Algorithmic stablecoins tried to build money with no bank behind it. What they achieved, why most collapsed, and the lesson that outlives them.
45 billion dollars disappeared in three days in May 2022, and the software that destroyed it was working exactly as designed.
That is the part people tend to skip. There was no hack, no stolen key, no developer vanishing with the treasury. The code did precisely what it had been written to do, and doing it correctly is what killed it. The project was called Terra, its token was called UST, and it was the most ambitious attempt anyone has made to build money that requires no permission.
I want to talk about that attempt properly, because it has become a punchline, and punchlines are where good questions go to die.
What they were actually trying to build
I have written before about how stablecoins became crypto’s quiet giant, and the model that won is a resolutely dull one. Someone holds a dollar in a reserve account and issues you a token that stands for it. It works because the dollar is genuinely there.
The algorithmic version looked at that arrangement and found it faintly insulting. All that capital, sitting in an account, doing nothing but proving a point. And worse, it puts a company at the centre of everything: an entity with a bank, a jurisdiction, a compliance department, and a telephone number a regulator can ring.
So they tried to remove it. Rather than reserves, you pair the stablecoin with a second token that floats freely. If the stablecoin trades above a dollar, the system allows people to create more of it by destroying some of the partner token, and the fresh supply pushes the price back down. If it trades below, the process runs in reverse, supply shrinks, and the price climbs. Arbitrage does the work. Nobody needs to be trusted, because nobody is in charge.
That is not a small idea. That is an attempt to build a currency the way you would build a law of physics.
Two different escapes
Here is the distinction I think gets lost, and it matters enormously if you have ever wondered whether any of this frees you from the currency system you were born into.
There are two separate things you might want to escape. One is the unit itself, the pound or the dollar, and the quiet erosion of what it buys. The other is the machinery built around it: the accounts that can be frozen, the intermediaries taking a slice at every step, the institutions deciding who gets access and on what terms.
Almost every stablecoin, this one included, still tracks a dollar. On the first count, then, nothing changes at all. But on the second count, the algorithmic design was the only serious attempt at a clean break. No issuer. No reserve account. No company that could be sanctioned, leaned on, or run badly by people you will never meet.
That is what it achieved, and for a while, within limits, it genuinely worked. Value crossed borders in seconds with no institution standing in the middle taking a cut. It was capital efficiency, because no money sat idle demonstrating solvency. And it could not be switched off, for the excellent reason that there was no switch.
Then it emerged that there might be a reason every other system keeps somebody in charge.
Why it broke
UST was propped up by a lending protocol paying close to twenty per cent. At its peak in April 2022, it was the third-largest stablecoin in the world, worth roughly $17.5 billion.
The unravelling took days. Confidence wobbled, holders moved to the exit, and the design permitted anyone to destroy one UST in exchange for a dollar of the partner token regardless of what that token was actually worth. So they did, in enormous volume, and the partner token was inflated into nothing. Chainalysis published a clear account of the trades that started it.
The structure had a circular problem sitting underneath it the entire time. The stablecoin was held up by confidence in the partner token. The partner token was worth something because of confidence in the stablecoin. Two people leaning on one another. It looks like a building right up until one of them shifts their weight.
The tell I use now
I will be honest about what I felt watching it, because it was not superiority.
A yield of twenty per cent attached to something described as stable is a contradiction wearing a very good coat, and I know precisely how persuasive a good coat can be. Impatience has cost me money before. The version of me who lost it was not stupid; she was in a hurry, and being in a hurry makes you read the number and skip the sentence underneath it.
What I do now is boring, and it works. I ask where the return is coming from, and I keep asking until I get an answer involving somebody somewhere actually producing something. If the honest answer is that the return comes from new people arriving, then the return is the new people, and it will last exactly as long as they keep arriving.
Where it stands now
Regulators have reached a settled view, and the view is no. The GENIUS Act in the United States leaves algorithmic designs outside its definition of a payment stablecoin altogether, and Europe’s MiCA rules effectively shut them out of that market. Britain has been characteristically less dramatic and rather more thorough, with the Bank of England setting out its rules for sterling-denominated systemic stablecoin issuers in June 2026. Nobody has written a law containing the word forbidden. They have drawn the perimeter so that reserves you can count sit inside it, and clever mechanisms do not.
The idea did not vanish, though. It grew up and found work. The surviving descendants hold real collateral, only more of it than they issue, so there is an actual cushion rather than a promise. A newer variety holds volatile assets and hedges them with offsetting positions so the swings cancel each other out. That is genuinely ingenious, and it depends on those hedges remaining available and affordable, which is a different risk rather than no risk at all.
What I actually take from it
Not that the idea was foolish. It was not, and I suspect some version of it eventually will work.
The lesson is narrower and far more useful, and it is not really about crypto at all. Terra’s stability depended on Terra staying popular. Growth was the collateral. Which means it was never stable; it was momentum in a costume, and momentum only resembles stability for as long as it is still moving.
You will find that shape in more places than you would expect, once you know how to spot it: a business whose margins hold up while it keeps expanding. A property market propped up by the belief that property goes up. A career that works while the industry around it grows. None of these is a fraud. They are simply arrangements that require tomorrow to resemble yesterday, and every so often tomorrow declines.
So the question worth carrying is not whether something is going well. It is whether it would survive going quiet.
The Jacqueline Brand — knowledge builds confidence, confidence builds wealth. This is editorial commentary for inspiration, not financial or professional advice. Always do your own research. The Collection
© 2026 TheJacquelineBrand. All rights reserved. Please do not reproduce or republish without written permission.
