Skip to main content

This site is for educational purposes only. Nothing here constitutes financial advice.

Back to all case studies
Depeg / Stablecoin Collapse
2022-05-09$40B+ market cap destroyed

TerraUSD (UST) & LUNA Collapse

May 2022 — algorithmic stablecoin death-spiral wipes out $40B+ across UST and LUNA in roughly 72 hours.

TerraUSD (UST) was, at its peak in early 2022, the third-largest stablecoin by market cap (~$18.7B) and one of the most-cited examples of 'algorithmic stability' working at scale. Its sister token LUNA traded at over $100, valuing the broader Terra ecosystem at approximately $40 billion.

Between May 7 and May 13, 2022, UST decoupled from the dollar and collapsed to under $0.01 while LUNA fell from over $80 to fractions of a cent — a roughly $40 billion market-cap evaporation in 72 hours. The mechanism was the failure of the arbitrage mint/burn system that was supposed to maintain the peg, which instead accelerated the collapse under stress.

The Terra collapse is the canonical case study of algorithmic stablecoin failure and remains the largest single-event loss of value in crypto history outside of broader market crashes.

Timeline

  1. 2020-09
    Terra launches UST stablecoin using mint/burn mechanism: 1 UST always mintable by burning $1 worth of LUNA and vice versa.
  2. 2021-03
    Anchor Protocol launches offering 19.5% APY on UST deposits. Becomes primary source of UST demand.
  3. 2022-02
    Luna Foundation Guard (LFG) raises $1B to back UST with BTC reserves.
  4. 2022-05-07
    Large UST sale on Curve's UST-3pool causes brief depeg to $0.98.
  5. 2022-05-08
    Withdrawal pressure from Anchor accelerates. UST briefly drops to $0.96.
  6. 2022-05-09
    LFG deploys ~$3B of BTC reserves to defend the peg. Reserves are depleted within hours without restoring confidence.
  7. 2022-05-10
    UST drops to $0.61. LUNA price collapses as arbitrageurs mint LUNA against cheap UST, but new LUNA floods the market.
  8. 2022-05-11
    Terra blockchain halts to prevent governance attacks. LUNA below $1.
  9. 2022-05-12
    Terra restarted; UST below $0.20. LUNA in hyperinflation as billions of new tokens are minted.
  10. 2022-05-13
    UST under $0.10; LUNA effectively zero. Total destruction of value across both tokens reaches ~$40B.
  11. 2022-05-28
    Terra hard-fork creates new chain (LUNA 2.0) without algorithmic stablecoin; original chain rebranded LUNA Classic.
  12. 2023-02
    SEC charges Terraform Labs and founder Do Kwon with fraud.
  13. 2023-03
    Do Kwon arrested in Montenegro on forged-passport charges; extradition proceedings begin.

Mechanism

The intended design: arbitrage-maintained peg. UST was an algorithmic stablecoin pegged to $1. Stability was supposed to be maintained by a mint/burn arbitrage: anyone could burn $1 worth of LUNA to mint 1 UST, or burn 1 UST to mint $1 worth of LUNA. If UST traded above $1, arbitrageurs would burn LUNA and mint UST to sell; if UST traded below $1, they would buy UST and burn it for LUNA. The system had no external collateral; LUNA itself was the absorption mechanism.

Anchor as demand engine. Anchor Protocol offered 19.5% APY on UST deposits — far above any sustainable rate. The yield was funded by LFG reserves (limited) and by emission of further LUNA (which depended on LUNA's price holding). Anchor created artificial UST demand that masked the underlying fragility. By May 2022, approximately 75% of all UST in existence sat in Anchor.

The trigger. On May 7, a series of large UST sales on Curve's UST-3pool (an automated market maker pool with UST and USDC/USDT/DAI) pushed UST briefly below $1. This triggered withdrawals from Anchor, which created additional sell pressure on UST. Whether the initial sales were a coordinated attack, a defensive de-risking by a large holder, or a sequence of independent decisions remains debated and may never be fully resolved.

The death spiral. When UST traded below $1, arbitrageurs began the intended mechanism: buy cheap UST, burn it for $1 of newly-minted LUNA, sell the LUNA. This flooded the market with new LUNA, driving LUNA's price down. As LUNA's price fell, more LUNA was needed to absorb each UST being redeemed, accelerating LUNA dilution. The mechanism that was supposed to stabilise the peg instead became self-reinforcing in the wrong direction. LFG's $3B BTC reserve was deployed to defend the peg but depleted within hours — finite reserves against infinite mint-and-sell pressure.

Why the reserves didn't help. LFG's roughly $3B in BTC and stablecoin reserves was, in early May 2022, less than 15% of UST's market cap. Once depleted (within roughly 24 hours of deployment), the algorithmic mechanism was the only thing left, and that mechanism actively accelerated the collapse. Critics had warned for over a year that the reserves were structurally insufficient; defenders argued the algorithmic mechanism would self-correct. The defenders were wrong.

Impact

Terra's collapse was the trigger event for the broader 2022 crypto credit crisis. Three Arrows Capital (3AC), a Singapore-based hedge fund heavily exposed to LUNA, collapsed in June 2022, defaulting on hundreds of millions in loans to Voyager, Genesis, BlockFi, and Celsius. Each of those entities subsequently failed or filed bankruptcy. The cascade continued through 2022, peaking with FTX in November. Regulatory responses were immediate: MiCA's stablecoin provisions were tightened to effectively prohibit algorithmic stablecoins from being offered to EU retail users; US state regulators (notably NYDFS) heightened scrutiny of stablecoin issuers; the SEC's enforcement actions against Terraform Labs set precedents for treating algorithmic stablecoins as securities.

Operational lessons

  1. 1Algorithmic stablecoins without external collateral have failed every time at scale. UST was not the first algorithmic stablecoin to collapse (Iron Finance preceded it in 2021), and the mechanism is structurally fragile. Future algorithmic stablecoin designs should be assumed to have the same failure mode unless they have external collateral.
  2. 2Unsustainable yield products fund themselves until they don't. Anchor's 19.5% APY was always going to end. The question was only when. Users who treated the yield as risk-free missed that the entire mechanism depended on LUNA's market value persisting.
  3. 3Reserves only work if they're proportionate to outstanding supply. LFG's $3B reserve against $18B of UST was 15% — enough for marketing, not enough for stress. Effective reserves for fiat-backed stablecoins approach 100% precisely because partial reserves don't survive a run.
  4. 4Death-spiral mechanics are visible in the design before the trigger. Critics had publicly identified the mint/burn arbitrage's failure mode for over a year before the collapse. The Terra defenders' confidence didn't change the mechanism's properties.
  5. 5Contagion timelines are measured in weeks. From the May 9 UST depeg to the June 2022 3AC default to the November 2022 FTX collapse was approximately 6 months. Financial-system contagion in crypto runs faster than in traditional finance.

Aftermath

Do Kwon was arrested in Montenegro in March 2023 on charges of using a forged passport. He was held there pending extradition disputes between the US and South Korea (which also sought him on capital-markets charges). He was eventually extradited to the US in late 2024. The SEC's civil case against Terraform Labs resulted in a April 2024 trial verdict finding Terraform Labs and Do Kwon liable for fraud and ordering disgorgement of approximately $4.5 billion. The Terra blockchain itself continues to exist as LUNA Classic (still trading at tiny fractions of a cent) and LUNA 2.0 (a non-algorithmic-stablecoin successor with limited traction). The 'do not invest in algorithmic stablecoins' lesson became one of the few areas of broad market consensus.

Sources & further reading

We prioritise primary sources. Where a topic moves quickly (regulation, security incidents), we re-check sources on the cadence shown by the page's "Next review" date.

Related on Block Clarity Hub