Pentagon widens classified AI partnerships with Nvidia, Microsoft, AWS and others, excludes Anthropic

The U.S. Department of Defense expanded or renewed AI agreements with multiple major vendors, including Nvidia, Microsoft, Amazon Web Services, SpaceX, OpenAI, Google and Reflection AI, while keeping Anthropic barred from Defense work following a prior supply‑chain risk designation ^{[1] [2] [3] [4]}.

Reports say the contracts enable deployment of AI systems on the military’s most secure classified networks at Impact Levels 6 and 7 and support the Pentagon’s GenAI.mil platform, which has attracted more than 1.3 million users; the expansion has been highlighted as strengthening Nvidia’s ties to the Pentagon and potentially boosting its market prospects, while officials cited security concerns in vendor selection ^{[1] [5] [4]}.

The moves reflect the Pentagon’s focus on rapidly fielding generative‑AI capabilities on classified networks while applying security assessments that continue to shape which commercial AI providers can participate in defense programs ^{[4] [1]}.