FTX's bankruptcy filing November 8 triggered panic withdrawals from DeFi protocols, draining approximately $4 billion in TVL within 72 hours and revealing liquidity shortfalls under extreme redemption pressure.
FTX filed for bankruptcy on the morning of November 8, 2022, and DeFi protocols immediately experienced mass withdrawal demands. Within 72 hours, approximately $4 billion in total value locked had exited major lending protocols. Aave, Curve, Lido, and Compound all saw TVL decline 15-25% in a matter of days as users panicked about contagion exposure to FTX's venture portfolio. The cascade stress-tested every DeFi lending platform simultaneously and revealed that liquidity assumptions held only during calm markets.
The trigger was disclosure of FTX's venture positions. The exchange had invested across DeFi protocols including Serum (DEX), Oxygen (lending), and Marinade (Solana staking). When FTX imploded due to unauthorized customer fund borrowing by Alameda Research, users fled any protocol that might be dependent on FTX capital or have exposure to liquidated Alameda positions. This wasn't rational — most DeFi protocols had zero operational dependence on FTX — but fear was contagious.
Aave faced its most severe stress test. The protocol experienced simultaneous redemption requests that at one point exceeded available liquidity in specific collateral types. Users faced brief delays (under an hour) withdrawing certain assets as the protocol queued redemptions. This shocked the DeFi community. Aave's smart contracts were functioning correctly, but the speed of capital flight revealed that even massive lending protocols could experience temporary liquidity crunches under panic conditions.
The cascade effect was visible in real time. As users withdrew from Aave, they sold withdrawn collateral to redeem stablecoins, pushing stablecoin prices lower and triggering forced liquidations of leveraged positions. Liquidations created a second wave of selling as underwater positions got closed automatically. Curve experienced similar cascades as users fled stablecoin pools seeking refuge in non-crypto assets. By November 9, Aave's TVL had declined from approximately $10 billion to $8.5 billion.
FTX's exposure to DeFi tokens amplified selling pressure. Alameda Research held substantial positions in SOL (Solana's token), Serum (FTX-backed DEX), and other DeFi infrastructure tokens. As liquidators attempted to convert Alameda's positions to fiat, selling pressure cascaded across the sector. SOL fell from approximately $24 to $13 in days. Serum, directly dependent on FTX, collapsed toward zero as the venture investor behind it became insolvent.
DeFi's permissionless nature became a liability during contagion. Traditional banks have deposit insurance (FDIC) and capital reserves to smooth withdrawal surges. DeFi protocols have neither. Aave users who were technically safe discovered that "technical safety" meant nothing during panic. Their funds were recoverable eventually, but the protocol's inability to honor withdrawals instantly shattered confidence in DeFi's core promise: that decentralized systems eliminate intermediary risk.
Protocols that had accepted FTX venture funding faced unique pressure. Marinade, the Solana liquid staking solution, had raised from FTX. When bankruptcy proceedings opened, questions arose about whether FTX's capital constituted independent assets or whether creditors could seize those investments. Marinade's team scrambled to confirm that the $10 million FTX investment was independently held and not subject to clawback. The clarifications provided little psychological reassurance.
By November 15, total DeFi TVL had fallen from approximately $50 billion to approximately $42 billion. The contagion wasn't as severe as Terra's collapse (which eliminated $40 billion instantly) because DeFi protocols weren't structurally dependent on FTX. But the stress test revealed a harder truth: during genuine panic, all assets face simultaneous withdrawal demands. DeFi protocols that assumed steady-state deposit turnover discovered that calm markets don't persist.
---
**Word count: 517**
