What Has Actually Happened When Stablecoins Depegged

Five events, five ways to fail

People who swap into stablecoins are usually trying to avoid volatility. That's a reasonable aim, but know that what you avoided was price volatility, and what you took on is a different set of risks: the issuer's, the reserve assets', the regulator's, and a fourth only seen at scale in 2025 — the pricing and liquidation logic of the venue you happen to be standing on.

UST: the mechanism was broken at its root

UST was an algorithmic stablecoin, a design fundamentally different in principle from reserve-backed stablecoins, and ethereum.org's stablecoin overview separates the types reasonably clearly. It didn't hold a dollar by holding reserve dollars; it did so through a mechanism of minting and burning against another token. Logically, as long as the market is willing to arbitrage, the price returns to a dollar.

The problem is that the mechanism depends on confidence.

When the price starts falling, arbitrageurs have to take on UST, mint the paired token and sell it, and that action pushes the paired token's price down. The further the paired token falls, the more of it has to be minted, and the faster it falls. Once that loop starts, it accelerates itself.

In May 2022 that is exactly what happened. UST depegged and never returned, the paired token went to essentially zero within days, and the market value of the whole system evaporated. Nothing went unexpectedly wrong at any single point — this design walks to that outcome under extreme conditions. It looks clever while things are going up, and accelerates its own collapse on the way down.

USDC: the reserves were real, but where they're held is also a risk

USDC's situation was completely different. It has genuine reserve assets (composition and reports are on Circle's transparency page); the problem was that part of those reserves sat in a bank that failed.

In March 2023 that bank was taken over, the market worried the reserves couldn't be retrieved, USDC was sold off on the secondary market, and the price briefly fell to around 0.87 dollars (quotes differed between venues). A few days later, once the deposit situation was resolved, the price returned to about a dollar.

Holders ultimately suffered no permanent loss from the depeg itself in that episode, but two things about it are worth remembering.

One: during a panic you may not be able to sell at all. With the price at 0.87, swapping out at a dollar was impossible. Anyone who genuinely needed the money had to accept the discount, and that is a real loss.

Two: the knock-on effects reach other products. Other stablecoins pegged against USDC moved with it, and positions collateralised in USDC faced liquidation pressure. You thought you were just holding one stablecoin, and you were pulled into a whole chain of consequences.

The most immediate impression during those days was that the markets page looked wrong: pairs quoted in it were all priced high, and at a glance it looked like certain coins were surging when in fact the quote currency was what had moved. We did nothing at the time; in hindsight, doing nothing was right. Though honestly, that was more luck than judgement.

BUSD: a product may disappear without anything going wrong

The third one is a different nature entirely.

The third case was neither a mechanism problem nor a reserve problem. In February 2023 the issuer was directed by a regulator to stop new issuance (the New York Department of Financial Services' notice at the time is still on its site), and the stablecoin then entered an orderly wind-down.

For holders it didn't go to zero and didn't depeg significantly, but you had to convert it into something else before a certain date. Which is a reminder that there's another kind of stablecoin risk: it may stop being offered. That layer has nothing to do with how well a platform is run, and everything to do with whether the coin you chose is compliant and what regulatory environment its issuer sits in.

Two new ways to fail, added in 2025

The three events above happened between 2022 and 2023. Then came more than two years without a major incident, and plenty of people filed “stablecoin risk” away as a closed chapter. Two events in the autumn of 2025 reopened it — and both failed in ways none of the earlier three did.

USDe: the coin was fine, the venue you were standing on wasn’t

On 11 October 2025 the crypto market went through a liquidation cascade of rare size. In the middle of it, Ethena’s USDe printed around 0.65 dollars on Binance alone, while trading near a dollar on other venues at the same moment.

Ethena’s founder afterwards pointed at how Binance was pricing the asset at the time; Binance’s own analysis column also filed the episode as a pricing-mechanism failure under extreme conditions rather than a problem with USDe’s collateral. With hindsight, neither its redemption path nor its backing had broken.

So why did people lose real money? Because the exchange prices your liquidations off that local number. If you were posting USDe as margin on Binance, or running looped leverage on a lending protocol, the moment 0.65 appeared your position was closed out — whether the coin itself was sound had stopped being relevant to you.

The xUSD cluster: a chain reaction among yield-bearing stablecoins

A month later, in November 2025, another episode showed what the words “high-yield stablecoin” actually cost.

An external fund manager for Stream Finance disclosed roughly 93 million dollars of asset losses, and the platform suspended deposits and withdrawals on 4 November. Its yield-bearing stablecoin xUSD came off peg immediately, with public reports putting the low somewhere between 0.23 and 0.5 dollars.

The instructive part is what came next. Another protocol, Elixir, had over 68 million dollars of exposure to Stream — roughly 65% of the reserves behind its own stablecoin, deUSD. When xUSD went, deUSD’s backing went with it and a run followed. Elixir later wound deUSD down and compensated holders in USDC. The panic then spread to USDX, built on a similar model.

Straighten the chain out and it reads: one stablecoin’s reserve was another stablecoin, and that one’s reserve was money invested somewhere it could be lost. On ordinary days nobody sees the chain, because it presents itself as an attractive APY. On the day it snaps, everyone discovers simultaneously that they were holding the same single risk.

Which is why this site keeps asking the same first question: where does the yield come from? The percentage on its own tells you nothing. What tells you something is the sentence behind it — who is paying this, and what are they doing to earn it. That layer is covered in what the platform does with the coins you park.

One boundary worth stating: neither 2025 episode involved USDT or USDC, the mainstream stablecoins backed by cash and short-dated treasuries. Both involved “yield-bearing” and “synthetic” newcomers. They sit on the same screens and carry the same label, and they pay noticeably more — and that difference is precisely the extra thing you are carrying.

What happens on your account during a depeg

Talking about "a depeg" in the abstract doesn't convey its practical effect. Broken down concretely, you find it doesn't only affect people holding that coin.

Your stablecoin balance doesn't change, but its value does

A depeg doesn't reduce the number in your balance. You still hold the same count of coins; each one just no longer swaps for a dollar. That's psychologically subtle; the account looks fine while the purchasing power has already shrunk.

Positions collateralised with it get liquidated

If someone borrowed against that stablecoin as collateral, a falling price directly triggers margin calls and then forced liquidation. That's why one stablecoin's trouble tends to drag a whole area with it — liquidated positions sell other assets, pushing prices lower still.

Pair prices start "looking strange"

It isn't only the holders who are affected.

Pairs quoted in the depegged coin show abnormal jumps. People mistake this for some coin surging or crashing, when what actually moved was the quote side. Placing orders at moments like that easily fills you at the wrong price.

Earn products may suspend subscription and redemption

Related products may pause operations in extreme conditions and resume once prices stabilise. Which means you may not be able to run even if you want to, and in hindsight, the people who did run didn't necessarily do better than those who stayed.

Put those four together and the conclusion is clear: what actually hurts people in a depeg is rarely the price itself; it's the chain reaction and the forced actions it causes.

Why algorithmic stablecoins keep failing

UST wasn't the first algorithmic stablecoin to fail, and there were similar attempts before and after. The failure pattern is remarkably consistent, and worth spelling out, because projects like this will keep appearing.

Its stability depends on someone being willing to arbitrage

They all trip over almost the same thing.

Every algorithmic stablecoin's core assumption is that when the price falls below a dollar, someone will buy in and use some mechanism to convert back into a dollar of value, pushing the price up. That assumption holds in calm periods, because the arbitrageur really does capture the spread.

The problem is that an arbitrageur's behaviour depends on their confidence in the whole system. Once they doubt the mechanism can hold, the rational choice is to leave, not to arbitrage. So the moment arbitrageurs are needed most is precisely the moment they least want to act.

Reflexivity makes the fall accelerate itself

Worse, most of these designs mint more of the paired token as the stablecoin falls.

That pushes the paired token's price down, and its price is the source of confidence in the whole system. The more it falls, the more must be minted, and the lower the price goes. That's a positive feedback loop, and once started it's very hard to stop from outside.

Traditional finance has a similar phenomenon called a bank run, but banks have deposit insurance and a lender of last resort. Algorithmic stablecoins have neither: there's no external reserve on-chain to step in and buy, and no institution obliged to bid at that moment. So once the fall starts accelerating, the only thing that stops it is the selling pressure exhausting itself.

High yield is usually the fuel

These projects usually offer yields far above the market to attract capital. That yield either comes from new money coming in or from the project's own token reserves, and neither is sustainable. So "mechanism-based stablecoin plus an abnormally high APY" is, on its own, a complete warning signal.

When a project claims "our mechanism is different from UST's, we improved it", the question worth pressing isn't what they improved, it's: when everyone wants out at once, who is on the other side? If there's no satisfying answer to that, the ending can be the same however elegant the mechanism.

What the five cases say together

Looked at together, the five give you a few conclusions more useful than any single case.

The failure mode never repeats

One mechanism, one custody, one regulatory, one exchange-pricing, and one where yield-bearing stablecoins held each other as reserves. Five events, five different causes. Which means you can't guarantee safety by "avoiding whatever went wrong last time" — the next cause will likely be new again. What you can defend isn't the cause, it's the exposure: what share of your total position any single stablecoin or platform holds is the number you actually get to decide.

You can't tell in advance whether it recovers

During the two days USDC sat at 0.87, the market was equally full of people saying it was never coming back; in the early days of the UST collapse, plenty of people believed the mechanism would pull the price up. We know the outcomes now; nobody knew them at the time. That's why we don't write "what to do during a depeg"; it would require information you couldn't possibly have had.

The real loss happens at the moment you're forced to act

Across the five events, the three groups with the largest permanent losses were: people holding mechanism-based stablecoins, people who sold cheap in a panic, and people liquidated at a local venue price. The second group's loss comes entirely from "having to sell": needing the money, being liquidated, or simply not being able to take it. That one can be avoided by arranging things beforehand.

So what should an ordinary person do

You don't need to be an expert, but a few things are worth doing.

Don't hold just one stablecoin

The five events had different causes, which itself shows you can't avoid all risk by "picking the safest one". Holding two or three mainstream stablecoins is more practical than researching which one is perfect.

Look at what the issuer discloses

One with proof of reserves, regular reports and a clearly stated reserve composition is more dependable than one that says nothing. That information is usually on the issuer's own site — Tether's transparency page, for instance. Scope and update frequency differ, so look before deciding how much to believe.

The rest is a few pieces of dull work.

Don't rush to act during a depeg

In the USDC episode, people who sold cheap in the panic took a real loss while those who held were back where they started days later. Of course, we only know that in hindsight, the people who held through UST lost everything. So the real answer isn't "here's how to handle a depeg", it's arranging your position in advance so that even if one of them fails, you're never forced to act.

Tell a small wobble from a real depeg

Stablecoin prices wobble slightly around a dollar almost every day, and a few basis points is normal; it reflects current buying and selling, not a problem at the issuer. Panicking at 0.9995 leaves you numb by the time something actually deserves attention.

The signals that do deserve attention are roughly these: a clear deviation (say over one percent) that persists; accompanying substantive news about the issuer or custodian; large abnormal redemptions on-chain or at exchanges; other related assets moving in step. A small price wobble on its own isn't on this list.

So what to look at is why the price deviated. A small deviation with no findable cause is most likely a liquidity matter and will come back quickly; a deviation with a clear cause is the one to take seriously.

Fold this into your return expectations

If you've been reading stablecoin Earn APYs as "a risk-free return", these five events are the correction: it isn't. It's a return that requires you to carry issuer risk, custody risk, regulatory risk, and the pricing risk of whichever venue you are standing on. Understand that, and you'll look at those APY numbers more coolly.

The structural differences at the product level are in the structural differences between products; how to choose between stablecoins is covered separately in USDT, USDC or FDUSD.