California signs deal for discounted access to Anthropic's Claude

The California government has established a partnership with Anthropic to provide state agencies with discounted access to its AI chatbot, Claude. Under the agreement, California government agencies will have access to Claude at half price, with Anthropic also providing training and support. The tool is intended to assist state employees with drafting documents and analyzing information. The partnership follows an executive order issued by Governor Gavin Newsom in March, which aimed to accelerate the use of AI to improve government efficiency while maintaining safety standards. “AI should not replace the human work of government; it should help our workers move faster, solve problems more effectively, and deliver better results for Californians,” Newsom said in a statement.

The state-level agreement highlights a divergence from federal policy in Washington. Earlier this year, Anthropic and the U.S. Department of Defense clashed over a contract that would have permitted the agency to deploy Claude for any lawful use. During negotiations, Anthropic sought to carve out protections against government surveillance and autonomous weapons without human oversight. However, Defense Secretary Pete Hegseth refused the terms, and the department signed a contract with competitor OpenAI instead. The federal government subsequently declared Anthropic a “supply-chain risk”—a designation used by the U.S. government to restrict contracting with specific entities—which prevented the company from working with other contractors associated with the department. Commenting on the federal approach, Newsom noted at the time that while policymakers in Washington were designing contracts in the shadow of technology misuse, California was focused on establishing its partnerships correctly.

While the federal government’s designation restricts Anthropic’s ability to work with other contractors, California officials took a different path. Chris Given, California’s Chief Information Officer and Department of Technology director, stated that the “supply-chain risk” designation did not come up during the negotiations for the state’s contract. The state’s decision to move forward independently highlights how local administrations are establishing their own frameworks for AI adoption, bypassing federal security classifications.

Why it matters

The deal highlights a growing divergence between state and federal AI procurement strategies, as California embraces Anthropic while the U.S. Department of Defense restricts the company over security concerns.