Anthropic Wins in Court as Pentagon’s AI Blacklist Collapses

Anthropic Wins in Court as Pentagon’s AI Blacklist Collapses
The Pentagon’s attempt to punish Anthropic over its boundaries on military AI use has run into a constitutional wall, turning a dispute over national security into a test of executive power.

A federal judge in California ruled that the Pentagon’s designation of Anthropic as a supply-chain risk was unlawful, blocking enforcement of the blacklist. The decision, as summarized by CNBC, treats the government’s action not as a routine procurement judgment but as retaliation against a technology company for taking positions the administration disliked.

The conservative framing is more direct and more focused on the conduct of Defense Secretary Pete Hegseth: the ruling “strikes down Hegseth’s ban on Anthropic AI” and calls it unconstitutional “retaliation.” That account emphasizes the administration’s failed attempt to exclude Anthropic from defense work. The liberal account, by contrast, places the decision within the broader question of how Washington should govern advanced AI while preserving access to private-sector technology.

The two perspectives nevertheless converge on the central fact: the Pentagon’s blacklist cannot stand in its current form. They diverge over its meaning. For critics, it is evidence that the administration crossed the line from regulating a contractor to punishing dissent. For the national-security establishment, the underlying dispute reflects a genuine concern about whether an AI supplier will meet military requirements.

Anthropic welcomed the ruling while leaving the door open to cooperation, saying, “We remain focused on working productively with the government to harness AI for our national security so all Americans benefit from this technology.” That formulation captures the unresolved tension: the company is challenging the Pentagon in court, but neither side can easily afford a complete break as the military race to deploy AI accelerates.