A federal appeals court has upheld the Pentagon's decision to bar Anthropic from military contracts due to supply chain security risks. Defense Secretary Hegseth argued that Anthropic's safety restrictions could jeopardize military operations. Anthropic says the designation has already cost it billions.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
A U.S. federal appeals court upheld 2-1 the Pentagon's designation of Anthropic as a national security supply chain risk, after the AI startup refused to allow its technology to be used for autonomous weapons and mass surveillance, costing it military contracts. Behind the ruling, the tension between security positions, political motives, and intelligence agencies' actual use remains unresolved.
A Washington federal appeals court, by a 2-1 vote, upheld the Pentagon's designation of AI startup Anthropic as a national security supply chain risk. According to CNBC, the court found that the Pentagon has the authority to bar Anthropic from military contracts based on national security considerations. This ruling provides judicial backing for the Defense Department to exclude a specific AI vendor from the procurement system.
The core dispute in the case is not technical capability, but Anthropic's refusal to allow its technology to be used for autonomous weapons and mass surveillance. Defense Secretary Pete Hegseth had argued that Anthropic's safety restrictions could jeopardize military operations. The court's majority opinion clearly accepted this logic: when a vendor's usage policy conflicts with the military's operational needs, the Defense Department can treat it as a supply chain risk.
Anthropic opposes the ruling and says it is weighing its next steps. The company had previously made progress on another legal front: in late August, a San Francisco federal judge blocked a parallel designation made under a different law, calling it illegal retaliation against Anthropic's AI safety stance. The two litigation tracks are based on different laws and heard by different courts, and the divergent outcomes mean this dispute is far from over.
According to Anthropic, the supply chain risk designation has already caused the company billions of dollars in losses and is damaging its planned IPO. This figure comes from the company itself and has not been independently verified by a third party, so it should be treated as Anthropic's unilateral claim.
A striking contradiction is this: U.S. intelligence agencies are heavy users of Anthropic's AI models. If Anthropic is designated a national security supply chain risk, how the intelligence agencies' reliance on its models squares with that label remains an open question. This fact does not directly negate the court ruling, but it reveals a disconnect between security review and actual practice.
Some tech industry figures and former military officials have backed Anthropic in the dispute. Meanwhile, there are reports that the Trump administration appears to be punishing Anthropic for being seen as "left-leaning" and "woke," and President Trump himself has explicitly expressed similar attitudes. These claims come from reporting and outside observation, and are not the direct basis of the court ruling.
This article is based on THE DECODER's account of CNBC reporting. The court ruling, the vote margin, the origin of the case, Hegseth's argument, Anthropic's response, and its loss claims all come from this single source. Anthropic's "billions of dollars in losses" and "IPO impact" are its unilateral claims, not independently verified. The description of the Trump administration's motives is outside observation in the reporting, not a judicial finding.
The court ruling confirms the Pentagon's discretion over supply chain security, but the temporary block Anthropic obtained in another court, the intelligence agencies' actual reliance on its models, and the allegations of political motive together form a dispute that has not yet closed. The boundary between security positions and national security procurement still awaits subsequent litigation and actual policy to clarify.
Primary report
Primary source