California Signed a Statewide Claude Contract at Half Price

California just signed Anthropic as its first statewide AI vendor at half price — while Washington still blacklists the same company and keeps its frontier models on a federal guest list, proving America's AI policy is now a jurisdiction fight, not a national strategy.

What Newsom signed:

• On June 29, Governor Gavin Newsom announced a first-of-its-kind partnership giving all state agencies, cities, and counties access to Claude at a 50% discount, plus free workforce training and technical support from Anthropic, per the Governor's office.
• Claude becomes the first AI productivity tool available statewide through California's new Statewide Information Technology Shared Services (SITeS) portal — a centralized procurement hub with transparent pricing.
• Anthropic's models already power Poppy, California's in-house AI assistant, and agencies including the DMV and Department of Health Care Services have been piloting Claude for customer service and internal workflows.

What Washington did instead:

• In March, Defense Secretary Pete Hegseth designated Anthropic a supply-chain risk to national security — a label previously reserved for foreign adversaries — after the company refused Pentagon terms allowing Claude for mass domestic surveillance or fully autonomous weapons without human oversight, per CNBC.
• The Pentagon walked away and signed with OpenAI instead.
• In June, Commerce Secretary Howard Lutnick used export controls to keep Anthropic's Fable 5 dark for the public while restoring Mythos 5 only to roughly 100 vetted U.S. partners — the same week OpenAI's GPT-5.6 launched to a government-approved guest list of about 20.

Why the split is the story:

• California's CIO Chris Given told POLITICO that the supply-chain risk designation "just didn't come up" during contract negotiations, per TechCrunch.
• A federal judge had already blocked the Pentagon designation in March, with U.S. District Judge Rita Lin ruling the action appeared designed to punish Anthropic for public criticism rather than protect national security, per Reuters.
• Sacramento is not alone: Austria asked the EU to host Anthropic inside the bloc, Brussels flew to D.C. to petition for Mythos access, and India wrote bank kill-switch rules while Washington punted — every level of government is now jurisdiction-shopping for frontier AI.

What they said:

• "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 the announcement.
• "This partnership is about using technology the California way: responsibly, transparently, and in service of people," he added.
• The deal builds on Newsom's March executive order requiring AI vendors to demonstrate responsible policies on bias, civil rights, and misuse prevention before winning state contracts — the first major agreement under that framework.

The commercial stakes:

• For Anthropic, a statewide reference from the state where it is headquartered is a public-sector anchor at the exact moment enterprise buyers are watching Washington's kill switches.
• For California, half-price Claude is a procurement lever — but agencies remain responsible for accuracy, transparency, and privacy on a case-by-case basis under state AI policy.
• For every other governor, the message is clear: you can buy from a vendor Washington is fighting if your procurement office decides the federal label is not your problem.

Convina's view:

Newsom didn't pick a chatbot. He picked a side in the AI cold war inside America. Sacramento signed Anthropic at a discount, trained its workforce, and routed procurement through a state portal while Given shrugged off a Pentagon label a federal judge already called punitive. That is not efficiency policy — it is digital secession by contract. Washington spent June building guest lists for Mythos and GPT-5.6; California spent June putting Claude on every agency desktop. The labs cannot serve both masters forever. Either federal AI policy becomes a floor every state accepts, or America's largest economy just proved the floor is optional — and every enterprise buyer with a compliance department now has to decide which government actually governs their model vendor.