Tokenization’s Last Mile Isn’t On-Chain — It’s in the Risk Ledger

On October 28, DTCC’s subsidiary DTC launched the DTCC Tokenization Service (DTS), minting “digital twins” of securities custodied at DTC: 34 securities are tokenizable at launch and 20 institutions have signed on. The legal foundation is the no-action letter (NAL) the SEC issued on December 11, 2025 — a narrowly scoped, three-year pilot authorization covering Russell 1000 constituents, major index ETFs, and U.S. Treasuries[1][2].

Some headlines proclaimed that “Wall Street’s $30 trillion just went on-chain.” But the authorized scope is not the tokenized volume — and what really matters is the segment of this pipeline that remains unsolved.

Five days after launch, two banks made two very different moves: Société Générale accepted $10 million of DTCC-tokenized U.S. Treasuries as eligible collateral; Morgan Stanley confirmed on its official account that it is “among the first banks to use DTS” — yet its own tokenized equity trading venue, Trajectory Cross, was pushed from H2 2026 to Q1 2027, citing “limited product availability and thin institutional demand.”

Collateral is moving on-chain while the trading side slips. That contrast is what this piece explains.

1. Break DTS into Three Layers

The most common mistake in discussing “tokenization” is conflating different layers. DTCC’s service splits into three, with wildly different maturity levels.

Layer 1: Registry — solved, and solved in a very “un-Web3” way.

What DTS mints is a “tokenized entitlement”: 1:1 with the traditional security custodied at DTC, with ownership rights, investor protections, and economic entitlements identical to the conventional version, sharing the same CUSIP. Registered ownership always stays with DTC’s nominee (Cede & Co), and the underlying securities cannot leave the “Digital Omnibus Account” until the corresponding tokens are destroyed — the mechanism that prevents double-spending[1][3].

Most notable is how the books are kept: on-chain transfers are tracked by an off-chain DTC system called LedgerScan — it runs on public cloud, scans the underlying blockchains, and records token movements and registered-wallet positions in near real time. LedgerScan’s records constitute DTC’s official books and records[1].

In other words: the chain is not the ledger; the off-chain scan of the chain is. The “risk ledger” in our title is not a metaphor — it is a mechanism written into the authorization itself.

Layer 2: Transfer — partially solved, and more constrained than appreciated.

On supported chains, between DTC-registered wallets, a token transfer is itself settlement — it can happen 24/7 without waiting for T+1. These peer-to-peer transfers bypass DTCC’s post-trade infrastructure, but every wallet that touches a token must be on the allowlist[3][5]. There are also three hard constraints:

1. Transfers are only possible between DTC-registered wallets — participants register designated third-party wallets for their clients through an interface; arbitrary self-custodied addresses are not allowed;
2. Cross-chain moves and conversions between “conventional form ↔ tokenized form” go through DTC’s conversion service, open only 02:00–18:00 (U.S. Eastern) on business days (per Ledger Insights’ description of the DTS conversion service);
3. Only two chains are connected so far — Canton Network and DTCC’s own private Besu network; others (including Stellar and Circle’s Arc) are expected to join through 2027[6].

So “24/7” currently holds only for peer-to-peer transfers between registered wallets on the same chain — not “move in and out of tokenized form anytime, anywhere.”

Layer 3: Recognition — essentially untouched, and the most expensive layer.

The SEC’s no-action letter contains a subheading literally titled “No Collateral or Settlement Value”: DTC will not assign collateral or settlement value to any Tokenized Entitlements for the purpose of computing participants’ Net Debit Caps and Collateral Monitors — i.e., they are not used in DTC’s own risk management[1].

Which means: the token is minted and transferable, but in DTC’s own capital and collateral calculations it is neither “collateral” nor “money.” So who recognizes it? Banks recognize it themselves. On November 2, Société Générale’s Prime Services & Clearing division announced that, in a controlled transaction, it accepted $10 million of DTCC-tokenized U.S. Treasuries as eligible collateral — explicitly within the bank’s own governance, risk, and control framework, and already reflected in its existing back-office and collateral systems[7][8].

This was not a first out of nowhere: SocGen joined Canton as a Super Validator in May, and in July’s first live phase it had already pledged tokenized Treasuries to Citadel Securities for trading exposure. What changed on November 2 — for the first time, it recognized the token as collateral in the role of the accepting party[5][7].

SocGen’s global head of prime services, Salim Nemouchi, put it in words that are almost this article’s thesis: “This transaction moves tokenization beyond issuance and begins to connect tokenized securities with the financing markets where clients can derive tangible value.”

That is the core of the whole piece: tokenization’s scarce resource has shifted from “can it be minted” to “who recognizes it.”

2. Why Value Lands on Collateral First, Not Trading Volume

With the three layers side by side, where value flows before 2027 is not hard to infer.

The minting layer is becoming a public good. As market infrastructure, DTCC provides “digital twins” as a service to participants; its business model is not seigniorage. Everyone can get twins — differentiation doesn’t happen at this step.

The trading layer is demand-constrained, not technology-constrained. Morgan Stanley pushed Trajectory Cross to Q1 2027, publicly citing “limited product availability and thin institutional demand” — a demand-side thermometer, not a technology problem. It is simultaneously building its own wallet infrastructure, and its head of digital asset strategy Amy Oldenburg frames the long-term tokenized asset market at $3–10 trillion by 2030[9][10].

The collateral layer doesn’t need “trading volume” — it only needs “speed.” Margin movements, repo collateral, intraday liquidity management — none of these require continuous 24/7 matching; they only require “faster than today’s T+1 and business-day windows.” DTS’s 24/7 same-chain transfers slot exactly into that gap. SocGen’s $10 million is small, but it is a symbol — and SocGen has stated it will develop “repeatable collateral mobility” and, subject to approval, explore on-chain repo financing[7].

In other words: this round of tokenization is a “collateral mobility” story, not a “retail trading volume” story. Conflating the two systematically overestimates trading-type revenue in 2027 while systematically underestimating collateral and treasury-management revenue.

3. The Part the U.S. Left Blank, Hong Kong Built Differently

The U.S. design leaves the “money side” blank: in the DTS roadmap, on-chain cash netting remains a 2027-and-beyond plan, and stablecoins and tokenized deposits for settlement are still at the “exploration” stage. In the conventional world, NSCC’s netting compresses routine securities settlement amounts by roughly 98% — to replicate that efficiency on-chain, the “on-chain money” layer has to be built first[6].

On November 2, Hong Kong took a different path: the HKMA announced that Ensemble^TX is entering a new phase, supporting 24/7 interbank settlement of tokenized deposit transactions in tokenised central bank money (tokenised CeBM), with participating banks expanding to 12 and industry pioneers to 20. HKMA Chief Executive Eddie Yue characterized it as “moving within two years from experimentation and pilots to a new phase of driving the commercialization of tokenization applications”[11][12].

The same day, the SFC welcomed the expansion of collateral eligibility by two HKEX derivatives clearing houses: HKFE Clearing Corporation (HKCC) and The SEHK Options Clearing House (SEOCH) began accepting China Government Bonds (CGBs) and Chinese policy bank bonds as eligible non-cash margin collateral — including onshore CGBs and policy bank bonds held under Northbound Bond Connect, as well as offshore CGBs — while cutting the accommodation charge on non-cash collateral to 0.15% p.a., effective that same day, November 2, 2026. Rico Leung, the SFC’s Executive Director of Market Supervision, said this not only lowers funding costs for market participants but also expands offshore use cases for these bonds, and together with the fee cut can unlock capital efficiency and make RMB assets more attractive to overseas institutional investors[13].

This locks directly into our thesis: the economic value of collateral mobility ultimately shows up as basis points in funding costs.

Viewed side by side, the division of labor between the two jurisdictions is clear:

These are not mutually exclusive routes — they are complementary halves. The U.S. turned “how assets go on-chain” into infrastructure; Hong Kong turned “how to settle in central bank money” into infrastructure. For cross-border issuers and licensed platforms, the real opportunity sits at the seam: assets in the U.S., money and distribution in Asia.

This table compares structural differences in market infrastructure capabilities only; it is not an assessment of any regulatory regime’s merits.

4. Five Overlooked Risks

1. Two markets for the same CUSIP. The digital twin and the conventional form share the same code and rights and are theoretically interchangeable; but outside the conversion windows, their liquidity, spreads, and borrow availability will diverge. As one market commentator put it, this creates “two stock markets for the same company” — that is not rhetoric, it is structural friction.

2. The gap between “24/7” marketing and reality. As noted above, same-chain transfers are 24/7, but cross-chain moves and form conversions run only 16 hours on business days[5]. Any description of DTS as “round-the-clock settlement” needs this footnote.

3. The demand-side cold start. Only 34 securities are tokenizable at launch (DTCC’s service page states “34 securities available to tokenize at launch”; Ledger Insights uses the same figure). The named participants in July’s first live phase were five — JPMorgan, CME, Société Générale, Citadel Securities, and BNP Paribas — while 20 participants have formally signed up, notably without BNY (the largest custodian) and Goldman Sachs (the latter took part in the July pilot). Industry working-group membership grew from 50 in May to over 150 in October, but “joining the discussion” and “signing up to use it” are two different things[3][6].

4. Regulators’ “slow down.” The IMF’s Tobias Adrian argued earlier that atomic settlement, 24/7 markets, and smart contracts could accelerate the transmission of liquidity stress and market shocks — faster than regulators can respond. In his view, the most disruptive change happens inside regulated institutions, not on DeFi rails[14]. And the authorization itself has a fuse: the no-action letter automatically expires three years after DTC launches the Preliminary Base Version — counting from October 28, roughly October 2029[1].

5. The turf war between issuer-sponsored and third-party wrapped tokens. The industry body representing transfer agents has lobbied the SEC to prioritize issuer-sponsored tokenized shares over third-party stock tokens, arguing that synthetic and third-party models can blur investor rights and introduce platform and custody risks — precisely what the official shareholder register exists to avoid[14]. Where this lobbying battle lands still comes back to the opening question of this piece: who gets to define the “official register.”

5. EX.IO Research Views

First, to judge a tokenization project’s value, ask which layer it sits on. Minting is becoming a public good; transfer is fragmented by business-day windows and wallet allowlists; the genuinely scarce layer is recognition — whose risk framework accepts it, whose custody can hold it, whose settlement can cash it out. Projects that only say “we’re on-chain” are talking about layer one; those that can explain “who recognizes it, how it monetizes, how defaults are handled” are playing at layer three.

Second, collateral is the first real paying use case of this cycle. Margin and repo collateral don’t need retail flow — they need time efficiency and auditable title. That also explains why the first meaningful bank-level move came from a prime brokerage and clearing desk, not a trading desk.

Third, the money side is scarcer than the asset side. The U.S. built a $30 trillion asset-side authorization, but on-chain money is still under exploration; Hong Kong built 24/7 settlement in tokenised central bank money first, while asset-side scale is still accumulating. Whoever connects the two at the seam takes the pricing power in cross-border collateral and distribution.

Fourth, “on-chain recording” is not “the on-chain ledger.” DTC generates its official books through the off-chain chain-scanning system LedgerScan — proof that within the regulated system, auditable record-keeping authority remains on the infrastructure side. Any “the chain is the truth” narrative still needs an off-chain official record in front of courts, regulators, and risk systems.

Fifth, don’t mistake “launch” for “scale.” 34 securities, 20 signed institutions, a 16-hour daily conversion window — these three numbers define real capacity before 2027.

6. Six Signals Worth Tracking

1. DTC Collateral AppChain (expected early 2027): whether tokenized collateral gets plugged into DTC’s own risk and netting systems — the watershed between “bank-level recognition” and “infrastructure-level value.”
2. Whether Asian banks accept DTCC-tokenized Treasuries as collateral: if licensed banks in Hong Kong or the HKEX system start recognizing this token, Asia’s recognition layer truly opens.
3. Whether cross-chain and conversion windows lengthen: whether the 16-hour business-day limit loosens once networks like Stellar and Circle Arc connect.
4. EnsembleTX participation expansion and stablecoin settlement trials: the timing of growth beyond 12 banks and of regulated stablecoins entering settlement experiments.
5. Whether Morgan Stanley’s Trajectory Cross slips again: the most direct demand-side thermometer — more reliable than any “institutions are optimistic” statement.
6. The no-action letter’s three-year clock (~October 2029): whether — and on what terms — regulators renew it will decide whether this architecture is a transitional experiment or a permanent foundation.

References

1.        SEC, DTC No-Action Letter (2025-12-11; includes “No Collateral or Settlement Value,” LedgerScan, and the three-year sunset clause) — https://www.sec.gov/files/tm/no-action/dtc-nal-121125.pdf

2.        DTCC, “Paving the Way to Tokenized DTC-Custodied Assets” — https://www.dtcc.com/press-releases/2025/paving-the-way-to-tokenized-dtc-custodied-assets

3.        DTCC, Tokenization service page — https://www.dtcc.com/products-and-services/digital-assets/tokenization

4.        DTCC, “DTCC Tokenization Service Goes Live,” 2026-10-28 — https://www.dtcc.com/press-releases/2026/dtcc-tokenization-service-goes-live

5.        Ledger Insights, “DTCC processes first live tokenization transactions with JP Morgan, CME, BNP Paribas” — https://www.ledgerinsights.com/dtcc-processes-first-live-tokenization-transactions-with-jp-morgan-cme-bnp-paribas/

6.        Ledger Insights, “Alpaca, Citi, JP Morgan among early adopters as DTCC Tokenization Service goes live,” 2026-10-28 — https://www.ledgerinsights.com/alpaca-citi-jp-morgan-among-early-adopters-as-dtcc-tokenization-service-goes-live/

7.        Société Générale, “Societe Generale Completes Tokenized US Treasury Collateral Transaction on Canton,” 2026-11-02 — https://wholesale.banking.societegenerale.com/fileadmin/user_upload/groupe_AMER/pdf/Societe_Generale_Completes_Tokenized_US_Treasury_Collateral_Transaction_on_Canton.pdf

8.        CoinMarketCap Academy, “Societe Generale tokenized Treasuries collateral on Canton” — https://coinmarketcap.com/academy/article/societe-generale-tokenized-treasuries-collateral-canton

9.        CoinMarketCap Academy, “Morgan Stanley DTCC tokenization ATS delay 2027” — https://coinmarketcap.com/academy/article/morgan-stanley-dtcc-tokenization-ats-delay-2027

10.     Morgan Stanley official post (X) — https://x.com/MorganStanley/status/2117727271377125454

11.     HKMA, “HKMA advances Project Ensemble pilot phase to 24/7 settlement in tokenised Central Bank Money,” 2026-11-02 — https://www.hkma.gov.hk/eng/news-and-media/press-releases/2026/11/20261102-5/

12.     Ledger Insights, “Hong Kong’s EnsembleTX pilot now supports 24/7 tokenized central bank money” — https://www.ledgerinsights.com/hong-kongs-ensembletx-pilot-now-supports-24-7-tokenized-central-bank-money/

13.     SFC, “SFC welcomes expanded use of China Government Bonds as margin collateral by HKEX” (26PR180), 2026-11-02 — https://apps.sfc.hk/edistributionWeb/gateway/EN/news-and-announcements/news/doc?refNo=26PR180

14.     Blockhead, “DTCC completes its tokenization pilot — the IMF wants everyone to slow down,” 2026-08-13 — https://www.blockhead.co/2026/08/13/dtcc-completes-its-tokenization-pilot-the-imf-wants-everyone-to-slow-down/

Disclaimer

This article was prepared by EX.IO Research for informational and educational purposes only. It does not constitute any investment advice, offer, solicitation, or recommendation of any particular virtual asset, security, or investment product. It does not constitute a solicitation or offer of services to residents of any jurisdiction (including mainland China). Data herein are drawn from third-party public sources and are stated as of the dates indicated; regulatory frameworks and market infrastructure rules may change — please refer to the latest publications of the SFC, HKMA, SEC, and other relevant regulators. This article does not constitute research coverage of, or an investment rating on, any listed issuer; neither the author nor the firm has received compensation from any institution mentioned. Readers should assess the relevant risks independently and consult independent professional advisers where necessary.

On the “~$30 trillion” figure: this number is a third-party estimate (Ledger Insights[6]) of the asset scale corresponding to the scope of the SEC no-action letter — it is neither the tokenized volume nor DTC’s custody volume. DTCC’s press release cites over $114 trillion in custody, while its service page carries a separate ~$30 trillion figure — the two are inconsistent, and this article does not adopt the latter.