Between May 8 and May 12, 2022, the Terra ecosystem suffered the largest single-protocol value destruction in cryptocurrency history, with the algorithmic stablecoin TerraUSD (UST) depegging from $1 to under $0.10 and the sister token Luna collapsing from approximately $80 to fractions of a cent, erasing roughly $40B of market capitalization in 96 hours and cascading into the bankruptcies of Three Arrows Capital, Voyager Digital, and Celsius Network over the following two months. The proximate trigger was a sequence of large UST withdrawals from Curve Finance's 3pool and from Anchor Protocol, which broke the peg by approximately 20 basis points and induced reflexive selling of Luna as arbitrageurs and panicked holders converted UST through the protocol's mint-and-burn mechanism. The deeper cause was the structural impossibility of Terra's design: a stablecoin whose stability rested on the market value of a sister token whose value was itself largely a bet on the stablecoin's stability. The Luna Foundation Guard's $3.5B Bitcoin reserve, deployed in defense of the peg on May 9, was insufficient to halt the spiral and was itself absorbed into the collapse. Do Kwon was eventually arrested in Montenegro in March 2023, extradited to the United States on December 31, 2024, pleaded guilty to fraud in August 2025, and was sentenced to 15 years in prison in December 2025.
Timeline of events
On the morning of May 7, 2022, the Terra ecosystem appeared healthy: UST circulating supply was approximately $18.7B, Luna's market cap was approximately $25B, and Anchor Protocol held approximately $14B in UST deposits paying a fixed yield of 19.5%. Beginning at approximately 21:00 UTC on May 7, a sequence of large UST sales hit the Curve 3pool (UST paired with USDT, USDC, and DAI), totaling approximately $285M in net UST sold over roughly two hours; the source of these sells has been the subject of substantial subsequent forensic work and remains partly contested, with Terraform Labs alleging coordinated short-attack and outside investigators including Nansen attributing the flow to seven specific wallets without identifying their ultimate beneficiaries. The 3pool, which was thinly liquid relative to UST's circulating supply, depegged to approximately $0.985 by 23:00 UTC. Anchor Protocol depositors, observing the depeg on Curve, began withdrawing UST at scale; Anchor TVL fell from $14B to $11.2B in the next twelve hours. Each Anchor withdrawal returned UST to the open market, deepening the depeg. By May 9, UST traded at $0.60-$0.70 with extreme volatility; the Luna Foundation Guard announced and executed a $1.5B Bitcoin sale to defend the peg; the defense produced a temporary recovery to $0.93 before the second wave of selling broke the peg permanently. By May 10, UST was at $0.30; by May 11, at $0.13. Luna, which had begun the week at $80, was at $1.20 on May 11 and at $0.0001 by May 13 as the protocol's elastic-supply mint-and-burn mechanism produced effectively unlimited Luna issuance against UST redemptions. The Terra blockchain was halted twice on May 12 to prevent governance attacks against the now-near-zero-value validator set.
Attack mechanism
The mechanism is a textbook reflexive death spiral, a failure mode that academics and skeptical practitioners had described in detail for years prior. UST's stability was guaranteed by an arbitrage mechanism: any holder of UST could burn 1 UST and mint $1 worth of Luna, regardless of Luna's market price. So long as Luna had non-trivial market depth, this mechanism would absorb UST sell-pressure: a UST seller would burn UST to mint Luna and sell the Luna for dollars, with arbitrageurs ensuring the round-trip held the UST price near $1. The failure mode is straightforward: as Luna's price falls, the mint-and-burn mechanism issues progressively more Luna per UST redeemed, increasing Luna supply and further depressing Luna's price; this in turn requires still more Luna issuance per subsequent UST redemption, in a feedback loop that has no equilibrium below a threshold. The threshold, sometimes called the bank-run point, depends on Luna's available market depth versus the UST supply potentially seeking exit; in May 2022, with Luna's market cap at roughly $25B against UST's $18.7B and with Curve and CEX market depth far thinner than those headline numbers, the threshold was substantially higher than the protocol's defenders publicly acknowledged. Once depeg pressure exceeded the absorbable depth, the spiral was deterministic. The Anchor Protocol layer accelerated the spiral: Anchor's 19.5% fixed yield, far above sustainable real-rate levels, meant that approximately 75% of all UST was held in Anchor as a yield-bearing position rather than as transactional medium of exchange, so depeg-induced withdrawals translated into immediate sell-pressure rather than into rebalancing of payment-rail balances.
Root cause analysis
The root cause is the structural incoherence of an algorithmic stablecoin whose stability rests on the market value of a sister token whose value rests substantially on the stablecoin's stability. This is not a contingent design defect that could have been patched; it is a category error. Stablecoins maintain their peg through one of three mechanisms: (1) full fiat collateral held in trust (USDC, USDT broadly), (2) over-collateralization in liquid crypto assets (DAI, sUSD, LUSD), or (3) a hybrid of the above plus market-making and seigniorage. UST's mechanism was a fourth class - pure algorithmic, sister-token-burn - that had failed in every prior implementation (Basis, Empty Set Dollar, IRON/TITAN). The Terra team's defense, articulated publicly by Do Kwon, was that Terra's scale and Luna's deeper liquidity changed the calculus; the May 2022 collapse demonstrated empirically that scale amplified rather than mitigated the reflexive failure. A secondary root cause was Anchor's unsustainable yield: 19.5% fixed against borrower demand that paid materially less, with the gap subsidized by a Yield Reserve that the Terraform Labs treasury had to repeatedly top up. Anchor had become a de facto demand-aggregator for UST, and its yield was a marketing expense for UST adoption rather than a reflection of underlying lending demand. When Anchor's yield-reserve sustainability came into question (a subject of public discussion as early as late 2021), the implicit promise underwriting UST's stability was already failing.
Initial response and recovery
There was no recovery. Terraform Labs' first public response on May 9 was the Luna Foundation Guard's deployment of approximately $1.5B in Bitcoin reserves, sold on Binance and elsewhere to buy UST in an attempt to restore the peg; this defense achieved a temporary partial restoration before being exhausted. The LFG subsequently disclosed that approximately 80,000 BTC (a substantial fraction of the announced reserve) had been deployed, leaving the foundation with negligible remaining capacity. Do Kwon's second response was a proposal, eventually executed on May 28, 2022, to fork the Terra blockchain, distinguishing the original chain (renamed Terra Classic, ticker LUNC) from a new chain (Terra 2.0, ticker LUNA) on which a fresh airdrop would attempt to compensate prior holders pro-rata. The new chain launched but achieved minimal traction; the Terra Classic chain continued to exist with token prices reflecting effectively-zero economic value. UST holders received no protocol-level recovery; their losses were total. The cascade through to lender insolvencies began within weeks: Three Arrows Capital, which held substantial Luna and UST exposure and had borrowed heavily against it from BlockFi, Genesis, Voyager, and others, defaulted on margin calls in mid-June and filed Chapter 15 in early July; Voyager Digital filed Chapter 11 on July 5; Celsius Network filed Chapter 11 on July 13. The pattern of cascading failures created the broader 2022 crypto winter that culminated in FTX's November collapse.
Funds tracking and laundering
Unlike a hack post-mortem, Terra's collapse did not involve theft of identifiable funds; the value destruction was largely market-cap evaporation as Luna's supply expanded by orders of magnitude in 96 hours. There were, however, substantial flows worth tracking. The Luna Foundation Guard's Bitcoin reserve - which the foundation had publicly committed to as a peg-defense mechanism - was sold during the crisis; subsequent forensic work by on-chain investigator ZachXBT and by Chainalysis confirmed the headline disposition but raised contested questions about timing and beneficiaries of certain transfers. Separately, allegations have persisted that early Terra-affiliated insiders, including some Terraform Labs principals, exited UST and Luna positions in advance of the collapse; the South Korean prosecutors investigating Do Kwon assembled detailed wallet-level timing evidence on this point, which formed part of the criminal case against him. Anchor Protocol depositors who exited in the first 24 hours of the depeg - approximately $3-4B in withdrawals - generally received UST that they could swap for stablecoins at gradually-deteriorating rates; depositors who waited beyond approximately May 10 generally took total losses. There has been no recovery of value into the holder community in any structural sense; the LUNA 2.0 airdrop produced tokens whose initial trading value was a small fraction of pre-collapse holdings and whose subsequent performance has been broadly downward.
Legal and regulatory aftermath
Do Kwon was the subject of multiple parallel investigations beginning within weeks of the collapse. South Korean prosecutors obtained an arrest warrant in September 2022; an Interpol Red Notice followed in late September. Kwon left South Korea before the warrant could be served; his subsequent movements - through Singapore, the UAE, and ultimately Serbia - were the subject of substantial public reporting. He was arrested at Podgorica airport in Montenegro on March 23, 2023, attempting to board a flight to Dubai using a falsified Costa Rican passport. After an extended extradition contest in which both South Korea and the United States sought transfer, Montenegro ultimately extradited Kwon to the United States on December 31, 2024, where he faced federal charges including securities fraud, wire fraud, and conspiracy. On August 12, 2025, Kwon pleaded guilty in the Southern District of New York to one count of conspiracy to commit commodities, securities, and wire fraud and one count of wire fraud, agreeing to forfeit over $19M; on December 11, 2025, Judge Paul Engelmayer sentenced him to 15 years in prison - above the 12 years prosecutors had sought - with the possibility of applying for transfer to South Korea, where he faces further charges, after serving at least half the sentence. Separately, a New York jury found Terraform Labs and Kwon liable for civil fraud in April 2024 in the SEC's case, and a final judgment of approximately $4.5B was entered in June 2024 as part of a settlement; the judgment is largely uncollectable given the bankruptcy of Terraform Labs. Daniel Shin, Terraform Labs co-founder, faced parallel charges in South Korea; his trial proceeded separately. The case has become a touchstone for cross-jurisdictional white-collar crypto prosecution.
Industry implications
Terra's collapse reshaped the stablecoin and DeFi landscape across four dimensions. First, it functionally ended the algorithmic-stablecoin design space as a venue for serious capital allocation; subsequent algorithmic projects (FRAX moved toward fully-collateralized; Synthetix moved away from sUSD as a primary product) have transitioned away from pure-algorithmic mechanisms, and new launches such as crvUSD and GHO are over-collateralized by design. Second, it accelerated regulatory attention on stablecoins specifically, with the U.S. GENIUS Act (enacted July 2025), the EU MiCA framework's stablecoin chapters, and various Asian jurisdictions' licensing regimes all reflecting post-Terra concerns about reserve adequacy and redemption mechanism design. Third, it triggered the cascade of credit-fund and lender failures (Three Arrows, Voyager, Celsius, BlockFi, Genesis) that defined the 2022 crypto winter and ultimately bled into FTX's collapse, in a chain of contagion that traced largely back to over-collateralized lending against Luna and to Anchor-yield-driven deposit flows. Fourth, it ended the Anchor-style fixed-yield-marketing era; subsequent yield products in DeFi have been variable and tied to underlying real-borrower demand, with limited exceptions. The deeper consequence has been a recalibration of how the industry talks about decentralized stability: the post-Terra consensus is that pegging fiat-denominated value to a crypto-native asset whose value is itself a bet on the peg is a structural impossibility, full stop, and that any future stablecoin design must rest on either fiat collateral or substantial over-collateralization in liquid, exogenous crypto assets.
Verdict and lessons
The Terra collapse is the empirical demonstration of a result that economists and skeptical crypto practitioners had argued analytically for years: an algorithmic stablecoin backed by a sister-token-burn mechanism cannot maintain its peg under sufficient stress, because the mechanism's stabilizing feedback inverts into destabilizing feedback below a threshold that is structural rather than contingent. The lesson is not that algorithmic stablecoins are difficult; it is that they are impossible in the form Terra implemented, and that a design space that has failed in every prior instantiation should not be expected to succeed at scale. For practitioners, the operational lessons are: stablecoins should be treated according to their collateralization mechanism and their reserve transparency, with algorithmic and partial-collateral designs assigned a meaningful risk premium even when their on-chain price holds; deposit yields materially above the prevailing real rate should be assumed to reflect either a marketing subsidy or a hidden risk and should not be used as a basis for size; and exposure concentration in any single ecosystem - particularly one that combines a stablecoin, a yield product, and a sister-token in mutual dependency - should be capped at a fraction of what a comparable diversified position would justify. For the industry, Terra established the working consensus that the post-2022 stablecoin design space contains exactly two viable categories: fiat-collateralized (with reserve transparency) and significantly over-collateralized in liquid crypto, with everything else regarded as experimental until proven otherwise.
Root cause
TerraUSD's algorithmic stability mechanism rested on a sister-token-burn arbitrage with Luna; this design contained a reflexive failure mode in which depeg pressure beyond Luna's available market depth produced unbounded Luna supply expansion, collapsing both tokens. Anchor Protocol's unsustainable 19.5% fixed yield concentrated approximately 75% of UST supply in a single yield-aggregator, converting depeg events into immediate exit-flow rather than rebalancing.
Recovery and aftermath
No protocol-level recovery for UST/Luna holders. Terra 2.0 fork (May 28, 2022) airdropped new LUNA tokens to prior holders pro-rata, but the new chain's market cap stabilized at a small fraction of prior levels. A jury found Terraform Labs and Do Kwon liable in the SEC's civil fraud case in April 2024, and an approximately $4.5B judgment was entered in June 2024, largely uncollectable. Kwon pleaded guilty to criminal fraud in August 2025 and was sentenced to 15 years in prison in December 2025. Cascading lender bankruptcies (Voyager, Celsius, BlockFi, Genesis) produced partial creditor recoveries via Chapter 11 over 2022-2024.
Lessons
- An algorithmic stablecoin pegged via sister-token burn contains a reflexive failure mode that becomes deterministic under sufficient stress; this is structural, not contingent
- Yields materially above the prevailing real rate signal either marketing subsidy or hidden risk and should not be the basis for size
- Reserve adequacy must be measured against potential exit-flow, not against headline circulating supply; thin liquidity in stress is the binding constraint
- Concentration in mutually-dependent token-stablecoin-yield trios creates correlated tail risk that diversified-exposure rules of thumb materially underestimate
Precedent
Functionally ended algorithmic stablecoins as a serious capital allocation venue; reshaped global stablecoin regulation (the U.S. GENIUS Act, MiCA stablecoin chapters); triggered the 2022 crypto winter that culminated in FTX. Established a working post-2022 consensus that stablecoins must be either fiat-collateralized with reserve transparency or significantly over-collateralized in liquid crypto.