Pentagon Excludes Anthropic in Deal With OpenAI, Google, SpaceX and Others for Classified AI Work

The Pentagon has announced new agreements with major artificial intelligence companies to deploy advanced AI tools on classified military networks. However, it has left out Anthropic amid a fight over military use of its technology.

The Defense Department said Friday it had reached agreements with leading AI firms, including SpaceX, OpenAI, Google, Nvidia, Reflection AI, Microsoft, Amazon Web Services, and Oracle, to bring their systems into classified network environments for “lawful operational use.”

The new deals are part of a broader push by the Trump administration and Defense Secretary Pete Hegseth to accelerate the military’s adoption of artificial intelligence.

In its announcement, the Pentagon said the agreements would help establish the United States military as an “AI-first fighting force” and improve troops’ ability to maintain “decision superiority” across the battlefield. The companies will be integrated into Impact Level 6 and Impact Level 7 networks, highly sensitive Defense Department environments used for secret and top-secret work.

The Pentagon said the tools are expected to “streamline data synthesis, elevate situational understanding, and augment warfighter decision-making in complex operational environments.” But the announcement is also notable for not including Anthropic, the maker of Claude and one of OpenAI’s most prominent rivals.

Anthropic has been locked in a dispute with the Pentagon over the use of its AI. The company objected to that standard over concerns that its technology could be used for domestic mass surveillance or fully autonomous lethal weapons.

The Pentagon then labeled Anthropic a supply-chain risk earlier this year, barring its use by the department and contractors. Anthropic sued in response. Reuters said that Anthropic’s tools had been widely used across the Defense Department and that some Pentagon staffers, former officials, and IT contractors were reluctant to give them up because they viewed them as superior to alternatives.

The outlet also reported that newer AI firms said the Pentagon had accelerated the process of bringing vendors into top-secret data environments, cutting a process that once took 18 months or longer to under three months. The Pentagon said the deal helps avoid dependence on any single vendor. Its announcement said the department wants an architecture that prevents “AI vendor lock” and preserves long-term flexibility for the Joint Force.

The Pentagon said GenAI.mil, its official AI platform, has been used by more than 1.3 million department personnel in five months, generating tens of millions of prompts and deploying hundreds of thousands of agents. The inclusion of Reflection AI also drew attention as the lesser-known startup raised $2 billion in October and is backed by 1789 Capital, a venture firm in which Donald Trump Jr. is a partner and investor.