What is Project Agorá?
Project Agorá is a BIS Innovation Hub initiative testing whether tokenized commercial bank deposits and tokenized central bank reserves can settle wholesale cross border payments atomically on one programmable platform. Eight central banks and more than 40 financial institutions, convened by the BIS and the Institute of International Finance, completed a working prototype.
On May 27, 2026, the Bank for International Settlements published the prototype report for Project Agorá (often written Project Agora, from the Greek for public square), its flagship exploration of tokenized cross border payments, and the Bank of Canada announced it was joining as the project’s eighth central bank.
To date, Agorá is the largest project the BIS Innovation Hub has ever convened.
| Metric | Value | Source |
|---|---|---|
| Central banks participating | 8, across five major reserve currencies | BIS |
| Private institutions convened by the IIF | 41+, incl. JPMorgan, HSBC, UBS, Swift, Mastercard, Visa | BIS list |
| Prototype development | Two years, completed May 2026 | Ledger Insights |
| Settlement speed once funds are locked | Seconds | Ledger Insights |
| Priority friction points addressed | 9 of 9, materially or partially | BIS report |
| Production timeline | None. The project is experimental. | BIS |
The problem Agorá attacks is correspondent banking. Today a wholesale payment from Toronto to Seoul hops between intermediary banks, each running its own compliance checks in sequence, each holding the funds while it does. The result is settlement measured in days, costs measured in basis points that compound, and counterparty risk at every hop.
Information moves instantly and for free. Money still does not.
From hops to one operation
Nine priority friction points identified by participants; the prototype materially or partially addressed all nine.
Tokenization
Tokenization records a claim on a programmable ledger. A tokenized bank deposit is the same deposit, the same legal claim on the same bank, in a form software can act on directly. The BIS report is explicit: tokenization changes the record, not the legal character of the money behind it.
A conventional deposit sits in one bank’s database, so moving value between institutions means messages, reconciliation, and waiting. A token can carry its own settlement logic. Payment and delivery execute as one operation. Compliance runs against the transaction instead of chasing it. None of it requires a cryptocurrency or a public blockchain, which is why the people building it are central banks and regulated lenders. We made the longer case in Why fintechs should not be afraid of tokenisation.
How Project Agorá actually works
The prototype combines two kinds of tokenized money on one programmable platform: commercial bank deposits and central bank reserves. Deposits are where the world’s payments actually live, and reserves are the only settlement asset with no credit risk. The architecture that delivers this has the following two layers:
Two layers, one boundary
Coordination is shared. The money layer never is. That split is the entire design.
The unifying ledger. Tokenized commercial bank deposits sit on a shared ledger where all participants coordinate. This is where the workflow lives: payment orchestration, sequencing, the parallel compliance checks that replace today’s sequential ones, and real time visibility of payment status for every party in the chain.
The jurisdictional ledgers. Tokenized central bank reserves do not sit on the shared ledger. Each currency area runs its own independent ledger, operated under its own central bank’s authority. Workflow on the unifying ledger never directly triggers actions on a jurisdictional ledger. The central bank’s money, and the machinery that moves it, stays inside the central bank’s jurisdiction.
Privacy is held at both the balance and transaction level while still supporting regulatory compliance. Tokenization did not alter the legal character of reserves or deposits, so no new legal category of money was created. Legal analysis found settlement finality achievable across all seven original jurisdictions, with further work needed to map the technical and contractual requirements onto each legal framework.
The modular design also leaves room for what comes next: conditional and always on payments, and compliance functions (AML, sanctions screening, fraud detection) built into the platform as data sharing frameworks mature.
The design decision that matters
The two layer architecture was driven by the earliest and most consequential requirement the central banks set: they would not place their money on a ledger they did not control.
Eight of the world’s most powerful monetary authorities, holding five major reserve currencies between them, looked at a shared global settlement platform and made domestic control the precondition of participation. The entire architecture flows from it. This is also where Agorá departs from mBridge, the earlier BIS linked project that placed multiple central banks on a single shared ledger, an approach Agorá’s designers deliberately did not repeat.
mBridge shared the ledger. Agorá shares only the coordination.
Sovereignty was not a constraint the architects worked around. It was the design brief.
Canada arrived at a version of the same answer. Project Samara, completed in March 2026 by the Bank of Canada with RBC, TD, and Export Development Canada, settled the country’s first tokenized bond in wholesale central bank deposits. Built on Hyperledger Fabric, the platform ran separate bond and cash ledgers, achieving atomic settlement through smart contracts and an interledger mechanism rather than a shared book. The cash leg was wholesale Canadian dollars created and managed by the Bank of Canada itself.
Both projects landed in the same place: central bank money on its own ledger, under the issuer’s control, linked to the asset leg rather than merged with it. That is what makes it a requirement rather than a preference, and it is why settlement infrastructure has to be designed around it from the first decision.
What the Bank of Canada joining Agorá actually means
“We know that the Canadian economy could benefit from innovation in cross border payments,” said Senior Deputy Governor Carolyn Rogers. “Tokenization has the potential to make these payments faster, cheaper and more efficient and secure. Project Agorá is a unique opportunity to test the technology across several jurisdictions and currencies, with the participation of private sector financial institutions.”
Canada gets a voice in the next phase, which includes real value transactions in selected currencies. It puts the Canadian dollar on a path toward its own jurisdictional ledger, operated under the Bank of Canada’s authority, interoperable with the platform the reserve currencies are testing. This signals to Canadian financial institutions that the Bank of Canada considers this the direction of money movement.
Now the caveats, because they are where the strategy lives.
Canada joined after the design was set. The prototype is finished. The architecture, the token standards, the orchestration model, the privacy approach: all of it was specified over two years in which Canada was not in the room. We arrive as adopters of decisions made by others.
No Canadian institution helped build it. The closest thing to a Canadian name on the participant list is TD Bank N.A., the US subsidiary, participating on the US dollar rail under US jurisdiction. Canadian banks, as Canadian entities, contributed nothing to the prototype.
Most members arrived with domestic rails. Canada arrived with research. Switzerland has settled over CHF 750 million in wholesale CBDC on SIX Digital Exchange. Japan has the Japan Open Chain running in production with bank issued stablecoins, which we examined in detail here. These countries are connecting domestic tokenized infrastructure to Agorá. Canada, so far, is connecting intentions.
Agorá is not a product. BIS Innovation Hub projects investigate feasibility. There is no production commitment and no timeline for one.
The layer Agorá will never build
Looking at the two layer architecture: where do tokenized Canadian dollar deposits come from?
Agorá is an application, not a base layer. It is a coordination protocol that runs on top of national settlement infrastructure, orchestrating payments across ledgers that already exist. It does not provide those ledgers. Every participant is expected to arrive with its own sovereign Layer 1: the jurisdictional reserve ledger its central bank operates, and the domestic platform where its commercial bank money is tokenized. That sovereign base is the precondition Agorá’s design assumes.
For Canada, it does not exist yet.
Agorá’s unifying ledger coordinates tokenized commercial bank money. The project assumes each jurisdiction shows up with its money already tokenized, its issuers already accountable, its compliance already legible to regulators. The jurisdictional ledger for reserves is the Bank of Canada’s to operate. The tokenized deposit layer, the rails Canadian institutions will actually transact on every day, is missing today.
Two layers spoken for. One is not.
Agorá assumes the domestic layer exists. Someone has to build it, on Canadian terms.
That is the sovereign choice, and it is the same one we mapped in Sovereign by Design. If Canadian deposits get tokenized on foreign operated platforms because nothing domestic exists when the procurement decisions land, then Canada will have joined a project whose founding principle is domestic control. This surrenders domestic control of the layer that touches every Canadian transaction.
The questions that decide it are concrete. Who operates the validators, and do they reside, legally and physically, in Canada? Does the legal structure qualify for Basel Group 1 treatment, so regulated institutions can actually hold the instruments? Is compliance native to the settlement layer or bolted on after? Can the platform interoperate with an Agorá type unifying ledger without a rebuild?
CADX™ and the Open Value Network™ were designed as Canada’s answer to exactly these questions. A Canadian dollar backed stablecoin with 1:1 reserves and enforceable redemption. A permissioned settlement network with compliance as a native property of the rail. Canadian validators, Canadian law, Canadian settlement finality, built to plug into the interoperable future Agorá describes.
The uncomfortable conclusion
Tokenized deposits plus tokenized reserves on programmable, interoperable platforms is where wholesale settlement is going. Project Agorá proved this with the signatures of eight central banks and 41 global institutions underneath it.
What Agorá did not settle is who built Canada’s piece of it. The project’s own architecture reserves that decision for each jurisdiction. Sovereignty was the design requirement, and sovereignty is not something the BIS can test on Canada’s behalf.
The Bank of Canada just accepted the invitation to the interoperability layer. The domestic layer is still unclaimed, the procurement window is still open, and the countries Canada will be interoperating with claimed theirs years ago.
Questions we keep getting
Is Project Agorá a CBDC?
No. Agorá tokenizes money that already exists: commercial bank deposits and the wholesale reserves commercial banks hold at central banks. It creates no new retail currency and the platform is not open to retail participants.
Which central banks are in Project Agorá?
Eight: the Federal Reserve Bank of New York, Bank of England, Bank of France (representing the Eurosystem), Bank of Japan, Bank of Korea, Bank of Mexico, Swiss National Bank, and, as of May 27, 2026, the Bank of Canada.
Is Canada part of Project Agorá?
Yes. The Bank of Canada joined on May 27, 2026, the same day the prototype report was published, and will participate in the next phase. No Canadian financial institution took part in building the prototype.
When will Project Agorá go live?
There is no production timeline. BIS Innovation Hub projects test feasibility, and Agorá remains experimental. A decision to build production infrastructure has not been made.
What does Project Agorá mean for Canadian payments?
Directly, nothing yet. What it changes is the standard: it confirms that tokenized settlement with domestic control is where the reserve currency economies are heading, which raises the cost of Canada leaving its own domestic tokenized layer unbuilt.
If Project Agorá is just an application, why does Canada still need a sovereign Layer 1?
Agorá is a coordination layer, not infrastructure. It orchestrates payments across national ledgers that already exist; it does not create them. Its own design makes domestic control the precondition: each participant brings its own jurisdictional reserve ledger and the domestic platform where its commercial bank money is tokenized. A sovereign Layer 1 is that base. Without one, Canada connects to Agorá on infrastructure someone else governs, which is the opposite of what Agorá’s architecture was built to protect.
What happens if Canada’s tokenized layer is built on non sovereign technology?
Control of the Canadian dollar leg, along with the settlement data and economics that come with it, moves to operators outside Canadian jurisdiction; a node hosted on a foreign hyperscaler can pull Canadian settlement under foreign law such as the US CLOUD Act. A deposit token on a nonqualifying chain risks Basel Group 2 treatment, a 1250% risk weight that keeps regulated institutions from holding it at scale. And once procurement commitments to foreign rails accumulate, they are far more expensive to reverse than to avoid. Canada would have joined a project whose founding principle is domestic control while surrendering that control at home.
For policy and regulators. The sovereign choice has a two to three year window before procurement commitments lock in, and Agorá’s architecture just confirmed that domestic control is the international standard. Read Sovereign by Design and start the policy conversation before the default becomes the decision.
For institutions and partners. If you are evaluating settlement infrastructure, the questions above (validator residency, Basel treatment, native compliance, interoperability without a rebuild) are the ones to put to any vendor. CADX™ and the Open Value Network™ are built to answer them on Canadian rails.