Central bank technology projects usually end with a report. In July 2026, Project Agorá ended with settled transactions — real value, real currencies, real institutions — and that distinction is the reason the exercise deserves attention.

What happened

Twenty-eight financial institutions and central banks across Asia, Europe and North America completed real-value transactions totalling approximately CHF 800,000. Participants included JPMorgan, Citi and UBS. The programme ran 17 transaction scenarios with individual values between CHF 9,000 and CHF 125,000 or local currency equivalents, settling payments in six currencies on a shared, multi-currency programmable platform using tokenised central bank reserves and tokenised commercial bank deposits. Average completion time was 80 seconds.

The sum is deliberately trivial. The point was never volume. The point was demonstrating that a tokenised platform can settle genuine cross-border obligations between regulated institutions across jurisdictions, on a single ledger, without the correspondent banking chain that currently carries that traffic. The live run takes the project past the prototype stage reached in its findings report of 27 May 2026.

Project Agorá is convened by the Bank for International Settlements with a group of central banks and private participants. Its premise is that the cross-border payments problem is architectural. Money moves slowly and expensively between countries because it passes through a chain of correspondent relationships, each with its own compliance checks, cut-off times and liquidity requirements. Tokenising reserves and deposits onto a common platform collapses that chain.

Why Asia has more at stake

Asia carries a disproportionate share of global cross-border payment friction: a large share of world trade, an unusually high number of currencies relative to its economic integration, and enormous remittance flows. It has also spent a decade building an answer. ASEAN+3 has made real progress connecting national fast-payment systems and QR standards bilaterally, linking Singapore with Thailand, India, Malaysia and others.

Those linkages work, but they scale awkwardly. Each new connection is a separate negotiation between two central banks and two sets of banks. Regional analysis notes that while connectivity has advanced considerably, the bilateral model produces a growing web of arrangements rather than common infrastructure.

Tokenised platforms take a different route. Rather than connecting systems pairwise, participants transact on shared infrastructure supporting round-the-clock atomic settlement, programmability and potentially offline use. Atomic settlement means both legs of a transaction complete together or neither does, removing settlement risk without an intermediary to guarantee it.

The competing rails

Agorá is one of several answers to the same question. Stablecoins and tokenised deposits are moving from the periphery of the crypto economy into trade finance, as firms look for faster and more programmable ways to move money and settle obligations across borders. Hong Kong has licensed its first stablecoin issuers. Wholesale central bank digital currencies are progressing in several jurisdictions.

These are not equivalent instruments. Tokenised reserves carry central bank credit. Tokenised deposits carry commercial bank credit inside an existing regulatory perimeter. Stablecoins carry issuer credit and whatever the reserve backing supports. Treasury teams evaluating settlement options need to tell them apart rather than treating “tokenised” as one category.

What to do about it

Nothing urgent. Agorá is a proof of concept and production infrastructure is years away. The useful response is to work out which payment flows are actually expensive, typically the low-value, high-frequency, multi-currency ones, and ask banking partners about their tokenisation roadmap. Correspondent relationships priced on the assumption that there is no alternative may become negotiable sooner than expected.

Summary

Project Agorá settled real cross-border transactions in July 2026 across 28 institutions using tokenised reserves and deposits, demonstrating that shared multi-currency ledgers can work in practice. For Asia, which carries heavy cross-border payment friction and has pursued bilateral fast-payment links as the answer, the result points toward a different architecture. Nothing changes this year, but the assumption that correspondent banking is permanent is now weaker than it was.