Anthropic supply-chain designation upheld in AI military dispute

Investing.com -- A federal appeals court on Friday upheld the Department of War’s decision to remove artificial intelligence developer Anthropic (NASDAQ: ANTP) from its supply chain, delivering a major legal victory to the Trump administration in a landmark battle over the military’s use of frontier AI models.

At the heart of the dispute is Anthropic’s persistent refusal to grant the Department an unrestricted "all lawful uses" contractual term for Claude. While the AI vendor relaxed its public usage policies to accommodate classified threat assessments and cyber operations, it insisted on maintaining strict prohibitions against deploying Claude for mass domestic surveillance and lethal autonomous warfare, a stance CEO Dario Amodei defended as core to the company's corporate identity.

The tension escalated dramatically in early 2026 after Secretary of War Pete Hegseth mandated that the military achieve "AI Dominance" free from vendor-imposed constraints. Tensions boiled over following an incident during a sensitive January operation in Venezuela, where Anthropic questioned whether its software’s deployment violated terms of service, raising alarm among defense officials about the reliability of mission-critical systems during active combat operations.

Anthropic argued that its usage restrictions did not constitute a "supply chain risk" under the 2018 statute, asserting that the law was designed to target covert, malicious acts by foreign adversaries rather than open contractual disagreements with domestic partners. The court rejected that interpretation, holding that the statutory definition broadly encompasses any manipulation of design or operation that disrupts or denies system functionality, regardless of the vendor's underlying motives.

The court also dismissed Anthropic’s constitutional claims, ruling that post-deprivation process satisfied Fifth Amendment standards given the rapid evolution of AI technology and active military engagements. Additionally, the panel found no evidence of First Amendment retaliation, determining that the administration moved to exclude Anthropic due to its contractual intransigence rather than its public advocacy for AI safety.

Circuit Judge Karen LeCraft Henderson dissented, warning that the majority’s broad reading transforms a national security statute intended to prevent foreign espionage into a coercive tool to strip vendors of contractual rights. Henderson argued that the statute requires evidence of intentionally subversive or deceptive conduct, which the government never established against Anthropic.

Despite the D.C. Circuit’s ruling, the practical impact on Anthropic's defense footprint remains temporarily muted due to parallel litigation in California. A federal judge in San Francisco previously issued an order lifting the government's ban in a separate lawsuit brought under a different statutory authority, a ruling that currently remains in effect.

Against that complex legal backdrop, the broader commercial trajectory for the AI pioneer appears largely unfazed. Even as the Pentagon actively shifted contract volume toward rival provider OpenAI, robust private sector demand has continued driving Anthropic's enterprise valuation upward, underscoring a growing divergence between the company's surging commercial momentum and its fraught relationship with military procurement officials.

Original Article

You May Also Be Interested In

• Cognition AI hits $1 billion run rate amid demand for coding software - Bloomberg
• Bernstein cuts Deutsche Telekom on T-Mobile U.S. risks; names BT top pick
• Canadian boards have an AI problem — what is it?