In a notable change for military technology, the Department of Defense has started evaluating alternative AI models to replace Anthropic’s Claude, which was previously contracted for military use. This testing began on March 1, 2026, shortly after Defense Secretary Pete Hegseth identified Anthropic as a supply-chain risk due to ongoing safety concerns regarding the company's restrictions.
Anthropic's choice to keep limitations on Claude's functionalities—especially concerning mass surveillance and lethal autonomous weapon systems—has created friction with Pentagon officials. These officials are advocating for unrestricted use of AI technologies in national security. Tensions flared earlier this year when it was revealed that Claude had been used for intelligence analysis related to Iran, heightening concerns within the military.
In July 2025, Anthropic secured a lucrative contract potentially worth up to $200 million to integrate Claude into classified military networks. However, as the Pentagon sought broader applications for the AI, a rift grew between Anthropic's leadership and military decision-makers. By late February 2026, the relationship had deteriorated enough for Hegseth to take decisive action, prompting the current search for a replacement AI model.
Twenty-five military personnel are now using GenAI.mil, an independent platform set up for this evaluation. The competing models under review include those from OpenAI, Google, and Grok—Elon Musk’s AI venture. This new testing regime aims to assess how these alternative AI systems perform in workflows that were previously dominated by Claude.
The implications of this shift could reverberate throughout the defense sector and the broader AI field. The Pentagon's emphasis on AI capabilities that can be deployed without strict limitations underscores an urgency to adapt to emerging threats and technological advancements. As military needs evolve, efficiently and ethically deploying AI will remain a complex challenge.
Looking ahead, the results of these evaluations may not only redefine military workflows but also impact the commercial AI market as tech companies compete for government contracts. The interplay of AI ethics, military necessity, and corporate strategy will likely shape the conversation in the coming months, especially as more details about the competing models' capabilities come to light.
