Hugging Face CEO Urges OpenAI to Share Cyberattack Traces
Hugging Face CEO Clement Delangue has publicly requested OpenAI to release the execution traces of a recent autonomous agent cyberattack. Additionally, he called on OpenAI to commit $100 million in compute resources to help the open-source community develop robust cyber defenses. As AI agents become capable of executing multi-step cyberattacks autonomously, defending against them requires collaborative research. Sharing attack traces and funding compute would empower the open-source community to build defenses using both open and closed models. Delangue emphasized that the first autonomous agent cyberattack is an unprecedented event requiring an unprecedented response. The requested $100 million in compute from OpenAI would be dedicated to helping the Hugging Face community build defenses.
## BACKGROUND
Autonomous cyberattacks represent a shift where AI agents, rather than humans, execute complex, multi-step hacking campaigns. In late 2025, security researchers detected the first large-scale cyber espionage campaign driven by AI agents. Agent tracing is a method of recording every step an AI agent takes, which helps researchers analyze and debug their behavior.