On April 23, 2025, DTCC’s Great Collateral Experiment followed Margin Corp, a fictional institution, through a day in four financial centers. The staged demo used real on-chain transactions to show how firms could move and reuse collateral.
DTCC helps financial markets clear trades, settle transactions and service assets. The experiment asked a simple question: can a firm use assets it already owns to meet an obligation, wherever and whenever it arises?
Four cities. Four collateral problems.
Follow the four workflows from the experiment replay. Margin Corp, Bravo Bank and Sterling Bank are fictional participants. The savings figures are adjustable estimates, not results reported by DTCC.
The Time Zone Problem
SituationMargin Corp’s Tokyo team needs to meet a JSCC margin call while U.S. markets are closed. It mobilizes U.S. Treasury inventory from its U.S. broker-dealer branch.
Borrow cash overnight while owned collateral is out of reach.
Mobilize eligible tokenized Treasuries to meet the obligation, with 24/7 availability.
Adjust savings assumptions
$500m position and 1% incremental funding premium are assumptions. 250 comparable events per year is illustrative.
More from the demo
The underlying Treasuries stay in local safekeeping while tokens make them available for allocation. The same asset pool also supports the firm’s crypto margin obligations.
The presenter converts 190 million in Treasury quantity (27:49) and later allocates 39 million in USDC quantity (31:08). These are demo quantities, not a measured savings result.
The Multiple Account Problem
SituationMargin Corp has excess collateral against Société Générale at BNY and an €8.8m deficit against Wellington at Euroclear. It returns and reallocates the excess across those agreements.
Recall and repost collateral across separate providers; fund the deficit while waiting.
Reallocate excess from the Société Générale/BNY agreement to Wellington/Euroclear. Société Générale also posts EURCV against a €2m call.
Adjust savings assumptions
€100m, a 4% funding rate, 72-hour traditional delay, 15-minute on-chain residual delay and 75 events per year are assumptions; the replay shows an €8.8m deficit.
More from the demo
The demo also shows two-way margin calls, collateral substitution and dividend processing for pledged assets.
The €8.8 million deficit is stated at 37:28; the €2 million EURCV call at 41:17. Substitution and dividend processing are discussed in the scenario’s recap at 48:58.
The Crisis Speed Problem
SituationAfter a fictional bank failure shakes gilt markets, Wellington’s Evan Crosby offers ULTRA fund tokens to Margin Corp. The parties update eligibility; a separate Sterling Bank default is handled with Euroclear.
Demo institution
Coordinate agreement changes and potentially redeem fund shares for cash.
Counterparties agree new eligibility terms, then pledge tokenized money market funds through smart contracts.
Adjust savings assumptions
£200m, a 2.5% incremental stress premium, 14-day traditional delay and two events per year are assumptions. This models funding time, not default losses.
More from the demo
After the eligibility update, a separate Sterling Bank default triggers closeout calculations and an authenticated notice to Euroclear.
Money market fund eligibility is changed at 53:19; the default workflow begins at 56:20. The scenario does not establish a measured annual cost saving.
The Short-Term Loan Problem
SituationMargin Corp executes a four-hour USD repo with Bravo Bank using JGBs and Bitcoin, with Fnality digital cash. Separately, it moves excess collateral at FICC to cover a deficit at NSCC.
Demo institution
Use overnight repo or a credit line; return collateral before reposting it to another CCP.
Settle a four-hour repo with digital cash and unwind automatically. Separately, move eligible collateral directly from FICC to NSCC.
Adjust savings assumptions
Four-hour repo tenor is demonstrated. $300m, 5% annual funding rate, overnight alternative and 130 events per year are assumptions. CCP rebalancing is excluded.
More from the demo
The repo uses a 1.3 collateral ratio. Its automatic unwind returns the assets after four hours. The FICC-to-NSCC transfer is a separate workflow.
The four-hour repo starts at 1:03:37, the 1.3 ratio is stated at 1:04:29, and the automatic unwind at 1:06:21. The CCP transfer is a distinct workflow.
How the savings model works
Each card uses simple interest: position × annual rate × funding hours ÷ 8,760, plus fees. Savings are the difference between the two event costs; annual savings multiply that difference by the selected number of comparable events.
The rates, fees, delays and event counts are assumptions. Tokyo and London model extra funding premiums; Paris models funding a temporary shortfall; New York compares loan durations. Fees start at zero. The estimates exclude implementation, custody, conversion and risk costs, as well as the effect of collateral availability and haircuts. Each currency is calculated separately.
What sits behind the collateral?
The demo brings together Treasuries, Japanese government bonds, fund tokens and digital cash. Each works differently: a bond promises repayment, a share gives ownership rights, and a fund holds a portfolio of investments.
Explore 20 types of securities and financing arrangements below. Choose one to see what it does, the information it needs, and how it could be represented on Ethereum.
This broader reference extends beyond the assets shown in the demo. It is not a list of products supported by DTCC’s AppChain.
Common Stock
How it is structured
How it could work on Ethereum
See the data and terms
CUSIP and ISIN identify a security. LEI identifies a legal entity. The fields below depend on the instrument’s terms.
Read the field abbreviations
NAV: net asset value. WAC: weighted average coupon. WAM: weighted average maturity. WAL: weighted average life. MBS: mortgage-backed securities. ABS: asset-backed securities. CMO: collateralized mortgage obligation. CDO: collateralized debt obligation. ADR/GDR: depositary receipts. ETF: exchange-traded fund. UIT: unit investment trust. UMR: uncleared margin rules. CCP: central counterparty. A haircut discounts the asset’s value for collateral purposes.
How do these securities fit on Ethereum?
Sending a token is one part of the job. The receiving institution also needs to know what the holder owns, whether the transfer is allowed, and how payments or redemptions will reach them.
Start with the legal claim. A token could represent a security issued on-chain, a claim held through a custodian, or exposure created by a contract. The legal documents and ownership register determine which one it is.
What does the holder own?
Who can receive or move it?
How are interest, dividends and redemptions handled?
Can another system follow the same rules?
Where token standards help
ERC-3643
Combines an ERC-20-compatible token with identity and compliance checks. It supports restricted transfers, freezing and recovery, which can help an issuer control who holds a security.
Read the standard ↗ERC-7943
Gives applications a common way to check transfer permissions, frozen balances and forced transfers. Issuers can use different compliance systems behind those shared functions.
Read the standard ↗ERC-4626
Defines deposits, withdrawals and shares for a vault holding one ERC-20 asset. A securities fund with several assets, external valuations or delayed redemptions needs additional rules.
Read the standard ↗The harder part: keeping the same rights everywhere
A Treasury token needs its issue identifier and repayment terms. A fund token needs its NAV and redemption rules. A share needs its voting and dividend rights. These assets can use common transfer functions while keeping their own terms.
For institutions to use them across systems, they need shared identifiers, trusted valuations and agreed ways to process payments and corporate actions. Moving between chains must also preserve transfer restrictions and prevent the same claim from being counted twice. Private investor records need controlled access.
That is the industry’s interoperability challenge: making sure an asset keeps its meaning and rights wherever it moves.
These are possible design approaches. They do not describe DTCC’s choice of token standards.
The architecture, explained at EthCC
The Enterprise Ethereum Alliance invited DTCC to EthCC in 2025. On July 3, Zach Short, DTCC’s Head of Blockchain, presented DTCC Launchpad & The Great Collateral Experiment. The talk is the technical companion to these workflows.
April 2025 → Q4 2026
- April 23, 2025
Great Collateral Experiment - July 3, 2025
Technical presentation at EthCC - Q4 2026
Expected AppChain go-live
The AppChain announced in 2025 was built on Besu, an Ethereum-compatible blockchain client. In May 2026, DTCC announced integration of Chainlink’s Runtime Environment and data standard for pricing, valuation and automated collateral workflows, with production go-live expected in Q4 2026.
Sources: April 2025 event · DTCC retrospective · AppChain architecture · Q4 2026 production announcement · EthCC agenda