Newsom Signs First-in-Nation Frontier AI Safety Law in California

Governor Gavin Newsom signed Senate Bill 53, the Transparency in Frontier Artificial Intelligence Act (TFAIA), on September 29, 2025. The legislation makes California the first state to regulate frontier AI development, specifically targeting models trained with more than 10^26 floating-point operations or developers with annual gross revenues exceeding $500 million.

The law mandates that frontier developers publish transparency reports and disclose critical safety incidents to the California Governor's Office of Emergency Services. Large developers must additionally publish a frontier AI framework detailing cybersecurity practices and assessments of catastrophic risks. The act also establishes CalCompute, a state-run public cloud computing cluster designed to provide infrastructure and research resources to reduce asymmetries favoring dominant AI labs.

SB 53 follows a previous veto by Newsom of a broader version of the law to avoid stifling innovation. The current framework includes whistleblower protections and allows the California Attorney General to enforce civil penalties of up to $1 million per violation. While supported by companies like Anthropic, the law faced opposition from the Chamber of Progress, which argued it could hinder industry growth. The law takes effect on January 1, 2026, positioning California's regulatory approach in contrast to the deregulatory policies of the Trump Administration.