California Enacts First State Law Regulating Frontier AI Safety

On September 29, 2025, Gavin Newsom, the Governor of California, signed Senate Bill 53, also known as the Transparency in Frontier Artificial Intelligence Act (TFAIA). Effective January 1, 2026, the law establishes the first state-level safety framework for frontier AI in the United States. It targets developers of high-power models trained with more than 10^26 floating-point operations or those with annual gross revenues exceeding $500 million.

Under the act, large frontier developers must publish transparency reports and a frontier AI framework detailing cybersecurity practices and assessments of catastrophic risks, defined as events causing over 50 deaths or $1 billion in damage. The law mandates that critical safety incidents be reported to the California Governor's Office of Emergency Services within 15 days. To support research and reduce the dominance of major labs, the state is establishing CalCompute, a public cloud computing cluster.

Noncompliance may result in civil penalties of up to $1 million per violation, enforced by the California Attorney General's Office. The legislation includes whistleblower protections for employees and follows a previous veto by Newsom of a broader AI bill he feared would stifle innovation. While the law received endorsement from Anthropic, it faced opposition from the Chamber of Progress and other industry groups who argue that a patchwork of state regulations hinders innovation and should be managed at the federal level.