Pentagon’s AI Strategy Shifts as Anthropic Faces Blacklist

A dispute with Anthropic over the permissible uses of its artificial intelligence technology is forcing the Pentagon to diversify its AI partnerships, accelerating contracts with companies like xAI and OpenAI. The situation highlights the growing tension between the Defense Department’s desire for advanced AI capabilities and the ethical concerns of AI developers regarding surveillance and autonomous weapons systems.

Anthropic’s Standoff with the Pentagon

The Defense Department issued an ultimatum to Anthropic, demanding the AI company agree to its terms for military use or face being forced into compliance. This followed a meeting on February 25, 2026, between Defense Secretary Pete Hegseth and Anthropic CEO Dario Amodei, where Amodei outlined “red lines” prohibiting the use of Anthropic’s technology in autonomous weapons systems – where AI makes final targeting decisions – and mass domestic surveillance of American citizens [1].

The Pentagon maintains that its use of AI technology adheres to legal standards, stating that “legality is the Pentagon’s responsibility as the end user” [1]. Still, the impasse led to the Trump administration reportedly deeming Anthropic a “supply chain risk,” potentially forcing military contractors to discontinue using its models [1]. On February 28, 2026, President Donald Trump posted on Truth Social, directing all agencies to cease using Anthropic products within a six-month period [1].

Shifting Partnerships and the Role of Palantir

With Anthropic potentially sidelined, the Pentagon is actively pursuing partnerships with other AI developers. XAI has already reached an agreement to operate on classified systems, and OpenAI is reportedly working on a similar deal [1]. However, surpassing Anthropic’s Claude large language model in military favor will be a key challenge for these competitors.

A critical factor in the Pentagon’s AI strategy is its relationship with Palantir Technologies. Palantir has established itself as a central hub for data management within the Department of Defense, consolidating data feeds and providing access through its platform. One former employee of the Pentagon’s Chief Digital and Artificial Intelligence Office stated that Palantir has “taken over the data lake problem” inside the Pentagon, and that “Everything runs through Palantir. They’re the 1,000 pound gorilla in this space” [3].

Anthropic’s existing integration with Palantir provides a strategic advantage. According to a former DoD employee, this connection makes Anthropic more appealing than requiring Palantir to integrate with other AI providers [1].

Contractual Background

In July 2025, Anthropic, along with OpenAI, Google, and xAI, each received a $200 million contract from the Pentagon’s Chief Digital & AI Office (CDAO) [4]. This initial $600 million investment signaled the Defense Department’s commitment to exploring “agentic AI” capabilities. The Pentagon later awarded a further $200 million pilot program to Anthropic [1].

If the Pentagon proceeds with blacklisting Anthropic’s Claude AI platform, it could take three months or longer to regain access to comparable capabilities on classified networks [3].