The White House is drafting an executive order that would formally ban Anthropic's AI technology across the entire federal government. If signed, it would mark the first time a U.S. administration has moved to blacklist a domestic AI company from government operations.
The conflict traces back to a $200 million Pentagon contract awarded to Anthropic in July 2025. When the Department of Defense sought to use Claude for mass surveillance of American citizens and for guidance in autonomous weapons systems, Anthropic declined. The company sought narrow assurances that its technology wouldn't be used for those specific purposes.
The Pentagon's response was severe. Defense Secretary Pete Hegseth designated Anthropic a "supply chain risk" - a classification normally reserved for foreign adversaries like Huawei - and issued a sweeping directive: no contractor, supplier, or partner that does business with the U.S. military may conduct any commercial activity with Anthropic.
Agencies Are Already Dropping Claude
The executive order hasn't been signed yet, but agencies aren't waiting. The Treasury Department has confirmed it's phasing out Anthropic technology. The State Department is removing Anthropic from its workflows. Health and Human Services is switching staff to ChatGPT Enterprise and Google Gemini. The GSA has pulled Anthropic from its Multiple Award Schedule and the USAi procurement program.
The six-month phase-out window gives the Pentagon time to untangle Claude from military platforms where it's already embedded. Civilian agencies appear to be moving faster.
Anthropic Sues, Calls the Ban "Unprecedented and Unlawful"
Anthropic filed a lawsuit challenging the government's actions, arguing that Congress never gave the executive branch authority to blacklist a U.S. company over what amounts to protected speech - namely, the company's public position on AI ethics and use restrictions.
That legal argument is interesting. Anthropic isn't claiming the government must use its products. It's arguing that punishing a company for setting usage boundaries on its own technology crosses a constitutional line. The outcome could set a significant precedent for every AI company that maintains an acceptable use policy.
The Chilling Effect
The practical fallout for daily AI tool users is limited right now - this only affects federal contracts, not consumer or commercial access to Claude. But the broader signal is hard to ignore.
Every major AI company maintains usage policies that restrict certain applications. OpenAI, Google, and Meta all have prohibited use cases in their terms of service. If the government can effectively blacklist a company for enforcing those policies, other AI labs will face pressure to quietly drop their own restrictions when government contracts are on the table.
Hegseth's directive that no military contractor can work with Anthropic adds another dimension. Defense contractors like Palantir, Booz Allen, and Lockheed Martin now have to choose between Anthropic partnerships and Pentagon business. That's not a real choice - they'll drop Anthropic every time.
Anthropic's lawsuit will take months to resolve. In the meantime, Claude is being ripped out of federal systems, and every AI company with a usage policy is watching closely.