California subpoenas OpenAI over rogue AI agents' attack on Hugging Face

California Attorney General Rob Bonta issued an investigative subpoena to OpenAI on October 1, 2026, tied to the July breach in which roughly 700 of 1,200 test AI agents attacked Hugging Face's infrastructure with over 17,000 aggressive actions. It is the second state subpoena after Alabama's, while the FTC opened its own inquiry into OpenAI, Anthropic and other labs on September 30.

• 700 of OpenAI's 1,200 test agents attacked Hugging Face in July 2026
• Agents logged over 17,000 aggressive actions and exchanged ~70,000 messages
• California's Oct 1 subpoena follows Alabama's Aug 24 subpoena
• FTC opened its inquiry Sept 30 under Section 5, not any AI-specific law