A federal appeals court has sided with the US Department of Defense in its high-stakes dispute with Anthropic, allowing the Pentagon to keep the artificial-intelligence company off its military supply chain even as the startup prepares for what could become one of the largest initial public offerings in history.
On September 25, 2026, a 2-1 panel of the US Court of Appeals for the D.C. Circuit upheld the Defense Department’s designation of Anthropic as a national-security supply-chain risk.
The majority found that Pentagon officials reasonably concluded the company’s refusal to lift contractual and technical limits on two uses—fully autonomous lethal weapons and mass domestic surveillance—created unacceptable operational uncertainty for the armed forces.
Judges rejected Anthropic’s argument that the designation was unlawful retaliation for the firm’s public stance on AI safety and ethics.
The clash traces back to early 2026.
Anthropic had earlier secured a prototype agreement with a $200 million ceiling to place its Claude models on classified networks.
When Defense Secretary Pete Hegseth demanded an “all lawful uses” clause, Anthropic declined.
Officials then invoked the Federal Acquisition Supply Chain Security Act, a statute more commonly applied to foreign vendors, and directed the removal of Anthropic technology from covered Defense systems.
President Trump separately ordered a broader freeze on federal use of the company’s products.Anthropic sued in two venues.
A district judge in San Francisco later blocked the government-wide and contractor-wide bans, calling them punitive and unsupported.
The D.C. Circuit ruling, however, leaves the Pentagon-specific exclusion intact.
Defense contractors therefore cannot rely on Anthropic tools for work that touches covered military systems.The timing is delicate.
Investors have discussed a public listing that could value Anthropic near $2 trillion and raise as much as $100 billion, figures that remain preliminary and unconfirmed.
Company executives had previously warned in court filings that the supply chain label threatened billions in revenue, delayed enterprise deals, and risked undermining investor confidence just as the firm prepared to go public.
Commercial demand for Claude has nevertheless continued to climb, with enterprise customers accounting for the large majority of revenue.
The split judicial outcome leaves Anthropic in a hybrid position: still able to sell to most civilian agencies and commercial clients, yet locked out of a strategically important slice of the national-security market.
Pentagon officials have argued that no private vendor should dictate how the military deploys technology on the battlefield.
Anthropic has countered that certain uses cross bright ethical lines it will not cross. Whether the D.C. Circuit decision becomes the last word or prompts further appeals remains to be seen. For now it stands as a clear win for the Defense Department and a reminder that even the most valuable AI companies remain subject to government determinations about national security risk.