Key facts
- New AI models are increasing the frequency and sophistication of cryptocurrency hacks.
- DeFi hacks in April 2026 resulted in over $634 million stolen, the highest monthly total in over a year.
- Immunefi CEO Mitchell Amador believes the crypto industry faces a critical three-to-four-year period for cybersecurity development.
- Amador suggests crowdsourced security solutions could accelerate the development of AI-powered defenses.
- A recent exploit on Kelp DAO's bridge resulted in approximately $290 million in stolen funds.
New artificial intelligence models are creating a significant cybersecurity challenge for the cryptocurrency industry, leading to a surge in decentralized finance (DeFi) hacks. Mitchell Amador, CEO of the bug bounty platform Immunefi, described this as a 'vulnerability apocalypse,' attributing the rise in illicit activities to the proliferation of advanced AI models like Claude Opus 4.8 and ChatGPT 5.5.
Hacking activity reached a peak in April 2026, with over $634 million stolen from crypto platforms, marking the highest monthly loss since February 2025. Amador emphasized that the next three to four years will be a critical survival period for the crypto industry as cybersecurity teams work to harness these same AI models for defense. He suggested that adopting crowdsourced security solutions could accelerate this process, potentially reducing the timeline to less than two years.
These concerns follow the release of Anthropic's latest Claude Mythos model, Fable 5, which raised industry worries about its potential to aid cryptocurrency exploits. Anthropic has stated that Fable 5 includes safeguards to reroute cybersecurity-related topics to other models, such as Claude Opus 4.8. The industry remains highly sensitive to security risks, particularly after recent major DeFi exploits, including one on April 19 where approximately $290 million to $293 million was drained from Kelp DAO's bridge. LayerZero identified a single point of failure in Kelp DAO's configuration as the cause.