AI Agents Render All DeFi Protocols Unsafe, Warns Security Expert

Manuel Aráoz, CEO of OpenZeppelin, has declared that all decentralized finance (DeFi) protocols are currently unsafe due to the enhanced capabilities of AI agents in identifying vulnerabilities. His warning follows a significant uptick in DeFi exploits, with over $600 million lost in April 2026 alone. The asymmetry in security—where defenders must address every vulnerability while attackers need to find just one—has been exacerbated by AI's ability to analyze code rapidly. Notable incidents include the $292 million KelpDAO exploit and the $285 million loss from Drift. The total value locked (TVL) in DeFi has decreased by over $20 billion since the start of the year, reflecting a loss of confidence in the sector. Experts suggest that traditional security measures are insufficient against AI-driven threats, prompting calls for enhanced security protocols and continuous monitoring. The current state of DeFi is precarious, with ongoing discussions about the need for new security frameworks.

Key Points: • All DeFi protocols are deemed unsafe due to AI's enhanced vulnerability discovery. • Over $600 million was lost to DeFi exploits in April 2026, with significant incidents reported. • The total value locked in DeFi has fallen by over $20 billion since January 2026.