For most of the last decade, whenever the word "tokenisation" came up in a policy meeting, someone would eventually sigh and say, "another pilot." And they were usually right. We had proofs-of-concept the way we had monsoon puddles: everywhere, briefly, and gone by afternoon. A bank would tokenise a bond, issue a press release, and move on. The technology was real, but it lived in a sandbox, walled off from the systems that actually move the world's money.

That season is ending. What I find striking, watching from a central bank in a small open economy, is not any single breakthrough but a change of register. Tokenisation is quietly moving out of the innovation lab and into the core architecture of finance. Institutions are no longer asking whether the plumbing can be rebuilt in tokenised form. They are rebuilding it.

From experiment to infrastructure
Consider what has happened in a matter of months. JPMorgan's Kinexys platform, once the modest "JPM Coin," now processes billions of dollars in transactions every day for institutional clients, handling intraday repo, cross-border payments, and foreign exchange settlement on a permissioned blockchain. In June 2026 it went further, taking its dollar deposit token live on a public network for institutional clients with near-instant, around-the-clock settlement. Wells Fargo announced in August 2026 that it will begin offering tokenised deposits to corporate clients this autumn. Citi, Goldman, UBS, BNY, Fnality, Partior, the list of bank-grade token platforms that have crossed from pilot into production reads like a roll-call of the institutions that actually run the wholesale system.

This is the shift that matters. When enterprise platforms start handling token issuance, lifecycle management, and smart-contract execution across asset classes as ordinary business rather than experiment, the centre of gravity has moved. The interesting question is no longer "does it work?" It is "who settles, and in what?"

The quiet victory of the tokenised deposit
Here is where I want to be direct, because I think a genuine reordering is underway that has not fully registered in public debate.

Stablecoins deserve real credit. They proved, at scale, that programmable money is useful. The total stablecoin market sat around $308 billion in mid-2026, up roughly 19 percent over the year, and in a single quarter these instruments moved tens of trillions of dollars. For businesses in emerging markets weary of correspondent-banking delays, a digital dollar that settles in minutes on a weekend is not a gimmick. It is oxygen. The IMF has noted that among businesses using stablecoins, a striking share report cost savings of ten percent or more, mostly on cross-border payments.

But utility is not the same as soundness, and this is the distinction I keep returning to. Last year the Bank for International Settlements, in its annual report chapter on the next-generation monetary system, argued that stablecoins fall short on three tests any money must pass: singleness, elasticity and integrity. Singleness is the quiet miracle we take for granted, that a dollar is a dollar is a dollar, whoever issues the deposit that holds it. A privately issued token that might trade at ninety-nine cents in a stressed moment does not offer that guarantee. It is a claim, not settled money.

Which is why the tokenised deposit is emerging, in my reading, as the preferred settlement backbone for serious institutional and wholesale flows. A deposit token is not a new species of money. It is the deposit you already hold at a regulated bank, recorded on a programmable ledger, still a claim on the bank, still redeemable at par, still inside the two-tier monetary system that took a century to build. It keeps the bank's balance sheet relevant. And it delivers what everyone actually wants from the crypto experiment: real-time, atomic settlement, where the cash leg and the asset leg move together or not at all, eliminating the settlement risk that has haunted markets since Herstatt.

The GENIUS Act in the United States, signed in July 2025, sharpened this line by treating payment stablecoins as payment instruments and barring them from paying yield. The effect has been telling. Even as stablecoin usage hit record highs, capital began drifting toward tokenised Treasuries and bank-issued tokens. Regulation did not kill the stablecoin; it clarified its place, and cleared room for regulated money to occupy the institutional core.

The problem nobody puts on a slide: fragmentation
There is a trap in all of this, and it is the one I worry about most from where I sit.

If every large bank builds its own private chain, we will have simply rebuilt the correspondent-banking maze in shinier materials. Fifty walled gardens do not make a system. They make fifty new reconciliation problems. The whole point of a shared ledger is that money and assets sit together; that promise dissolves the moment each institution keeps its own isolated ledger and we are back to bridges, messages, and delay.

This is why the most consequential work happening right now is not any single token. It is orchestration, the connective layer, the cross-chain standards, the governance that lets these islands talk to one another safely. And crucially, it is the decision to embed identity, compliance and governance directly into the permissioned framework, rather than bolting them on afterward. Compliance as a property of the rail, not a checkpoint beside it.

The clearest evidence that this can work is the BIS Innovation Hub's Project Agorá. In May 2026 the project reported that its prototype had achieved atomic, multi-currency settlement of wholesale cross-border payments on a shared programmable platform, tokenised commercial-bank deposits and tokenised central-bank reserves, operating side by side, across jurisdictions, potentially around the clock. Smart contracts let institutions embed compliance requirements and conditional triggers directly into the transaction, collapsing the sequential correspondent chain into a single simultaneous act. Agorá is now moving to real-value testing. That is not a puddle. That is a foundation being poured.

Why a small economy should be watching closely
I will end where I actually stand, which is neither Wall Street nor Basel, but a central bank in a country of remittance-dependent households and thin financial markets.

For economies like ours, the temptation is to treat all of this as somebody else's future, a rich-country debate about rails we do not own. That would be a mistake. The architecture being set now, in these pilots that have stopped being pilots, will define the terms on which our banks, our exporters and our migrant workers plug into global settlement for the next generation. If the wholesale layer converges on tokenised deposits and interoperable central-bank money, the cost of notbeing interoperable will be paid by the smallest and the poorest, as it always is.

The lesson I draw is an old central-banker's lesson, dressed in new code. Money is a public good that runs on private plumbing, and when the plumbing changes, the change arrives quietly, underground, long before anyone notices the water pressure. Tokenisation is that kind of change. The headlines are about tokens. The story is about settlement. And settlement, in the end, is just trust made mechanical which is the one thing a central bank exists to keep sound.

The plumbing is changing. We should be in the room where it is being drawn.

The views here are personal and do not represent those of any institution with which the author is affiliated.