California Governor Gavin Newsom has officially launched a pilot partnership with Anthropic, granting state government agencies access to the AI assistant Claude. The agreement, signed on June 29, provides a 50% discount for state agencies, cities, and counties, along with full employee training and technical support from developers.
A Tool for Enhancing Efficiency, Not Replacement
Newsom emphasized that integrating AI into government processes is not intended to replace human labor. Instead, Claude is meant to accelerate routine tasks, improve customer service, and strengthen cybersecurity. Specifically, several state agencies are already using the model to scan and fix code, as well as to optimize internal processes.
Claude will be the first AI tool available through the single IT services portal SITeS of the California Department of Technology. This means all agencies will have standardized access to advanced algorithms without needing to handle integration themselves.
Strategic Course Toward Administrative Automation
This agreement is a logical continuation of the course Newsom set back in 2023, when he ordered a study of generative AI's potential in the public sector. In 2025, a directive followed to improve government efficiency, and in March 2026, AI procurement rules were updated with a focus on civil rights, privacy, and employee risks.
However, the context of Claude's deployment in California cannot be viewed in isolation from global events. Notably, on June 9, Anthropic released two versions of models from a new family—Fable 5 and Mythos 5. The first was described as a safe Mythos-class solution, while the second was a "private" model with relaxed restrictions. Yet, just days later, the company restricted access to tools for foreign nationals following a directive from the U.S. government. Some sources link this to the use of the models by a group connected to China. At the end of the month, Anthropic accused operators affiliated with Alibaba of distilling Claude.
Expert opinion: California's partnership with Anthropic is a significant precedent for the market. On one hand, it demonstrates regulators' trust in the technology; on the other, it sets a precedent for other states. However, it is worth remembering that any large-scale AI deployment in the public sector requires strict oversight of safety and ethics, especially given recent incidents involving access restrictions and leaks through distillation. The market will closely watch how Anthropic balances openness and security in a government context.