Pentagon Enters Secret AI Partnerships with OpenAI, Google, and Nvidia, Excluding Anthropic

The Pentagon has secured partnerships with OpenAI, Google, Microsoft, Amazon, Nvidia, Elon Musk’s xAI, and the startup Reflection, enabling the agency to utilize their AI technologies in classified environments, as announced recently. Notably absent from this list is Anthropic, which had previously collaborated with the Pentagon on classified projects but was deemed a supply-chain risk.

These new agreements build on existing collaborations with OpenAI and xAI, which have already established terms for the responsible use of their AI systems. Reports indicate that Google has also entered into a similar arrangement. According to sources, Microsoft and Amazon have long-standing connections with the Pentagon, while the contracts with Nvidia and Reflection are fresh additions.

Anthropic, which had a $200 million contract to manage classified information for the Pentagon, found itself at odds with the agency over its refusal to relax restrictions concerning mass domestic surveillance and fully autonomous weapons. This disagreement resulted in the exclusion of Anthropic’s products from federal use, prompting the company to file a lawsuit against the government, which led to a temporary injunction in its favor.

Emil Michael, the Defense Department’s chief technology officer, stated that Anthropic remains a supply chain risk but acknowledged the significance of its robust security model, Mythos, describing it as a “separate national security moment.” He emphasized the need to fortify networks, given Mythos’s unique capabilities in identifying and addressing cyber vulnerabilities.

In its announcement, the Pentagon highlighted that these agreements with the seven AI firms will facilitate the “lawful operational use” of their technologies, positioning the U.S. military as an AI-first combat force.