond has settled with e-HKD and e-CNY as tokenised central bank money.
No token on a public chain is eligible as central bank collateral or clearing-house margin anywhere. Every rail above is permissioned, every eligibility rule names a CSD or a bank. Public-chain representation and institutional eligibility are separate questions, and so far the answer to the second has been no. The ECB has said it will consider extending collateral eligibility to tokens issued entirely on DLT outside CSDs. Until it does, the enabling columns are green for L2 and grey for everything above and below it.
This is the structural point about the rails. They were built so that only CSD-issued tokens qualify, which entrenches L2 and strands native issuance. Europe has not solved the two catalysts for tokenisation. It has solved them for CSDs.
Applied to roughly forty initiatives, the factors produce four clusters and three empty zones. The picture below places each initiative by holding rung and by the role the institution plays in that initiative. Status is marked where an initiative is planned or in pilot rather than live. The set is curated, not a census: notable initiatives as of 3 November 2026, chosen to cover every rung and region, and the clusters describe the set rather than the market.
Four clusters.
Three empty zones.
Confusion exists exactly where tokenisation happens without the issuer. Asia and continental Europe are legible because nearly every initiative there is a funding instrument with a sponsor, a CSD and a regulator-chosen cash leg. The United States and the offshore market are illegible because most of what trades there is re-plumbing or third-party product, built around an issuer who was not asked, through channels the law has not finished defining.
The second thing the sort shows is that liquidity flows to the thinnest legal promise. The crowded cells on the map are the L4 wrappers and perpetuals, where the holder owns a note or a contract and not the share. The strongest rights sit in the emptiest cells. Reach accrues where the claim is weakest, which is the pattern every wrapper market has shown before this one.
The third is that the incumbents are keeping the chokepoint, function by function. DTCC retains settlement and converts it into a fee service. The ICSDs keep collateral eligibility keyed to themselves. Banks keep the cash leg on their own ledgers. The NYSE, facing a veto it cannot lift, signed the people who keep the register instead. Risk and distribution migrate to whichever balance sheet carries the cheapest claim, and the chokepoint stays where it was.
Prognosis, twelve to thirty-six months.
Falsifier. If by the end of 2027 a major clearing house accepts DTC tokens as eligible margin collateral under its rulebook and no top-twenty S&P 500 issuer has sponsored a tokenisation of its own shares, then consent was never the hinge. Collateral eligibility was, and this essay's second question should be dropped.
Securitize's page said that Wall Street has rules and that tokenisation done right plays by them. It did not say whose rules. The open question in every market on this map is whether the register or the clearing house writes them, and the sort above suggests the answer differs by factor: the register wins where the token funds, the clearing house wins where the token re-plumbs, and nobody wins where the token is somebody else's product.
This week supplied the sharpest version of the answer. Sixty-seven issuers used their veto. The exchange that lost the veto fight signed the registrars. And the derivatives regulator opened a product on which the veto does not run. The market did not resolve the consent question. It built an instrument for which the question is not asked.
A reader who takes nothing else from this essay can keep the two questions. Who holds the register? Did the issuer agree? Everything else — the offering regime, the claim type, the cash leg — follows from the answers, and the next announcement can be placed in a minute.
Written in a personal capacity. Analytical views only — not legal or investment advice.
Julian Gretzinger — Investor and writer on monetary history, real wealth mechanics, and financial markets. substack.com/@juliangretzinger