The Pentagon is making secret AI deals with OpenAI, Google, and Nvidia — but not with Anthropists.

The Pentagon has struck deals with OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk’s xAI startup Reflection, allowing the agency to use its AI tools in classified settings, according to a report. Announcement on Friday. Meanwhile, the Ministry of Defense has axed the Anthropic app – which it had previously used to obtain classified information – after declaring it a supply chain risk.

Emil Michael, chief technology officer at the Department of Defense He told CNBC on Friday Anthropic still poses a risk to the supply chain, but it’s called a risk Robust security model, Mythos“A separate national security moment,” he said, adding, “We have to make sure we strengthen our networks, because this model has special capabilities to find and patch cyber vulnerabilities.”

The Pentagon says in its announcement that the agreements with the seven artificial intelligence companies will allow “lawful operational use” of their systems, and “establish the United States military as the premier fighting force in the field of artificial intelligence.”