Manuel Aráoz, the co-founder of OpenZeppelin, one of crypto's best-known smart contract security firms, says he now considers "all of DeFi unsafe" since AI coding agents can detect bugs faster than developers can fix them.
Aráoz wrote in an X post on May 27 that he has "privately advising friends and family to exit all DeFi positions including low-risk 'blue chips' like Aave, MakerDAO & Compound."
The OpenZeppelin co-founder argues that DeFi is hard to defend because attackers need to find only one weakness, while teams have to catch every one before funds are drained.
Later, Aráoz clarified in another X post that he wasn't only talking about smart contract code. He said the broader problem is security, including parameter settings, mechanism design and operational security, adding that AI coding agents are also "superhuman" at finding those kinds of vulnerabilities.
- DeFi is under heavier scrutiny than centralized crypto firms because its code is public and often directly controlled by smart contracts.
- That gives researchers and users more visibility, but it also gives attackers a public map of what to study.
- Blockchain forensic firm Chainalysis, for instance, has repeatedly pointed to DeFi as a large target for hackers, while the European Banking Authority has warned that DeFi composability can let a weakness in one smart contract spread across several protocols.
There's no indication that Aave, Compound or MakerDAO has a newly discovered vulnerability tied to Aráoz's warning. Yet, some of those protocols have faced stress before.
- MakerDAO's "Black Thursday" crisis in March 2020 left about $4.5 million of unbacked DAI in the system after liquidation auctions failed during a market crash.
- Compound had a 2021 bug that wrongly distributed tens of millions of dollars in COMP rewards.
- Aave came under pressure in April of this year after hackers used stolen KelpDAO-linked tokens as collateral on Aave, leaving the protocol with about $230 million in bad debt.
The broader DeFi market has also been hit by fresh security concerns. In May alone, nearly $14 billion was reportedly pulled from DeFi after several hacking incidents, adding weight to warnings that yields across DeFi protocols may simply not be high enough to compensate users for the risks.
OpenZeppelin pushed back on Aráoz's warning, saying in an X thread that his comments on AI and DeFi security don't reflect the company's position and noting that he co-founded OpenZeppelin but left in 2019 after serving as CTO.
OpenZeppelin also added that most recent incidents came from operational security failures rather than smart contract bugs, and argued that the answer to AI-driven risk is stronger security, not a retreat from DeFi.
More context on lending: DeFi lending drops 51% from peak as Q1 borrows fall 14%
Update 5:27pm UTC. Added comments from OpenZeppelin