Hong Kong Says Customers Will Choose Their Digital Money. It Has Built the Settlement Layer for One of Them.
On page four of its new Project Ensemble business use case report, the Hong Kong Monetary Authority makes a concession that most central banks avoid putting in writing. Innovation, competition and, above all, customer usage will decide which forms of digital money gain traction, and tokenised deposits and regulated stablecoins may each appeal to different end-users.
That is a statement of pluralism. The infrastructure published alongside it is not. The interoperability standard published alongside it is built to settle tokenised deposits against central bank money, and the bridge to anything else is, for now, a single verb: explore.
Neither half of that is a criticism. It is a sequencing choice, and a defensible one. But it has a practical consequence for every bank in Hong Kong that the reports leave implicit. We've focused here on that gap between the stated principle and the built infrastructure, and on what the sandbox results say about where the real cost of tokenisation sits.
Why this matters now
The HKMA has just published three documents: a business use case report covering the sandbox experiments, a shared taxonomy of digital finance terms, and an interoperability standard that it describes as fully operational and live.
They land in a market that has moved quickly. Project Ensemble's pilot phase, EnsembleTX, began on 13 November 2025, with seven banks offering tokenised deposits and settling real-value transactions, interbank settlement initially running through the existing HKD Real Time Gross Settlement system. Five months later, on 10 April 2026, the HKMA granted Hong Kong's first two stablecoin issuer licences under the Stablecoins Ordinance, to Anchorpoint Financial and HSBC.
So both instruments the report compares are now live or licensed in the same city, under the same regulator. The question of how they relate is no longer theoretical.
The concession
The report's comparison of digital money forms is brief but precise. A CBDC is a direct liability of the central bank. A tokenised deposit is issued by a commercial bank, backed one-to-one by corresponding off-chain commercial bank money on a balance sheet subject to the liquidity coverage ratio, and is typically account-based. A regulated stablecoin is issued by a licensed institution, backed by a segregated pool of high-quality, highly liquid reserve assets, and is token-based.
Having laid those out, the HKMA declines to rank them. Its position is that each has features that may suit different users, and that customer usage will ultimately settle which one wins in which use case. For a central bank, that is a notably market-led stance. It is also consistent with what we found when we looked at where the money actually sleeps in stablecoins and tokenised deposits: the two are different legal and economic instruments, not competing versions of one product.
The architecture
Read the interoperability standard next to that passage and the emphasis shifts.
The standard opens by noting that private forms of tokenised money carry credit and settlement risks absent from central bank money, which it calls the settlement asset of choice for wholesale transactions. Its eight business domains include wholesale CBDC, tokenised deposits and tokenised assets, alongside domestic and cross-border payment, foreign exchange, custody and exchange platforms. Stablecoins are not listed as a domain. Its phased path runs from settlement anchored on HKD RTGS, to 24/7 multi-currency settlement through tokenised central bank money, to a possible longer-term single-ledger model.
The use case report confirms the same scope in practice. Every sandbox experiment used tokenised deposits as the cash leg, with a hypothetical e-HKD used for settlement between banks, and EnsembleTX's next step is to settle tokenised deposits against CBDC. Interoperability with other instruments appears once, in the final box of the report, where the HKMA says Project Ensemble will continue to onboard banks, widen use cases, and explore interoperability with other forms of regulated digital money. The report's concluding diagram describes the interbank settlement function as supporting major forms of regulated digital money, which reads as the intended direction rather than a description of what runs today.
The taxonomy, which the HKMA is careful to say is not a legal reference and contains no legally binding definitions, points the same way: its entry for digital money gives CBDCs and tokenised deposits as the examples.
None of this contradicts the concession on page four. It means the HKMA has said the market will choose between two instruments while building public settlement infrastructure for one of them first.
One bank, both instruments
The clearest illustration is a single institution. HSBC completed the first live cross-bank transaction in EnsembleTX in November 2025, integrating its tokenised deposit service with the interoperability layer. In April 2026 it received one of Hong Kong's first two stablecoin issuer licences.
So the same bank now holds both instruments the report compares. One of them plugs into the public interoperability layer the HKMA has built. The other sits outside that layer's current scope. Customers will choose, as the HKMA says, but they will choose between an instrument with a public settlement rail and one whose connections into the banking system each institution has to work out for itself.
The technology works. The workflow is the bill.
The report's most useful finding for banks is not about money forms at all. It is that every use case confirmed the cross-platform orchestration model is technically viable, with delivery-versus-payment settlement completing in seconds, and that this was the easy part.
The HKMA is direct about it: technical viability is only one precondition for adoption. Of the five challenges it identifies, the first is that automated end-to-end processes are needed to realise the benefit, and that in the early stages they are hard to achieve. Its example is the tokenised money market fund, where net asset value calculation and unit register reconciliation between distributors, fund managers and administrators still run off-chain or manually. The cost of fixing that is spelled out too: overhauling existing off-chain and manual workflows could require significant upfront investment, which weakens the cost-benefit case for tokenisation.
That is an unusually candid admission from a programme sponsor, and it relocates the problem. The token settles instantly. The fund still strikes its price once a day, and three parties still reconcile a register by hand. Moving the token on-chain does not move the workflow around it, which matches the pattern we traced in the IMF's mapping of hybrid tokenisation infrastructure, where the hard interoperability problem sits between ledgers and the systems institutions already run.
The cost nobody tested
The fourth of the five challenges deserves more attention than it will get. If tokenisation succeeds, the report notes, banks may need to hold substantial interbank settlement balances to support their clients' instant tokenised transactions, making liquidity management a challenge. It then states plainly that this aspect was not covered in the sandbox experiments.
The implication is straightforward. A bank that offers 24/7 instant settlement has to fund 24/7 settlement capacity, and nobody in the sandbox measured what that costs. The HKMA says it is exploring operational arrangements that would let banks manage liquidity cost-effectively in a 24/7, high-speed, high-value environment, and treats that as a prerequisite for banks to offer the service at all. Until those arrangements exist, the price of round-the-clock settlement is an open question on every participating bank's balance sheet.
Where the HKMA says to start
The report's strategy section is pragmatic about sequencing. To build critical mass, it recommends prioritising products with fewer parties and fewer manual steps, naming repo and bond trades between financial institutions. Interbank trading of those instruments helps institutions manage their own liquidity and also builds market liquidity for tokenised assets, which the HKMA expects to become the foundation of a broader ecosystem.
That is the opposite of where most tokenisation marketing points. The retail fund and the real-world asset are the headline use cases. The HKMA's advice is to start with the dull, high-value, low-party-count transactions where automation is achievable and the benefit is immediate.
What follows for each group
Banks. The public settlement layer is being built for tokenised deposits first, and the HKMA has signalled a long-term intention to widen it. A bank that wants to serve customers who choose the other instrument should not plan on that capability arriving through Ensemble on any particular timetable. Separately, the unmeasured liquidity cost of 24/7 settlement is a balance sheet question each bank should model before committing to round-the-clock service levels.
Stablecoin issuers. Hong Kong licensing and Hong Kong settlement infrastructure are, for now, separate tracks. A licence establishes what may be issued. It does not establish a connection into the interbank layer the HKMA is building. Tracking how each jurisdiction's licensing regime relates to its payment infrastructure is part of why we maintain the free Stablecoin Regulation Tracker across 200+ markets.
Corporate treasurers. The treasury use case, run by Ant International with HSBC and Standard Chartered, explored real-time intra-group transfers between banks in Hong Kong and between Hong Kong and another jurisdiction. The report is nonetheless clear that the sandbox did not include use cases specifically testing cross-border settlement, and lists supporting it as an open challenge, so multi-jurisdiction treasury flows should be read as prospective.
Policymakers. The report's own framing, that the market should decide, implies infrastructure that does not pre-decide. Its phrasing on wider interoperability is the part worth watching as EnsembleTX moves toward CBDC settlement.
Plan for both, because only one arrives pre-connected
The HKMA's position is internally consistent. It has said users will choose, built the first rail for the instrument that sits closest to the existing banking system, and signalled that others may follow. That is a reasonable order of operations for a central bank.
For a bank, the practical reading is simpler. If customer usage really does determine which money wins, a bank cannot assume its customers will all choose the instrument the public infrastructure was built around. The tokenised deposit side comes with a settlement rail. The stablecoin side, today, comes with a licensing regime but no equivalent public settlement layer, and connecting it to existing banking operations is each institution's own work. The sandbox has already shown which part of that work is hard. It is not the token.
Source: Project Ensemble Business Use Case Report, Project Ensemble Taxonomy and Project Ensemble Interoperability Standard, Hong Kong Monetary Authority, 2026. The business use case report credits the Securities and Futures Commission as co-lead of the use case experiments. The taxonomy states that it is not a legal reference and contains no legally binding definitions.
Additional data: EnsembleTX launch date, participating bank count and RTGS settlement arrangement from the HKMA press release of 13 November 2025. HSBC's first EnsembleTX cross-bank transaction from HSBC's announcement of 13 November 2025. Stablecoin issuer licences granted to Anchorpoint Financial and HSBC from the HKMA press release of 10 April 2026.
Methodology: We read the three documents together, focusing on how the use case report's comparison of digital money forms relates to the scope of the interoperability standard, and on the challenges section of the use case report. The observation that public settlement infrastructure has been built for one instrument while the HKMA states that the market will choose between two is our reading of the documents, not a claim the HKMA makes. Project Ensemble does not currently settle stablecoins; the HKMA states it will explore interoperability with other forms of regulated digital money.
This article is for informational purposes only and does not constitute financial, investment, or legal advice.