Trump orders U.S. Treasury to ban Anthropic and Claude

What happens when an AI lab tries to tell the military how its models can and cannot be used? You get kicked out of the entire federal government.

U.S. Treasury Secretary Scott Bessent announced that the Treasury Department is completely ending its use of Anthropic products, including the Claude platform, following a direct order from President Donald Trump.

Bessent stated that the American public deserves confidence that every federal tool serves the public interest, adding that under Trump, no private company will ever dictate the terms of national security.

A battle over terms of service

The Treasury's ban is part of a sweeping directive across the entire federal government. Trump ordered all agencies to immediately halt using Anthropic's technology after clashes over military applications.

In a statement reported by CNBC TV18, Trump slammed Anthropic as a radical left and woke company. He claimed the company made a disastrous mistake by trying to strong-arm defense officials into adhering to its commercial terms of service instead of the U.S. Constitution.

Trump maintained that decisions on how the military fights wars belong strictly to the commander-in-chief. Agencies currently relying on Anthropic have been given a six-month phase-out period, alongside warnings of further penalties if Anthropic fails to cooperate.

Why it matters

If you are an enterprise developer building software for federal agencies or defense contractors, this is a massive operational headache. Even with a six-month transition window, workflows built around Claude will now need to be rapidly migrated to alternative vendors or self-hosted models.

For the broader AI ecosystem, this marks the most high-profile clash yet between Silicon Valley AI safety principles and state authority. Anthropic has long positioned itself as the standard-bearer for responsible, constitutionally aligned AI guardrails.

We think this clash highlights an inevitable reality check: when private acceptable-use policies meet sovereign defense mandates, the government will always push back hard. Moving forward, AI labs hoping to win lucrative public-sector contracts will likely face immense pressure to strip away their usage restrictions or risk total exclusion from federal procurement.