California Defies Federal Blacklist to Partner with Anthropic for Statewide Claude Deployment

A Landmark Procurement Partnership for Claude

California has entered into a first-of-its-kind partnership with San Francisco-based artificial intelligence developer Anthropic to supply its generative AI assistant, Claude, to state, city, and county agencies. Under the agreement, public entities across the state can access Claude at a 50 percent discount. The partnership also includes free workforce training, expert technical assistance, and direct workflow input from Anthropic developers. State employees will use the productivity tool to summarize documents, analyze complex information, and optimize day-to-day operations. Nick Maduros, the Secretary of the Government Operations Agency, noted that the state aims to equip its teams with modern tools like Claude to improve service delivery, while Kate Jensen, Anthropic Head of Americas, expressed honor in putting their models to work for the people keeping California running.

Defying Federal Supply-Chain Restrictions

This partnership marks a significant escalation in California's efforts to shield its homegrown AI sector from federal policy. In early 2026, the United States Department of Defense designated Anthropic as a supply-chain risk and the Trump administration subsequently barred federal agencies from contracting with the company. This federal blacklisting followed a legal dispute in which Anthropic sued the federal government, alleging retaliation. The conflict began when Anthropic refused contract terms that would allow its AI models to be utilized for autonomous lethal warfare and the domestic surveillance of American citizens. In response to the federal ban, Governor Gavin Newsom signed an executive order directing California to conduct its own independent reviews of federal supply-chain risk labels. Under this order, the state will make its own procurement decisions, bypassing federal restrictions to support local startups and secure its position as a global technology leader.

A Strict Framework for Safety and Catastrophic Risk

While Governor Newsom has moved to defend local AI developers from federal actions, California has simultaneously established some of the nation's strictest regulatory guardrails. In September 2025, the governor signed landmark legislation targeting frontier AI models that meet high computational thresholds. Designed to prevent catastrophic events, the law mandates that developers publicly disclose safety protocols to ensure their systems cannot be used to shut down banking systems, hack power grids, or build bioweapons. The legislation defines a catastrophic risk as an event causing at least 1 billion dollars in damage or resulting in more than 50 injuries or deaths. Violations can carry fines of up to 1 million dollars, and companies are legally required to report critical safety incidents to the state within 15 days. The law also establishes whistleblower protections for tech sector employees and creates a public research cloud.

Preparing the Workforce for Automation

As state agencies accelerate their adoption of generative AI, California is also moving to address the economic and labor market disruptions caused by automation. In May 2026, Governor Newsom signed an executive order requiring state agencies to study labor shifts, including potential layoffs, changes in hiring practices, and emerging skills gaps. The order mandates that agencies deliver recommendations within 180 days on how to update California's WARN Act, which governs mass layoff notifications, to better protect workers affected by automation. This initiative builds on previous deployments of generative AI in state government. Since adopting the Request for Innovative Ideas procurement method in 2019, the state has launched several AI projects, including tools to reduce highway congestion, improve roadway safety, and enhance call center operations, as well as the California Poppy state government AI assistant. Additionally, a June 2026 report by Jake Brymner of the Institute for California AI Policy highlighted that local agencies are already using AI to optimize bus travel times, support social caseworkers, and predict housing instability for residents at risk of homelessness.

By bypassing federal blacklists to integrate Anthropic's models directly into local government, California is positioning itself not just as a regulator of artificial intelligence, but as an active geopolitical counterweight to federal tech policy.