Pentagon AI Deals

The Pentagon just drew a line in the AI industry — and where you stand on autonomous weapons determines which side you're on.

Today, the Department of Defense formalized agreements with seven AI companies — OpenAI, Google, NVIDIA, Microsoft, Amazon Web Services, SpaceX, and Reflection — to deploy their technology on its most sensitive classified networks, Impact Levels 6 and 7.

Anthropic was not on the list.

The reason matters. Negotiations broke down after Anthropic sought written assurances that Claude would not be used for fully autonomous weapons systems or domestic mass surveillance. The Pentagon declined. Anthropic walked.

That is not a failure of diplomacy. It is a values statement with a financial cost attached to it.

Most companies talk about responsible AI in press releases. Anthropic just demonstrated what it actually looks like when the stakes are real — and the stakes here include billions in government contracts and a direct line to U.S. military infrastructure.

President Trump said last week that Anthropic is "shaping up," which suggests the administration sees this as a negotiation, not a final answer. Whether Anthropic holds its position or eventually finds a middle ground will tell us a lot about whether safety guardrails are principles or leverage.

The broader question for every AI leader watching this: when your biggest prospective customer asks you to remove the guardrails, what do you do?

What would you do if a major contract required you to compromise on your product's core safety design?