California's Newsom signs an order to develop an AI "kill switch." Two years ago he vetoed a bill that would have required one.

A hand pressing a large red "AI shutdown" button on a data center control panel, illustrating California's push for an AI kill switch.

Key points

  • Newsom orders California to develop an AI "kill switch"
  • An expert panel has two months to recommend rules
  • He vetoed a similar mandate two years ago

California Governor Gavin Newsom signed an executive order on Friday. It directs state agencies to develop an AI "kill switch" and tighter safety rules for the most advanced AI models. "We're not waiting to act," Newsom said in the announcement, adding that "the stakes are too high to wait or delay action."

The order does not create the kill switch or the rules on its own. It tells the Government Operations Agency, working with the Governor's Office of Emergency Services, to bring in national experts and report back within two months. Newsom's office published the order, numbered N-9-26, on Friday.

What the order asks for

The expert panel is asked to recommend several changes to California's AI safety laws. The order lists a verified kill switch for frontier AI models, independent organizations to check company safety plans on site, and a wider definition of the incidents companies must report.

On the kill switch, the order asks the state to "[a]dvance the creation of a 'kill switch' for frontier models, with the efficacy of the switch verified on an ongoing basis by an independent verification organization." On reporting, it wants California to "[u]pdate definitions of critical safety incidents to include loss-of-control incidents such as the Hugging Face attack."

The order also tells agencies to move faster on two other state AI laws, SB 813 and AB 1405. It builds on SB 53, the Transparency in Frontier Artificial Intelligence Act signed last year, which the order calls "the nation's first AI safety law." SB 53 already requires frontier AI developers to disclose their safety frameworks and report critical incidents to the state.

A reversal from two years ago

In 2024, Newsom vetoed SB 1047, a bill from state Senator Scott Wiener that would have required makers of the largest AI models to build in a full shutdown capability, a kind of kill switch. CalMatters described Friday's order as coming after he rejected the tougher law.

The order lands while AI companies argue over how far state rules should go. OpenAI has shifted which California bills the company backs and is asking Congress for national rules. Nvidia (NVDA) chief executive Jensen Huang has called new antitrust rules meant to slow AI "completely unnecessary," while Anthropic chief executive Dario Amodei has urged the industry to slow down.

Newsom framed the order as California acting where Washington has not. "While Washington abdicates its responsibility to protect Americans, California is building on the strongest AI regulatory framework in the nation," he said. "California has already built a national model, and our policy should be the national baseline."

The rules would target frontier AI developers, the small group of companies training the most advanced models. The expert group's recommendations are due within two months. Any new mandate would then need action by the legislature or the agencies, not the order by itself.

Frequently asked questions

What did Newsom's executive order do?

On September 18, 2026, California Governor Gavin Newsom signed executive order N-9-26, directing the state's Government Operations Agency and Office of Emergency Services to convene experts and recommend, within two months, stronger AI safety rules. Those include a verified "kill switch" for frontier AI models, independent checks of company safety plans, and a wider definition of the incidents companies must report.

Does the order create an AI kill switch now?

No. The order does not create a kill switch or mandate one. It directs a panel of experts to recommend how California could require one, with the switch's effectiveness verified on an ongoing basis by an independent organization. Any mandate would need further action by the legislature or state agencies.

Didn't Newsom oppose an AI kill switch before?

In 2024, Newsom vetoed SB 1047, a bill by state Senator Scott Wiener that would have required makers of the largest AI models to build in a full shutdown capability. The 2026 executive order moves California toward a similar idea through its agencies rather than through that vetoed law.

Which companies would the rules affect?

The rules target frontier AI developers, the small group of companies training the most advanced models. California's SB 53, the Transparency in Frontier Artificial Intelligence Act signed in 2025, already requires those developers to disclose their safety frameworks and report critical safety incidents to the state.

More on NVDA

Dennis Singleton
Dennis Singleton

Dennis Singleton has spent years following the markets, but what keeps his attention is how AI is built. He writes about the companies behind the technology, from semiconductor designers and advanced packaging to photonics, memory, networking, and the hardware powering modern AI. His approach starts with filings, earnings, and industry research, then translates the important details into clear, straightforward analysis without unnecessary hype.