Related ToolsClaudeChatgpt

Pentagon Signs Classified AI Contracts with 7 Companies, Quietly Drops Anthropic

Anthropic
Image: Anthropic

The Defense Department signed classified AI agreements with seven tech companies on Friday: OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk's xAI, and a startup called Reflection. One notable name is absent - Anthropic, which the Pentagon had previously authorized for classified work.

These aren't standard software subscriptions. Classified AI contracts mean the tools operate inside government networks that handle restricted national security data - systems that can't touch ordinary commercial cloud infrastructure without special clearance. Getting onto that list requires vetting that takes months, which is part of why Anthropic's removal stands out.

Who Made the Cut and Why

The seven vendors cover a wide range. Google, Microsoft, and Amazon bring both AI models and the underlying cloud infrastructure already embedded across federal agencies - AWS GovCloud, in particular, already holds some of the most sensitive federal workloads in existence. Nvidia's inclusion is notable because it makes chips, not AI models, so its role here almost certainly relates to compute hardware rather than any software it publishes publicly.

The surprise addition is xAI. Grok, Musk's flagship model, is less than two years old. Landing a classified DoD agreement this early is a meaningful government validation - the kind that typically takes companies much longer to earn.

What Anthropic's Exclusion Signals

Anthropichasn't commented publicly on the change. Whether the Pentagon chose not to renew access or whether Anthropic declined to bid isn't clear from the announcement. Claude has an active enterprise push underway, and Anthropic has an existing commercial relationship with AWS, so this doesn't fit a pattern of deliberately avoiding government contracts.

For ChatGPT maker OpenAI, the contract cements a defense relationship that accelerated after the company removed restrictions on military use of its API in early 2024. For Anthropic, it's a quiet but significant commercial setback.

The U.S. federal government is one of the largest AI buyers globally. Classified contracts function as a proving ground - vendors who build deep integrations at the classified tier tend to win larger unclassified deployments next. Missing this round means Anthropic's competitors have a head start on that relationship-building while Anthropic watches from the outside.