U.S. judge blocks Trump administration's ban on Anthropic — a landmark clash over AI, speech and supply‑chain security

Ruling halts blacklist while case proceeds

It has been reported that a U.S. federal judge in San Francisco granted a preliminary injunction on Thursday that temporarily blocks the Trump administration from enforcing its directive banning federal agencies from using Anthropic’s Claude model. Judge Rita Lin (林萍) said the government offered no “legal basis” for treating Anthropic differently and warned that punishing the company for publicly contesting government contracting positions looks like unconstitutional retaliation. The injunction follows Anthropic’s lawsuit seeking to overturn the Pentagon’s designation of the company as a supply‑chain risk.

A first for a U.S. AI firm, with big implications

Anthropic — a high‑profile AI startup — has reportedly become the first American company publicly designated as a supply‑chain risk, a label historically applied to foreign adversaries. The finding has ripple effects: it reportedly requires defense contractors such as Amazon, Microsoft and Palantir to certify they are not using Claude in military work. During recent hearings Judge Lin even cited an amicus brief that used the phrase “attempt to implement corporate murder,” asking whether the government was effectively trying to weaken the company.

This is more than a narrow contractual dispute. The case pits national‑security officials’ desire to police the AI supply chain against companies’ speech and commercial rights, and could shape how Washington balances security, innovation and trade policy in an era of AI competition with geopolitical rivals. Final judgment may still take months, and the injunction merely pauses—but does not resolve—the underlying policy questions.

What happens next will matter to U.S. tech firms and to allies watching how Washington treats domestic innovators under the banner of supply‑chain risk. Will the courts erect clearer limits on executive action? Or will national‑security prerogatives expand to police AI providers more aggressively?