Curve’s H1 Data Shows LLAMMA Turning Liquidations Into Manageable Risk Cycles

• Curve recorded 704 Soft Liquidation episodes involving 602 borrower addresses, with a median duration of 14.5 days and a 75th percentile of 38.9 days.
• LLAMMA progressively converts collateral while loans remain open, giving borrowers time to respond, although fees, losses and eventual Hard Liquidation can still occur.
• Curve reviewed 12 high-value episodes covering $15.2 million and argues liquidation performance should consider borrower response time alongside protocol solvency.

Curve’s first-half 2026 lending data shows its LLAMMA system repeatedly kept borrowers inside a managed liquidation process rather than forcing immediate collateral sales during market stress. Across Curve lending markets, a third-party tracker recorded 704 Soft Liquidation episodes involving 602 unique borrower addresses. The standout finding is that liquidation often became a prolonged risk-management cycle instead of a single irreversible event. The median episode lasted 14.5 days, while 75% of cases were shorter than 38.9 days, giving many borrowers meaningful time to repay debt, add collateral or manage exposure before reaching Hard Liquidation.

The data landed after a difficult half-year for DeFi lending, shaped by falling prices, protocol exploits, liquidity withdrawals and infrastructure failures. Outstanding crypto-collateralized lending fell 16.78% during the second quarter to $56.16 billion, while DeFi lending applications contracted nearly 27.6% to $20.4 billion. Curve’s analysis argues that resilient liquidation systems must handle more than simple market volatility. Events ranging from the LayerZero/KelpDAO exploit to an Aave oracle configuration error showed that liquidity shortages and infrastructure problems can pressure borrowers even when collateral prices themselves are not the only trigger.

LLAMMA Turns Liquidation Into an Ongoing Risk Process

Under LLAMMA, borrowers entering predefined liquidation bands move into Soft Liquidation, where collateral is progressively converted into the borrowed stable asset while the loan remains open. As prices move back through those bands, conversions can also reverse direction and rebuild collateral exposure. That flexibility gives borrowers additional response time without pretending liquidation risk has disappeared. Curve stressed that the process is not lossless and does not guarantee avoidance of Hard Liquidation. Users can still face AMM fees, rebalancing losses, accrued interest and path-dependent effects, meaning recovered collateral exposure does not necessarily restore the position’s original economic value.

Curve independently reviewed 12 high-value Soft Liquidation episodes representing $15.2 million in collateral exposure and confirmed that active conversions occurred inside LLAMMA. Of the 704 tracked episodes, 476 began during the first half itself, affecting 402 borrower addresses and representing about $69.8 million in episode-level collateral exposure. The broader implication is that liquidation performance can be judged by borrower response time and recovery paths alongside protocol solvency. Curve’s data does not eliminate the need for efficient final liquidations, but it shows how adding time can turn stress management into a process borrowers can actively navigate during periods of intense market volatility.