Project Agora 2026 shows why tokenised bank deposits could beat stablecoins for serious wholesale cross-border money. They keep the customer’s legal claim against a regulated bank, settle the interbank leg with central-bank money and can make every currency update conditional on the others. Stablecoins remain more portable on open networks, but portability is not the same thing as final settlement.
The story also changed in July. Project Agora no longer sits only in the prototype category. Twenty-eight financial institutions and central banks completed real-value transactions worth about CHF 800,000 across 17 scenarios. The average journey from payment initiation to settlement was roughly 80 seconds. Those are controlled-test numbers, not a production-service promise. Even so, real money has now crossed the architecture.
The most important result is not the speed. It is the legal design. A tokenised deposit in Agora is still a deposit liability of the issuing commercial bank. A tokenised reserve is still a liability of the issuing central bank. The code records and moves account balances; it does not quietly replace those relationships with a new cryptoasset. That distinction is the whole game.
What Project Agora 2026 proved with real money
The July 2026 exercise did more than replay fake balances. The issued tokens represented real monetary value. Participating central banks issued and redeemed tokenised reserves, while financial institutions earmarked funds or booked equivalent amounts in their core banking systems before issuing tokenised deposits. Settlement covered corporate and interbank payments, single- and dual-currency flows, payment-versus-payment transactions and intragroup transfers.
Project Agora 2026 also tested the less glamorous machinery that decides whether a payment system can survive outside a presentation. About 250 staff worked across payments, compliance, legal, operations and risk. Instructions used familiar ISO 20022 messages, including pacs.008, pacs.009 and camt.053. Participants had a contractual testing agreement, a runbook and fallback procedures. That is a better signal than a fast demo transaction because it begins to test the organisation around the code.
Still, keep the champagne cork in. Transactions ran during predefined intraday windows. Cross-currency providers manually confirmed some corresponding amounts. The prototype was not deeply integrated with every RTGS and core-banking system. So 80 seconds measures a prepared test corridor, not the time a multinational can expect for an ordinary invoice tomorrow morning.
| Test result | What it supports | What remains unproven |
|---|---|---|
| Real-value transactions settled | The architecture can represent and move genuine monetary value | Production scale, continuous availability and commercial pricing |
| Average settlement of about 80 seconds | Coordinated validation and atomic settlement can compress a prepared workflow | End-to-end speed after exceptions, investigations and full legacy integration |
| 17 scenarios | The model supports more than one narrow payment type | The long tail of real correspondent-banking currencies and edge cases |
| ISO 20022 interaction | A migration path can reuse established payment messages | Working integration across every bank, RTGS and sanctions stack |
Tokenisation changes the record, not the debtor
A tokenised bank deposit is not cash placed inside a digital wrapper. It remains a debtor-creditor relationship between a bank and its depositor. Project Agora 2026 treats the tokenised balance as equivalent to the balance in a conventional deposit account. The existing account agreement remains the source of the holder’s rights.
Here is the unusual bit: the token itself is not supposed to confer a separate legal or economic right. It is the technical record used to debit, credit and transfer the underlying account balance. If the bank fails, the customer does not point to a free-floating token and claim a mystery asset. The customer points to the bank deposit relationship, subject to the relevant insolvency, resolution and depositor-protection rules.
That is why the legal wrapper can be more important than the blockchain. The token can be technically perfect while an account agreement is silent about ledger records, data sharing, unauthorised instructions or outage liability. Agora’s report openly says some agreements and platform rulebooks may need updating. Code performs the obligation. Contract law gives the performance legal meaning.
Switzerland has been testing the same basic proposition from another angle. Our review of Swiss deposit-token infrastructure explains why banks want programmable money without abandoning the bank-account claim. Agora takes that logic into a multi-currency public-private architecture.
Four tokens can show the same currency symbol and mean four different claims
The label on a wallet is a poor guide to legal risk. A tokenised deposit, stablecoin, wholesale CBDC and tokenised money market fund can all display a value close to one dollar, euro or franc. Yet each answers the question “who owes me?” differently. For treasury decisions, that is the first comparison to make.
| Instrument | What the holder owns | Primary debtor or asset pool | Main protection | Main weakness |
|---|---|---|---|---|
| Tokenised bank deposit | A deposit claim recorded on a programmable ledger | Issuing commercial bank | Bank supervision, capital, liquidity, resolution and eligible deposit protection | Bank credit risk, permissioned access and issuer-by-issuer interoperability |
| Fiat stablecoin | A redemption or contractual claim defined by the issuer and local regime | Stablecoin issuer and its segregated reserve arrangement | Reserve quality, segregation, redemption rights and issuer regulation | No ordinary deposit insurance; redemption access, liquidity and secondary-market price can diverge |
| Wholesale CBDC or tokenised reserve | A direct claim on the central bank | Issuing central bank | Central-bank balance sheet and sovereign monetary framework | Restricted to eligible institutions and controlled infrastructure |
| Tokenised fund | A security or fund unit representing an interest in a portfolio | Fund vehicle and its underlying assets | Securities law, custody, valuation, diversification and fund governance | NAV, liquidity, gating and market risk; it is an investment claim, not final money |
Why Project Agora 2026 gives bank deposits an institutional edge
Tokenised bank deposits start with an advantage that stablecoins must manufacture: they are already inside the banking system. The issuing bank has capital and liquidity obligations. It already knows the customer. It already maintains deposit records, processes sanctions controls and connects to central-bank settlement. Tokenisation can make that bank liability programmable without moving the customer’s claim into a separate reserve-backed issuer.
The second advantage is the singleness of money. A corporate treasurer does not want Bank A dollars trading at 99.7 cents against Bank B dollars during stress. Agora’s design coordinates deposit balances with tokenised central-bank reserves so the interbank leg retains a common settlement anchor. Stablecoins can maintain par through high-quality reserves and reliable redemption. However, each issuer still operates its own conversion mechanism, liquidity policy and distribution network.
Third, bank deposits bring balance-sheet elasticity. Banks can create deposits through lending and can draw on central-bank liquidity against eligible collateral. A fully reserved stablecoin generally needs cash or approved reserve assets before new units appear. That restraint can be a virtue for holders, but it is not the same machinery that supports flexible credit and payment liquidity across the economy.
Finally, tokenised deposits suit known-party workflows. A multinational paying a regulated supplier does not necessarily need a bearer token that can travel to any wallet. It needs the correct beneficiary, completed compliance, guaranteed receipt or cancellation and a defensible audit trail. Agora makes those conditions part of the payment workflow itself.
Atomic settlement is a legal event, not a magic button
In ordinary language, atomic settlement means every leg completes or none does. In a multi-currency payment, that can remove principal risk: a bank should not deliver one currency and discover that the corresponding currency never arrived. The technical mechanism locks the required balances, validates conditions and issues a common trigger.
Agora adds a necessary legal layer. Tokenised deposits sit on a unifying ledger, while tokenised central-bank reserves can remain on separate jurisdictional ledgers. Those ledger updates do not all happen at the exact same microsecond. The platform therefore needs its rulebooks to define the atomic settlement trigger as the moment when the payment becomes legally final and protected in insolvency across each jurisdiction.
That point deserves more attention than another throughput number. Technical irreversibility tells the software not to roll back a state. Legal finality tells an insolvency administrator, court and counterparty whose money it is. Serious cross-border money needs both.
CUSTOMER DEPOSIT CLAIM
|
v
TOKENISED COMMERCIAL-BANK LIABILITY
|
| coordinated with the interbank leg
v
TOKENISED CENTRAL-BANK RESERVES
|
| atomic trigger defined by rulebook
v
LEGAL FINALITY ACROSS EVERY RELEVANT LEDGERTokenised deposits inherit bank protection, but not unlimited safety
Calling a tokenised deposit “insured” without qualifiers is sloppy. Deposit-insurance regimes are generally technology-neutral, so recording an eligible deposit on a shared ledger should not by itself remove protection. Yet coverage still depends on the bank, account type, depositor category, currency, location and statutory limit. A large corporate or another bank may receive little or no protection even when a retail depositor at the same institution is covered.
Project Agora 2026 is a wholesale project. Its tokenised deposits can be held by financial institutions and corporate depositors, not ordinary retail customers. That means the strongest case is not “your token is insured.” It is “your claim remains inside a supervised bank relationship, with the same legal and prudential framework that applied before tokenisation.” Bank credit risk remains. The technology does not repeal it.
There is also a record-keeping problem hiding inside the promise. Deposit insurers and resolution authorities need a reliable single-customer view. A shared ledger must reconcile with the bank’s other books without creating two competing truths. If tokenised balances sit in a separate operating stack, recovery planning, sanctions evidence and depositor records must still work when the platform is unavailable.
Stablecoins still win the open-network contest
Stablecoins have one advantage banks should not wave away: distribution. A well-supported stablecoin can move between exchanges, wallets, merchants and public-chain applications without every user holding an account at the same bank. Tokenised deposits are more likely to remain permissioned, institutionally gated and tied to approved corridors. That makes them safer for some transactions and useless for others.
The better verdict is not that deposits kill stablecoins. It is that they attack the most valuable institutional use case. When a corporate payment already begins and ends in regulated bank accounts, replacing the bank claim with a stablecoin may add an issuer, a reserve pool and two conversion events without removing the need for banks at the edges. A tokenised deposit can modernise the middle while preserving the account relationship.
By contrast, the stablecoin can be the better rail when a recipient has no access to the issuing bank network, when a public blockchain application needs native cash-like value, or when 24/7 distribution matters more than integration with bank credit. Our analysis of Hong Kong’s licensed stablecoins reaches a similar conclusion from the other side: strong reserves and redemption rights improve the token, but they do not turn it into a bank account.
Wholesale CBDC is the anchor, not the product sold to the treasurer
A wholesale CBDC or tokenised reserve is the cleanest claim in the comparison because it is a liability of the central bank. That does not make it the obvious instrument for a corporate. In Agora, access to tokenised reserves stays under each central bank’s control and is generally limited to eligible institutions that already hold reserve accounts in that jurisdiction.
The corporate uses commercial-bank money. The banks settle their obligations with central-bank money. This division is the two-tier system, and Agora preserves it instead of offering every company a central-bank wallet. That choice keeps credit creation, customer service and compliance in commercial banks while retaining a public settlement anchor beneath them.
So the wholesale CBDC-versus-tokenised-deposit debate is slightly false. They perform different jobs in the same transaction. Deposit money faces the customer. Central-bank money settles between institutions. The architecture becomes interesting when both can execute under one coordinated workflow.
A tokenised fund is useful collateral, not final cash
A tokenised money market fund can look more attractive than either a deposit or stablecoin because it may pay market yield while remaining transferable on-chain. But the holder owns a fund security, not money issued by a bank or central bank. The token represents an interest in a portfolio. Its value, redemption and liquidity follow fund rules and securities law.
Even a fund designed to maintain a stable net asset value can face liquidity stress, valuation changes, settlement windows, transfer restrictions or redemption controls. Allow-listed wallets help the manager enforce investor eligibility, but they reduce open transferability. A tokenised fund can become excellent programmable collateral or a treasury investment. It still needs a cash leg when a payment must settle with finality.
This is why the UBS tokenised money market fund workflow matters without making the fund a new currency. Tokenisation can automate subscription, ownership and redemption. It does not erase the distinction between owning a fund unit and owning a bank balance.
The serious-money test: choose the claim before the rail
For treasury teams, the right instrument depends on the job. Start with the desired creditor and failure outcome, then consider speed. A payment tool that settles in seconds but leaves the legal claim unclear is not operationally mature. Neither is a beautifully regulated bank token that the counterparty cannot receive.
Measure the whole corridor, not the transfer fee
However, the cheapest-looking rail can be the expensive one once a treasury team prices trapped liquidity, manual repairs and the cost of turning the received token back into usable cash. Therefore, a fair comparison starts with the full operating chain, including the failure path.
All-in Corridor Cost = transfer fees + FX spread + liquidity carry + reconciliation labour + exception losses + redemption or off-ramp cost + legal and capital overhead.
Easy Global Banking corridor-cost model
This equation explains why Agora’s 80-second result should not dominate the investment case. A tokenised-deposit corridor could settle quickly and still disappoint if banks must pre-fund every currency, duplicate sanctions reviews or repair ledger mismatches by hand. Conversely, a slower rail can remain economical when it reaches almost every beneficiary and has a familiar exception process.
Deposit Rail Fit Score: a six-factor corridor diagnostic
Give each factor 0, 1 or 2 points. Zero means absent, one means moderate and two means decisive. This is an editorial screening tool, not a BIS model or a substitute for legal and risk review.
Open a worked example
| Treasury job | Strongest starting point | Reason | Question that can still stop the transaction |
|---|---|---|---|
| Large corporate payment between participating banks | Tokenised deposit plus tokenised reserves | Preserves bank claims and enables atomic interbank settlement | Do both banks and currencies share a governed corridor? |
| Public-chain payment to a crypto-native counterparty | Regulated stablecoin | Broader wallet and application reach | Can the recipient redeem lawfully at acceptable cost? |
| Interbank settlement asset | Wholesale CBDC or tokenised reserve | Direct central-bank liability and final settlement anchor | Is the institution eligible and is liquidity available? |
| Yield-bearing on-chain liquidity or collateral | Tokenised money market fund | Portfolio income and securities-law wrapper | Can the fund redeem when payment cash is needed? |
| Ordinary supplier payment outside token networks | Conventional bank transfer | Universal account reach and familiar evidence | Are beneficiary and address data ready for the payment chain? |
The last row is not a concession. Existing rails keep improving. From November 2026, for example, the new SWIFT structured-address rule changes how payment data must travel through CBPR+. Tokenisation must beat a moving conventional system, not a frozen picture of correspondent banking from ten years ago.
Open the legal claim before choosing the token
I want my ordinary bank relationship to remain intact
I need value that can reach public-chain wallets
I need the lowest-credit-risk wholesale settlement asset
I want on-chain yield or collateral
What must happen before Agora becomes infrastructure
Project Agora 2026 has answered the feasibility question more convincingly than most tokenisation projects. It has not answered the ownership, governance and economics questions for a production network. Who operates the unifying layer? Who pays for idle liquidity? Which currencies run around the clock? Who bears loss after a software defect, false compliance attestation or erroneous payment?
Scale also creates concentration. A common platform can remove duplicate processes, but it can become a shared point of operational dependence. Production design needs cyber resilience, fallback settlement, clear default rules, liquidity facilities, enforceable cross-border finality and a credible way to add smaller banks without turning membership into a club for global institutions.
My base case is a hybrid future. Project Agora 2026 points to tokenised deposits taking high-value bank corridors where legal finality and account continuity matter most. Regulated stablecoins will keep the edge in public-chain distribution. Tokenised funds will supply yield-bearing collateral. Central-bank money will sit underneath the serious settlement layer. The winners will interoperate; none will replace the entire stack.
Five operating questions worth resolving
Is Project Agora live?
What is a tokenised bank deposit?
Are tokenised bank deposits covered by deposit insurance?
Why not use a stablecoin instead?
Is a tokenised reserve the same as a CBDC?
Important: This article is general information, not legal, investment or payment advice. Project Agora remains experimental. Product rights, deposit protection, access rules, settlement finality and tax treatment depend on the instrument, account agreement and jurisdiction. Review the binding documents before moving real value.
References
- BIS Innovation Hub: Project Agora overview and July 2026 real-value testing update (opens in new tab)
- BIS: Project Agora – A shared programmable platform for wholesale cross-border payments (opens in new tab)
- BIS Annual Economic Report 2026: Anchoring trust in money (opens in new tab)
- BIS Bulletin 73: Stablecoins versus tokenised deposits (opens in new tab)
- BIS Bulletin 115: The rise of tokenised money market funds (opens in new tab)
