Meridian FX APAC: a payment can be coordinated before its legal outcome is settled
The latest settlement experiment makes the boundary between software coordination and enforceable payment outcomes worth examining.

The most interesting question about a cross-border payment experiment is not always whether its systems can communicate. It is what a bank can safely do after receiving their answer. Can it release goods, discharge a debt or make another payment? A successful message exchange and an enforceable conclusion are different products.
On 13 November 2026, the Bank of England published the Meridian FX APAC policy report. It distinguishes technical coordination from legal finality and makes no commitment to live deployment. The accompanying technical report describes a test: once the first settlement instruction executes, cancelling reservations alone cannot reverse the transaction. This article examines that boundary, rather than treating the experiment as a launched payment service.
Three meanings of “done”
Consider a hypothetical importer buying foreign currency. Its bank might first receive confirmation that money has been reserved. Later, the system might report that a transfer has been processed. The importer ultimately needs something stronger: a reliable basis for treating its obligation as discharged. These are not interchangeable answers.
A reservation is useful precisely because it limits what the owner can do next. But the very same restriction can become costly if a process stalls. A processing receipt establishes that a machine reached a state; its commercial meaning depends on the rules attached to that state. Legal certainty concerns which claims survive and which obligations have ended, including when a participant fails.
The longstanding Principles for Financial Market Infrastructures address finality and linked settlement separately: linked obligations should settle conditionally on each other, while final settlement must be clear and certain. That distinction offers a useful analytical lens, not a finding that any prototype already satisfies the standards.
For a buyer of settlement services, the practical product is therefore not just faster communication. It is a sufficiently dependable answer to the question: “What can I now rely on?” A technically impressive system that leaves this answer unclear may save processing time while preserving expensive legal and operational caution. Extra buffers do not necessarily signal distrust of software; participants may simply lack a clear commitment about responsibility.
Design the difficult minute, not only the successful second
Imagine, as an analytical example rather than a reported test result, that one system has completed a transfer but the acknowledgment disappears. A coordinator sees silence. Silence does not prove failure; nor does it prove success.
Blindly sending the instruction again risks acting twice. Abandoning it risks leaving a valid payment outside the coordinator’s records. Waiting indefinitely preserves uncertainty while tying up resources. The sensible response depends on being able to establish the authoritative state and on having a previously agreed procedure for acting on it.
This is why an attractive average processing time says little about the hardest operational question. Participants need to know who can declare an incident, whose record governs, who may authorize recovery and when customers receive a definitive explanation. Those choices influence staffing, reconciliation and liquidity planning even if software normally works perfectly.
A reasonable counterargument is that such arrangements can evolve after technical feasibility has been demonstrated. That is true for experimentation. It becomes less convincing as a basis for production commitments: a commercial participant must price and manage exceptional outcomes before, not after, it depends on the service. A question left open in an experiment cannot necessarily remain unanswered in a contract.
Liquidity savings have a clock
A second hypothetical illustrates the trade-off. Several banks may reduce their funding needs by coordinating offsetting payments instead of funding every payment separately. But if completing the group requires waiting for the slowest participant, another bank may lose the use of money it needed elsewhere.
There is no contradiction. Less money required at the intended settlement moment and longer money tied up before that moment can coexist. A useful commercial evaluation would compare both the amount and duration of restricted funds, especially under stress. A headline about gross processing speed or net funding requirements cannot substitute for that comparison.
Different customers may rationally choose different services. A participant with predictable flows may tolerate batching; one facing time-critical obligations may value certainty of timing more. Interoperability does not erase these preferences. It makes the rules governing their interaction more important. Efficiency needs to specify for whom, under which conditions, and who bears the cost of waiting.
A narrower claim is a stronger foundation
Our assessment is that the valuable next question is institutional: can each participant explain the same exceptional transaction consistently to its own operators, customers and authorities? That is more demanding than displaying a common success message, but more useful than claiming that all systems must become identical.
Evidence that would strengthen a production case includes clear recovery responsibilities, demonstrable reconciliation after interrupted communications and a defensible account of the legal outcome in every participating jurisdiction. These are evaluation criteria, not claims about completed Meridian work. Nor do they establish superiority over existing services: costs, coverage and participant needs still require comparison.
The experiment should therefore be read as a reason to ask better questions, not as proof that cross-border settlement risk has disappeared. Connecting ledgers can expand what is technically possible. Turning that connection into a dependable financial service requires explaining what happens when the connection stops behaving as expected.
Information only. No investment, legal, tax, or financial advice.