Federal court backs Pentagon's decision to flag Anthropic as a supply chain security risk

A federal appeals court in Washington has upheld the Pentagon's decision to bar AI startup Anthropic from military contracts, ruling 2-1 that the agency was right to classify the company as a national security supply chain risk. The ruling, reported by CNBC, deepens a months-long legal fight over whether an AI company can limit how its models are used by the military — and the stakes extend well beyond one startup. The designation has reportedly cost Anthropic billions of dollars and is complicating its planned IPO.

Why the Pentagon Barred Anthropic

The dispute traces back to Anthropic's refusal to allow its technology to be used for autonomous weapons and mass surveillance — two use cases the company has drawn hard lines around. Defense Secretary Pete Hegseth argued that these safety restrictions could jeopardize military operations, framing them as an unacceptable risk to missions that depend on reliable access to AI tools.

The Pentagon's supply chain risk classification effectively locks Anthropic out of defense contracting, a market where AI companies increasingly compete for work in logistics, analysis, and intelligence support. Anthropic rejects the ruling and says it is weighing its next steps, leaving the door open to further legal action.

A Split Picture in the Courts

The appeals court decision sits in tension with a separate ruling from late August. A federal judge in San Francisco blocked a parallel classification under a different law, calling it unlawful retaliation against Anthropic's AI safety stance. That ruling suggested the government's moves were punitive rather than security-driven; the Washington court's 2-1 decision now suggests the opposite.

For readers following the case, the divergence matters: Anthropic faces one set of legal standards in San Francisco and another in Washington, and the company's practical access to federal work may hinge on which interpretation ultimately prevails.

Intelligence Agencies Still Use Anthropic's Models

A notable wrinkle: US intelligence agencies are heavy users of Anthropic's AI models even as the Pentagon brands the company a supply chain risk. How those two positions square with each other remains an open question.

Parts of the tech industry and former military officials have publicly backed Anthropic. Meanwhile, the Trump administration's motivations have been openly political — President Trump has explicitly suggested the company is being punished for being "left-leaning" and "woke."

Where the Fight Stands

The case has moved from a question of procurement policy to a broader test of whether AI safety commitments can trigger national security penalties. Anthropic's options include seeking further review of the appeals court ruling or continuing to challenge related designations that the San Francisco judge already found unlawful. Watch whether Anthropic files an en banc petition, and whether the IPO timeline shifts as the legal uncertainty persists.