Product4 publishers3 min readPublished
Defense sellers should plan Pentagon systems without Claude
Federal appeals judges voted 2-1 to uphold the Pentagon's blacklisting of Anthropic over the limits it puts on how Claude can be used. Teams with Claude inside military deliverables should budget for a replacement while Anthropic weighs further review.
The Product Desk · Product desk
What happened
- The majority found the designation reasonable after Anthropic refused to let its products be used for autonomous weapons or mass surveillance, and rejected Anthropic's retaliation claim.
- Last month in San Francisco, US District Judge Rita Lin struck down a parallel designation under a different law, finding the administration had unlawfully retaliated against Anthropic.
- Anthropic says the designation has cost it billions of dollars in lost business and damaged its reputation ahead of its planned IPO.
- Anthropic says it is weighing further review, either a rehearing before all of the DC Circuit's judges or a petition to the Supreme Court.
Compiled by The Product DeskSomething wrong?How this is made
Why it matters
- precedent A defense buyer can now cite a model vendor's usage restrictions as a supply-chain risk without alleging bad intent. Any AI vendor's terms on weapons and surveillance use become part of a defense bid review.
- contradiction Two courts have reached opposite conclusions on whether the administration retaliated against Anthropic, so what a contractor may do depends on which statute a program office relies on.
- decision AI vendors bidding for defense work now have to choose between keeping bans on weapons or surveillance use and staying eligible for Pentagon contracts.
Picture the engineer whose document-triage tool for a defense program calls Claude. On Monday she has two court rulings in front of her that point different ways, and a program manager who wants one answer.
Both rulings can stand at once because Hegseth designated Anthropic under two different laws [16]. The one the DC Circuit upheld [20] rests on the Supply Chain Security Act. "In doing so here, the Secretary did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution," the court wrote [6]. Judge Lin's ruling on the other law also blocked the government-wide ban on Anthropic and Hegseth's order barring military contractors from doing any business with the company [10].
The majority's reasoning is about how the product behaves. "The Secretary raises the deeply sobering prospect of overly constrained AI models shutting down unexpectedly and thus causing important military operations to fail," the majority wrote in an opinion by Judge Gregory Katsas [4]. The ruling set Anthropic's concern beside it: "Anthropic raises the deeply sobering prospect of unconstrained AI models hallucinating inappropriate targets for lethal military force" [5]. It then left the choice to the executive branch, saying Trump and Hegseth "must determine how best to balance the competing risks" [19].
The feature Anthropic pitches is the one the Pentagon objected to. Anthropic presents its refusals as a safety position, saying AI is not yet reliable enough to be used safely in autonomous weapons [17]. Hegseth argued the same restrictions could compromise military operations [18]. According to Ars Technica, the judges held that the government could blacklist Anthropic for withholding the features even if the company had no malicious intent [7].
Teams tell themselves the San Francisco ruling means this unwinds. The record on the supply-chain designation runs the other way. It dates to March [2]. The same appeals court refused Anthropic an emergency stay in April [8] and upheld the designation on Friday [1]. An Anthropic spokesperson told CNBC, "Another federal court has already held the government's parallel designation unlawful" [12]. The statement is accurate, and it refers to the other law [16]. Commerce Secretary Howard Lutnick recently said the administration and Anthropic are "in tune" [14].
The panel found "ample support" for the Pentagon's view that integrating Anthropic's technology into military systems could pose a national security threat [15]. The reports do not spell out how far the upheld designation reaches into a contractor's internal tools.
I'd plan for Claude to stay out of anything delivered into a Pentagon system for as long as the designation stands. The tradeoff is paying now to re-test prompts and evaluations on a second model, against a designation Anthropic is still contesting.
Sort each program on two axes: whether Claude sits inside what the military receives or only in internal tooling, and whether swapping the model takes a sprint or a quarter of re-testing. Inside the deliverable and slow to swap is where the upheld finding lands [15], and the migration starts now. Inside and quick to swap, a tested second model named in the proposal covers the risk. Internal and slow to swap, the protection is Judge Lin's block on the contractor-wide ban [10]. Internal and quick to swap is the one quadrant where waiting for further review costs little.
What to watch
- Whether the supply-chain designation stays in force while Anthropic seeks rehearing by the full DC Circuit or petitions the Supreme Court.
- Any appeal of Judge Rita Lin's ruling, since her order is what keeps the ban on military contractors doing business with Anthropic blocked.
- Whether Commerce Secretary Howard Lutnick's 'in tune' remark turns into a negotiated deal that lifts the designation.