A federal appeals court upholds DOD's Anthropic blacklisting, finding Claude's integration with DOD systems is "a statutorily covered national-security risk" (Ashley Capoot/CNBC)

The US Court of Appeals for the District of Columbia ruled on Friday that the Pentagon can continue to block Anthropic's Claude AI models from military and federal government information systems. The decision reaffirms a supply-chain risk designation previously challenged and overturned by a lower federal court. Anthropic maintains that Claude's restrictions on certain tasks prevent the company from enabling the military to deploy the models for autonomous weapons or surveillance.

The Pentagon argues the restrictions violate contractual terms the government deemed essential. Both Anthropic and the US government have argued their positions in appeals courts, with the final decision in a 2-1 majority. The ruling could allow the Pentagon to continue excluding Claude from its supply chain indefinitely, potentially impacting revenue for Anthropic as customers grapple with concerns about doing business with a government pariah.

The Pentagon has not disclosed how the designations have affected its budget. Meanwhile, Anthropic is targeting a potential initial public offering later this year. The decision follows a decision in April where the same appeals panel declined to block the designation temporarily.