Pentagon expands classified AI partnerships with eight firms, excludes Anthropic amid dispute

The Department of Defense has expanded classified artificial intelligence work across eight companies while excluding Anthropic, according to a new analysis from the Center for Security and Emerging Technology (CSET). The change comes amid an ongoing dispute between Anthropic and U.S. defense officials.

CSET, a Georgetown University research center in Washington, D.C., reported the shift and provided expert commentary that was also shared with DefenseScoop. The reporting frames the move as part of the Pentagon’s broader effort to fold advanced AI capabilities into classified military projects and operations.

Sources close to the contracting process told CSET

Sources close to the contracting process told CSET that the Department has opted to distribute sensitive AI tasks among multiple private-sector partners rather than continue or expand work with Anthropic while disagreements persist. CSET’s write-up highlights how procurement decisions are being shaped by security reviews, policy questions and industry debates over AI safety.

The analysis from a D.C.-based think tank reflects the policy and institutional debate in the capital over how the U.S. military adopts powerful AI tools. CSET’s coverage connects the procurement shift to ongoing oversight and coordination efforts inside defense and national security circles in Washington.