Brawlers Guide
Contents

DTCC Tokenisation Turns the CSD Into the Rail

DTCC tokenisation turns the CSD into the rail. July production trades. October commercial service. Same CUSIP, two forms.

Heath Donald Tokenisation @heathdonald /post/dtcc-tokenisation-csd-rail

DTCC tokenisation is not a wrapper looking for a chain. It is the depository learning to mint a twin of securities it already holds. That change is leaving the lab.

On 15 July, DTCC converted securities sitting at The Depository Trust Company into tokens and used them in live production trades. More than 30 firms ran collateral pledges, securities lending, Treasury and repo delivery-versus-payment, equity DVP and DVD, token transfers and CCP margin. The conversions sat on DTCC's private Besu network and on Canton. The commercial service is scheduled for October. I wrote in August that settlement, custody and legal finality were the bottleneck. The depository just answered by becoming the rail.

This is a CSD extending itself, not a minting booth with a press release. DTC already custodies more than $114 trillion. The SEC's December 2025 no-action letter lets DTC offer a defined tokenisation service for three years, covering Russell 1000 names, major-index ETFs, and US Treasury bills, bonds and notes. The tokens are digital twins of assets already in the box. Same CUSIP. Same entitlements. Convertible back into the traditional form. If your token cannot be recognised by the place that already holds the security, you do not have an instrument. You have a brochure.

The depository is minting the twin

Issuance was never the hard part. Anyone with a legal opinion can wrap a fund. What DTCC did in July is different. It tokenised DTC-held securities, delivered them into participant wallets, and then ran the ugly workflows that keep a market open: pledge, lend, repo, margin.

Frank La Salla called it institutional rigor. Brian Steele was more useful. DTC-tokenised assets keep the same investor protections, entitlements and ownership rights as the traditional form. That is the product. Programmability is a feature. Legal continuity is the sale.

The industry working group was 50-plus firms in May and over 100 by July. The roster is not a crypto conference. BlackRock, J.P. Morgan, Goldman, Citadel, Vanguard, Nasdaq, NYSE, State Street, Circle, Ondo, Fireblocks, Virtu. When that room shows up for a production test, they are not buying a narrative. They are checking whether the twin will clear.

Nadine Chakar said after the July run that the safest path runs through trusted financial market infrastructure. I agree, with one caveat. Trusted here means a CSD a court already understands. It does not mean the chain with the loudest conference circuit.

Same CUSIP, two forms

DTCC's own description is blunt. Traditional form and tokenised form share a CUSIP, so the asset can move between the existing market and a pre-approved chain without becoming a different security. Tokens can flow between registered participant wallets on those chains, including outside banking hours. ComposerX LedgerScan is meant to keep a near-real-time view across traditional and blockchain ledgers. The smart contracts include mint, burn, force transfer, clawback, pause and freeze.

That last list will offend people who think tokenisation means censorship resistance. They should get over it. A Russell 1000 stock that cannot be frozen after a court order is not going through DTC. The design choice is the point: the token is the security, not a derivative that tracks it. A lot of the retail "tokenised stock" market is still a third-party instrument, sometimes backed 1:1, sometimes a derivative, often without the same rights. Do not confuse that pile with a DTC twin. They are different products wearing the same word.

The Fed's Lisa Cook put the US tokenised-asset stock around $25 billion in May, more than double the year before, with government bond funds the largest slice. BeInCrypto's July cut of the wider real-world asset market was about $60 billion across more than 7,000 products. Tokenised Treasuries were the only class that looked production-grade: about $15 billion across 100 assets, 99 percent on public chains. Figure's HELOC book was $18.3 billion on its own, and 97 percent of tokenised RWA value sat outside US retail. The on-chain number is still a rounding error against $114 trillion in the box. October is not everything on-chain. It is the box learning to speak token without leaving the box.

Collateral that can move at 2am

The economic argument is not 24/7 meme stocks. It is trapped collateral.

Forkast, writing at the end of August as the October launch came into view, cited DTCC figures of $300 trillion in global high-quality liquid assets with only 10 to 11 percent used as collateral. Digital Asset, which built Canton, has been telling the same room that tokenised workflows could lift balance-sheet efficiency 30 to 50 percent. Treat that second number as a vendor estimate, not a law of physics. The direction is still right. If a Treasury can be pledged, recalled and re-pledged without waiting for a settlement cycle, the inventory sitting idle in a custodian account starts to look expensive.

That is the same shape as the power problem. In AI infrastructure, the scarce asset is the interconnection queue, not the chip: https://brawlersguide.com/post/ai-infrastructure-power-procurement-bottleneck. Here the scarce asset is a CSD file that a court, a CCP and a custodian will all recognise at 2am. A wrapper without that file is inventory.

BlackRock is already selling the cash leg. It put a tokenised share class of the Select Treasury Based Liquidity Fund on Ethereum, with BNY Mellon as transfer agent, and launched a Daily Reinvestment Stablecoin Reserve Vehicle through Securitize aimed at GENIUS Act eligible reserves. BUIDL, the first tokenised money market, was around $2.5 billion. Cash that can sit next to a DTC twin is how delivery-versus-payment stops being a slide.

I argued in August that tokenisation's real bottleneck is settlement: https://brawlersguide.com/post/tokenisation-real-bottleneck-settlement. DTCC did not refute that. It absorbed it. The settlement file is now a service the depository intends to sell.

What operators should buy

If you are raising, allocating or building against this, stop underwriting the chain. Underwrite the twin.

A path into a DTC participant wallet beats a pretty token standard. If the thing cannot convert back into the traditional form under the same CUSIP, you are holding basis risk and calling it innovation. The cash token, or the RTGS window that can hit at the same instant as the security token, is the other half. Atomic DVP on a private Besu instance still needs a cash leg. A Treasury twin that can only move when the banking cut-off is open is a part-time instrument.

The control set matters more than the branding. Mint, burn, freeze, clawback. Ugly, and the reason a CCP will take the thing as margin. If your design cannot survive a court order, it will not survive October. Legal continuity has to be bought years ahead of the mint. The no-action letter is three years and a defined asset set. That is a runway, not a permanent statute. The projects that trade on day one will be the ones whose true-sale opinions and transfer-agency agreements were signed while this was still a working group.

Do not finance the wrapper. Lenders will underwrite the depository, the custodian and the cash leg. A tokenisation shop that cannot show a DTC path, or an equivalent CSD path in another market, is still a brochure with a contract address.

October is not the finish

The October launch is an entry point. Integration into risk systems, accounting and the collateral schedule is the work. DTCC is also pointing at a fourth-quarter go-live for its Collateral AppChain. Eligible networks include the private Besu instance and Canton, with Stellar on the public list. Multi-chain here is not ideology. It is a refusal to strand $114 trillion on one vendor's ledger.

I do not think this kills public-chain issuance. BlackRock will keep putting fund shares on Ethereum because that is where the digital-asset cash already sits. What changes is the idea that you can outflank the CSD. The depository is extending itself. Nobody is replacing it. The firms that win the next two years will hold a twin the registrar already understands, a cash leg that moves with it, and a freeze function a regulator can live with.

The issuance deck is still table stakes. The CSD file is the fight.

If you want the next note when October clears, get on the weekly digest. One post, straight to the inbox, or take the RSS at brawlersguide.com/rss.xml and read it wherever you already read.

Also in this edition