Anthropic Ban is Pentagon-Only, Commercial Access Unchanged

The Anthropic situation got some useful clarification today. Google and Microsoft both confirmed that Anthropic models remain available through their cloud platforms for civilian and commercial use. The ban is specific to Pentagon and defense contracts, where Anthropic was designated a “supply chain risk.”

So the picture is: the DoD is blocked from using Claude, but every other customer on Google Cloud and Microsoft Azure can still access it normally. The initial reporting made it sound like a broader prohibition, but it’s narrower than that.

The “supply chain risk” designation came after Anthropic refused demands around military AI applications, specifically mass surveillance and fully autonomous weapons. There’s a separate open letter from employees at OpenAI, Salesforce, Databricks, IBM, and Cursor calling on Congress to examine whether that designation was appropriate.

Nvidia GTC is now 10 days out. DeepSeek V4 continues to not exist in public.