A short history of stablecoins in five crises

Crisis one: the first designs, 2014 to 2018

In 2014 crypto traders had no stable asset to park money in between trades. Moving back to dollars meant a bank transfer that was slow, closed at weekends and often refused to crypto businesses. Three projects tried to fix that in the same year.

BitUSD, launched on the BitShares blockchain in July 2014, was backed by BTS, the chain's own token, posted at twice the value of the dollars issued. NuBits, launched in September 2014, used two tokens: NuBits as the dollar and NuShares as a second token meant to absorb swings in demand. The third was Realcoin, soon renamed Tether, which took the plain route: hold real dollars in a bank and issue one token for each. The first USDT were created in October 2014 on Omni, a layer on top of Bitcoin.

The two crypto-native designs held their pegs in calm markets and broke in rough ones. NuBits lost its peg in 2016 and again in March 2018, when it fell to about 0.25 USD and never came back. BitUSD went through an emergency shutdown, called global settlement, on 25 November 2018, after the price of BTS fell so far that the collateral no longer covered the tokens.

The reason was the same in both cases. The collateral lost value at the same moment, and for the same reason, as the token it was meant to protect. When people lost faith in BitShares, they sold both BTS and BitUSD, so the backing shrank exactly when it was needed.

Crisis two: Tether's missing audit, 2017 to 2021

Tether won the early race by being simple. A token that claimed one dollar in a bank for every USDT, listed on the large exchange Bitfinex, gave traders a dollar that moved at crypto speed. That was enough, even though Tether published very little about its reserves.

Scale ran ahead of oversight. In January 2018 alone Tether issued 850 mn new USDT, more than in any earlier month. Later that month it ended its relationship with its auditor, Friedman LLP, without a completed audit. For years the market argued about whether the dollars behind USDT existed.

The answer arrived through regulators. In 2021 Tether and Bitfinex paid 18.5 mn USD to settle with the New York Attorney General over misleading statements about USDT's backing. The same year the Commodity Futures Trading Commission fined Tether 41 mn USD. The CFTC found that, over a 26-month sample from 2016 to 2018, USDT had been fully backed by fiat reserves on only about 28% of days.

And yet USDT never lost its peg for good, and it is still the largest stablecoin. The lesson is uncomfortable. Traders paid for liquidity and wide acceptance, and reserve transparency came second. Transparency does not move the daily price. It matters on the day something breaks, which is why it matters so much to regulators.

Crisis three: Black Thursday, March 2020

By 2020 the main crypto-native dollar was DAI, issued by MakerDAO. A collateralized debt position, or CDP, is a vault where a user locks crypto worth more than the DAI they borrow against it. If the collateral's price falls too far, the system sells it at auction to repay the debt. In November 2019 DAI moved to a multi-collateral design that accepted more than one asset.

On 12 March 2020, now known as Black Thursday, ether roughly halved in value within about a day and a half. Hundreds of vaults fell below their limits at once. Ethereum was so congested that transaction fees spiked and MakerDAO's price feed updated late. The auctions that were supposed to sell collateral to many competing bidders drew almost none, because bidders could not get their transactions through.

One bot noticed it could bid zero DAI and win. It took about 8.3 mn USD of ether for nothing. Vault owners lost their whole collateral, and MakerDAO was left with debt that no collateral covered. To fill the hole, it created new MKR, its governance token, and sold it for about 5.3 mn DAI.

The lesson was narrower than "DeFi failed". The code did exactly what it said. It assumed that bidders would always show up, and that assumption broke on the one day the network was least able to carry them. MakerDAO's later answer was to add stable assets, and later tokenized Treasury bills, to its backing, which is how DAI came to depend partly on USDC.

Crisis four: Terra and UST, May 2022

Terra's UST was an algorithmic stablecoin. An algorithmic stablecoin holds no outside reserve. It relies on a sister token, here LUNA, and a rule: anyone could swap one UST for one dollar's worth of newly created LUNA, and the reverse. If UST fell below a dollar, traders were meant to buy it cheaply, swap it for LUNA worth a dollar, and pocket the difference.

Demand for UST came mostly from Anchor Protocol, which paid close to 20% a year on UST deposits. That rate was subsidised rather than earned from real borrowing. By early May 2022 UST supply was above 18 bn USD, and most of it sat in Anchor.

On 7 and 8 May large withdrawals and sales pushed UST below a dollar. Holders rushed to swap UST for LUNA, which created LUNA in huge amounts and crushed its price, which made every new swap worth less. A death spiral is this loop running in reverse: each exit weakens the backing for the next holder. The Luna Foundation Guard spent a reserve of about 80,000 bitcoin trying to defend the peg, and it was not enough. By 16 May UST traded at about 0.12 USD and LUNA was close to zero.

More than 40 bn USD of value disappeared within a week. The damage spread to lenders that had bet on the ecosystem: the hedge fund Three Arrows Capital collapsed, and the crypto lender Celsius froze customer withdrawals and filed for bankruptcy on 13 July 2022.

The cause was the design, and panic only set it off. LUNA had no value independent of UST, which was the BitUSD flaw of 2018 again, dressed in a 20% yield that kept most holders from asking where the money came from. After Terra, algorithmic stablecoins shrank to a tiny corner of the market, below 1 bn USD in September 2026, according to DefiLlama.

Crisis five: SVB and BUSD, 2023

March 2023 produced two different kinds of failure in one month, both at fully backed, regulated tokens.

The first was a depeg. On Friday 10 March US regulators closed Silicon Valley Bank. Circle, the issuer of USDC, held 3.3 bn USD of its reserves there, about 8% of the total. Banks were closed over the weekend, so Circle could not pay redemptions, and USDC fell to about 0.87 USD on Saturday. The damage spread through links between tokens. DAI traded near 0.88 USD because a large share of its backing sat in USDC through its Peg Stability Module, a contract that swaps DAI and USDC one for one. On Sunday the authorities said all SVB deposits would be protected. Redemptions resumed on Monday, and USDC returned to its peg within days.

The second was an orderly shutdown. Paxos issued BUSD, a stablecoin branded by the exchange Binance and worth about 16 bn USD at the time. In February 2023 the New York Department of Financial Services ordered Paxos to stop issuing new BUSD, citing its oversight of the Binance relationship. The reserve itself was not the issue. Minting stopped on 21 February 2023, holders could redeem for months afterwards, and BUSD wound down without a depeg.

Together the two cases taught the market that a full reserve is not the end of the story. USDC showed that even Treasury bills cannot hold a peg if the bank holding the cash is shut and redemptions stop. BUSD showed that a stablecoin can be closed by its supervisor when the issuer's partner is the problem.

What the depegs have in common

A depeg is a period when a stablecoin trades clearly away from its target price. The depth of the drop tells you little about what happens next. USDC fell 13% and recovered within days. UST fell 99% and never came back.

What decides recovery is whether, at the moment of stress, the stablecoin had assets that did not depend on the token itself, and whether a way to redeem against them was still working. USDC had a reserve of Treasury bills that did not care about USDC's price, and a desk that reopened on Monday. UST had only LUNA, whose value came from UST.

The same flaw returned in November 2025 inside DeFi, at a smaller scale. Stream Finance disclosed a loss of about 93 mn USD on 4 November 2025, and its stablecoin xUSD collapsed. Elixir's deUSD had lent much of its backing to Stream and fell with it. The two tokens were largely backing each other. A third, Stables Labs' USDX, broke below 0.60 USD days later, and none of the three recovered its peg.

The regulated era, 2024 to 2026

After 2023, the story moves from crashes to rules. The European Union's MiCA regulation (Markets in Crypto-Assets) applied its stablecoin rules from 30 June 2024 and required licensed issuers for tokens offered in the EU. Circle received a French e-money licence in July 2024 and brought USDC under MiCA. Tether did not seek a licence, and regulated EU exchanges removed USDT for European users. Coinbase's delisting took effect on 31 March 2025.

In the US, the GENIUS Act was signed on 18 July 2025. It creates a federal licence for payment stablecoin issuers, limits reserves to cash, short Treasury bills and similar assets, and bans issuers from paying interest to holders. By September 2026 the regulators had published proposals but no final rules, so the law takes effect on its backstop date of 18 January 2027. The rules covers both regimes in detail.

New designs arrived alongside the rules. PayPal launched PYUSD in August 2023. Ethena launched USDe, a synthetic dollar backed by a hedged crypto position, in February 2024. MakerDAO renamed itself Sky in August 2024 and launched USDS as DAI's successor.

Supply grew fast through this period and peaked at about 321 bn USD on 17 May 2026, according to DefiLlama. By early September it was about 3% lower. That is the first contraction in this history without a crash behind it. Higher interest rates on alternatives, money leaving DeFi and a move into tokenized Treasury funds have all been suggested as causes, and none is proven.

What to watch next

• Whether the 2026 dip turns into a trend: a plateau without a crash would mean growth now depends on real use rather than trading cycles. Where stablecoins stand in September 2026 tracks the numbers.
• The first final rules under the GENIUS Act: they decide which of today's issuers can keep serving US users after January 2027.
• Tether's US strategy: USDT's reserve includes gold and bitcoin, which the GENIUS Act does not allow, so Tether is launching a separate US token instead. How that token fares will show whether scale can survive strict reserve rules.
• Linked collateral in DeFi: the November 2025 cluster showed stablecoins backing each other again. Any new yield-bearing dollar whose backing sits mostly in another protocol repeats the BitUSD pattern.