Tokenised Deposits: The Cross-Bank Test
Tokenised deposits can move fast inside one bank. The ECB's Pontes bridge tests the harder crossing: par value, final settlement and the missing 24/7 rail.
Tokenised deposits are easy to admire inside one bank's ledger. The awkward moment comes when a customer at Bank A needs to pay someone at Bank B. On 6 October, the European Central Bank made that cross-bank transfer the centre of its case for public money in tokenised finance. The ECB has launched a settlement service called Pontes. It has not declared the problem solved for every bank or every hour of the day. That distinction should shape how we judge the next wave of deposit-token announcements.
A deposit token is a claim on a commercial bank, recorded in a form that can move through a digital network. That claim still depends on the issuing bank. A token from Bank A is not automatically the same asset as a token from Bank B, even if both say "one euro" on the screen. The monetary system works because bank money is convertible at par through a common settlement anchor. A new database does not supply that anchor by itself.
One bank's token is another bank's claim
In its 6 October speech, the ECB warned that deposit tokens could become separate instruments trapped inside closed networks without central bank money to connect them. Its stated principle is plain: a euro issued by one regulated bank should exchange at par with a euro issued by another. This is a sharper test than asking whether a transfer can be recorded quickly on a ledger. A fast internal movement may leave the actual interbank obligation untouched.
Imagine a manufacturer whose bank issues a tokenised deposit. The manufacturer pays a supplier who banks elsewhere. The supplier does not want a permanent exposure to the manufacturer's bank merely because its payment arrived as a token. Its own bank needs to accept or convert the claim, and the banks need a way to discharge what they owe each other. If that conversion happens through an expensive, delayed or uncertain workaround, the clever front end has passed the friction down the line.
"Instant payments" may describe the customer's confirmation, while the receiving bank still has liquidity and reconciliation work to do. The question is whether money remains interchangeable when private issuers build separate rails.
A bank can automate transfers between its customers and connect cash instructions to asset transactions it controls. The scale claim begins at the boundary: what happens when the other party and the settlement institution are outside the system? The contract for convertibility matters more than a demonstration video.
Pontes provides a public settlement anchor
The Eurosystem launched Pontes on 21 September. It connects eligible market distributed-ledger platforms with Eurosystem services so wholesale tokenised asset transactions can settle in central bank money. An initial group of participants and operators had completed onboarding. That is an operational launch, not proof that every deposit token already moves across Europe.
The Pontes design page describes a dual model. A transaction can use cash tokens on a Eurosystem DLT platform or settle in T2, the Eurosystem's real-time gross settlement system. For the T2 route, the cash leg reaches final settlement when the corresponding transaction completes in T2. The design also describes a Hash-Link mechanism for synchronised delivery versus payment across platforms. These details matter because a transaction's record, the cash leg and the legal moment of finality have to agree.
That connects with our earlier account of tokenisation's settlement bottleneck. An asset token can change hands on a ledger while its buyer's payment is still pending elsewhere. Pontes is aimed at bringing a trusted cash leg to the transaction. The ECB's latest speech extends the argument to deposits issued by different banks: central bank money can serve as the bridge between their private claims.
Do not turn those two uses into a claim that Pontes already handles every deposit-token payment. Its initial eligibility rules name particular classes of participants and operators. The ECB describes the wider cross-bank ecosystem as work to build. Anyone claiming mass interoperability should show the participants, assets, settlement flow and rules.
The clock is still part of the product
Tokenised finance is routinely sold as always on. The ECB's October speech says Pontes will gain longer operating hours and immediate finality on its DLT platform as it develops. It expects full implementation by mid-2028, with a planned 24/7 service, more programmability and multi-currency capability. These are future features, not a description of what the first release can do at any hour.
That is an important limit for a buyer evaluating a weekend or cross-time-zone transaction. An asset can sit on a network that accepts instructions at any time, while the cash settlement it depends on follows a different operating schedule. A receiving institution may then be asked to carry exposure until the public-money leg can complete. The exposure might be managed through prefunding, credit limits or other arrangements, but none should be assumed away by calling the transaction atomic.
Ask for the cut-off times and the state of the obligation before final settlement. Can a payment instruction be cancelled? When can the recipient use the funds without recourse? If one ledger is up and another is down, who has the authority to unwind or complete the transaction? These questions are dull right up until a large payment sits between systems overnight. It is easier to fix them in the rulebook than after a dispute.
The 2028 target is a timetable, not a warranty. Banks should say which capability is live, which is under development and which depends on other infrastructures joining. Buyers need that separation before moving material liquidity onto a new rail.
Appia is where the rulebook gets written
The ECB separates Pontes, its operational settlement service, from Appia, the work on standards and governance for a wider European tokenised market. In the same speech, it says Appia is working towards a blueprint in 2028. Its scope includes asset interoperability, collateral, cross-border transactions, resilience and legal interaction. That is a wider job than wiring two pieces of software together.
A bank deposit is a liability with an issuer and a set of rights. Its representation on another network must preserve who owes whom, how the claim transfers and what happens if a participant fails. If a platform can technically transfer a token but the receiving institution cannot recognise the claim or mobilise it, the transfer has not created useful market-wide liquidity. Common standards and enforceable arrangements decide whether the token can leave its home system and still work as money.
Our earlier look at tokenisation ownership approached the same problem from the asset side. A record of transfer does not settle every legal question about title or custody. On the money side, the corresponding issue is whether a recipient gets the bank claim it thought it bought and whether the interbank obligation has truly settled. The code can make the event legible. The rules have to make it binding.
There is a practical scorecard here for anyone evaluating a tokenised deposit network. Start with a payment between two unrelated banks, not a transfer between two wallets at one institution. Trace the issuing bank's liability, the receiving bank's claim, the cash settlement asset and the point of legal finality. Run the test outside ordinary business hours. Then ask what happens when a platform fails during the handoff. If the answers require bespoke bilateral deals for every new participant, the network may be a collection of islands with a polished map.
The ECB's October intervention is worth reading because it puts a constraint on the sales pitch. Public settlement infrastructure can help private bank money travel farther. Pontes has begun that work; Appia is meant to shape the shared rules; the 24/7 version remains planned. The test is whether a deposit token can cross a bank boundary at par, with a clear settlement and a recipient who knows exactly what it owns. Until then, a fast token is still somebody else's bank claim.