Summary:
On Feb 27, 2026, the U.S. government moved to phase out Anthropic across federal agencies after a public clash with the Department of Defense. The DoD said it would designate Anthropic a “supply chain risk,” escalating what had been months of behind-the-scenes negotiations into a high-visibility political and procurement fight.

Anthropic’s tweet response (what they posted):
Anthropic posted on X linking to its statement responding to Secretary of War Pete Hegseth’s comments. (Anthropic on X)

What Anthropic says the dispute is about (per their statement):
Anthropic says negotiations hit an impasse over two narrow exceptions it requested to “lawful use” of Claude: mass domestic surveillance of Americans and fully autonomous weapons. Anthropic also says it plans to challenge any “supply chain risk” designation in court and argues the Secretary lacks authority to extend such a restriction beyond DoD contract usage. (Anthropic statement)

Why it matters:
This is a real-world example of policy risk turning into platform risk: access to a major model provider can be restricted for reasons unrelated to model quality or price. For operators, it raises the urgency of continuity planning (multi-provider options, routing, audit trails) because procurement and legal teams will now treat “model vendor risk” like any other critical supplier risk.

Monetize: sell “Protected Workflows” (per department / per automation), not seats.