Hong Kong Just Put Stablecoins Inside Government Debt. 50,000 Users Is the Number That Matters

Hong Kong is doing the thing Washington keeps announcing and never quite finishing. It's putting regulated stablecoins to work inside government debt, and it's running the settlement clock around the clock while it does it. Not a consultative paper. Not a five-year roadmap. Working rails.

That's the story. The headline number, 50,000 users on HSBC's RedCoin, isn't.

The cash leg problem nobody wanted to fix

Tokenized bonds have been around for years. You can issue one, register it on a distributed ledger, and settle the security side in minutes. Everyone claps. Then the cash shows up and it's a commercial bank transfer through the old correspondent chain, subject to banking hours and a two-day wait. The tokenization was theater. The money still moved by fax and patience.

The Hong Kong Monetary Authority's plan is to fix that specific leg. Licensed stablecoins, inside the existing tokenized government debt trials, settling continuously. Delivery and payment in the same instant, on the same ledger. No funding window where your counterparty could theoretically evaporate.

From a compliance standpoint, the sequencing here's the interesting part. The HKMA didn't start with stablecoins and reverse-engineer a use case. It started with tokenized government debt, ran Project Ensemble as a sandbox from March 2024, issued HK$800 million in tokenized green bonds in 2023 and followed with a multi-currency deal reported around HK$6 billion the next year, and only now bolts stablecoins onto a structure that already cleared legal review.

That's backwards from how nearly every other regulator does it. Most write a stablecoin rulebook first and then hunt for something to point it at.

The precedent here's important. If a licensed stablecoin can settle a sovereign debt instrument inside Hong Kong's legal perimeter, the same plumbing reaches repo, collateral mobility, and margin calls at 3 a.m. on a Sunday. Collateral that can't move on a weekend is collateral that costs you money.

Fifty thousand users is a rounding error, and that's fine

Now the steelman for the skeptics, because there's a real one.

HSBC has a customer base in the tens of millions. RedCoin has 50,000. That's a fraction of a fraction, in one city, on a product most of the bank's clients couldn't name if you spotted them the first syllable. Scale this against the US Treasury market and Hong Kong's HK$6 billion tokenized issuance looks like a slow Tuesday afternoon.

Reading between the lines, there's a sharper critique too. HSBC built its own tokenized deposit rather than plugging into a third-party stablecoin. That's a global systemically important bank capturing the settlement rail, not ceding it. If every major bank issues its own walled-garden token, you get twenty interoperable-in-theory coins and zero actual composability. The crypto-native promise of disintermediation dies politely, in committee.

And there's the liquidity question. A stablecoin float measured in the hundreds of millions can't absorb institutional bond settlement in size. Redeem at scale and you're testing redemption mechanics in public, which is the worst place to test them.

The licensing regime is deliberately narrow, full backing, licensed issuers only, no offshore shortcuts. That's a feature if you're a regulator and a bug if you're an issuer trying to build float. Ask anyone who's tried to get a license in that market.

So what's the bear case? That Hong Kong builds beautiful plumbing for a bathtub.

Why the plumbing still wins

I don't buy it, and here's why.

Volume follows settlement certainty, not the other way around. The reason tokenized bonds stayed a demo for five years isn't that institutions hated the technology. It's that the cash leg introduced a settlement risk their treasurers couldn't sign off on. Remove that risk and the objection disappears. The size of the market was never the binding constraint. The legal certainty of the trade was.

Second, a bank like HSBC running a stablecoin with regulator sign-off at any scale changes the reference class. 50,000 users is small. But it's a live product with a real balance sheet behind it, and that's a different animal from a sandbox simulation. The key detail is who's holding the liability. Not a startup. A bank whose deposits already clear through the system.

The US passed stablecoin legislation in July 2025 and set a genuine framework. Credit where it's due. But rulebooks aren't rails. Hong Kong is running the experiment while Washington writes the manual, and the gap between those two activities compounds.

Who wins here? Licensed issuers with bank partners, custodians that already have the securities lending relationships, and HSBC, which gets to define what a compliant tokenized deposit looks like before anyone else sets the standard. Who loses? Offshore issuers whose entire pitch was that they didn't need permission. And anyone who still believes crypto settlement and traditional settlement are two separate universes heading in different directions.

They aren't. They're the same universe now, and one of them has a 24/7 clock.

The verdict

Hong Kong wins the plumbing race. It probably won't win the volume race for years, and that's a distinction most coverage will get wrong by conflating the two.

My call: by the time the volume arrives, the standard will already be set, and the standard is worth more than the flow. Settlement conventions have a nasty habit of outliving the institutions that created them. SWIFT is older than most of the banks using it.

Can a stablecoin actually become boring infrastructure? That's the test. Boring is the goal. Boring is where the money hides.

Watch the next tokenized bond deal out of Hong Kong, specifically the cash leg. If it settles in a licensed stablecoin with no overnight gap, the argument is over and the rest of the world is playing catch-up. If it settles the old way, then this was a very well-designed pilot, and pilots are what regulators produce when they can't produce decisions.

Fifty thousand users today. The number to check is the one printed on the next deal's term sheet.