Why does Immunefi's CEO believe AI is increasing DeFi hacks?

Immunefi CEO Mitchell Amador argues that advanced AI models are making it easier for attackers to identify software vulnerabilities and develop exploits. According to Amador, these tools have shifted the cybersecurity advantage toward hackers, contributing to a rise in DeFi attacks during 2026.

What is the "vulnerability apocalypse" in crypto?

The term refers to a period where increasingly powerful AI systems can rapidly discover and exploit security weaknesses in blockchain applications. Amador believes the crypto industry is entering a challenging phase where attackers can use AI faster than defenders can adapt.

How much crypto was stolen in April 2026?

More than $634 million was stolen from cryptocurrency platforms during April 2026. According to the report, this was the highest monthly total since the major wave of losses linked to the Bybit hack in February 2025.

Which AI models were cited as contributing to the threat?

Amador specifically referenced Claude Opus 4.8 and ChatGPT 5.5 as examples of increasingly capable AI systems. He argues that frontier AI models are significantly improving the ability of attackers to find vulnerabilities in smart contracts and blockchain infrastructure.

Why is DeFi particularly vulnerable to AI-assisted attacks?

DeFi protocols rely heavily on smart contracts that manage large amounts of capital. AI tools can help attackers analyze code, identify weaknesses and automate parts of the exploit discovery process, potentially increasing the speed and scale of attacks.

How long could the crypto industry face elevated security risks?

Amador estimates the industry faces a critical two-to-four-year period before defenders can fully leverage AI to build more secure systems. He believes security teams eventually will use the same technology to strengthen blockchain infrastructure.

Image for illustrative purposes only. Not a real photo.

AI blamed for surge in DeFi security breaches

Written by Heidi Cuthbert Published Jun 12, 2026, 1:59 AM

Advanced artificial intelligence models have created a “vulnerability apocalypse” for the cryptocurrency industry by making it easier for attackers to identify and exploit weaknesses in decentralised finance protocols, according to Immunefi chief executive Mitchell Amador.

Amador said the emergence of increasingly capable AI systems, including Claude Opus 4.8 and ChatGPT 5.5, has shifted the cybersecurity landscape in favour of hackers and contributed to a resurgence in crypto-related exploits during 2026.

“Crypto needs to survive the next three to four years,”

Amador said, warning that the industry remains vulnerable until security teams can use the same AI technologies to build significantly more resilient codebases.

Cryptocurrency hacking activity surged in April, with more than $634 million stolen from digital asset platforms, the highest monthly total since the major Bybit-related losses recorded in February 2025.

Amador said the timeline for improving security could be shortened to less than two years if the industry embraces more crowdsourced security initiatives and bug bounty programmes.

His comments followed the release of Anthropic's latest Claude Mythos model, Fable 5, which prompted debate within the crypto sector over whether increasingly powerful AI systems could accelerate the discovery of exploitable vulnerabilities.

Security concerns have intensified following several major DeFi exploits, including an April attack on Kelp DAO's LayerZero-powered rsETH bridge that resulted in losses of approximately $290 million.

LayerZero said the incident stemmed from Kelp DAO's use of a single verifier path for cross-chain messages, creating a single point of failure despite previous warnings against that configuration.