Negotiations between the Department of War and Anthropic, a leading artificial intelligence company, have stalled over demands from the AI firm for safety guardrails in the deployment of its technology, according to sources familiar with the discussions.
The dispute, which has been ongoing for months, centers on Anthropic’s insistence that the Department of War commit to specific protocols to mitigate potential risks associated with advanced AI systems. These concerns come as the Pentagon rapidly accelerates its adoption of AI across multiple domains, from intelligence gathering to warfighting operations. The Department of War released an “Artificial Intelligence Strategy” on January 9, 2026, outlining plans for rapid deployment and data access, and Secretary Hegseth subsequently announced an overhaul of the Department’s innovation and acquisition ecosystems on January 12.
The Pentagon’s push for AI integration follows a $200 million series of contract awards in July 2025 to Anthropic, Google, OpenAI, and xAI, aimed at providing advanced AI capabilities to address national security challenges. Dr. Doug Matty, the Chief Digital and AI Officer, stated at the time that the adoption of AI was “transforming the Department’s ability to support our warfighters and maintain strategic advantage.”
The current impasse with Anthropic highlights a growing tension between the Department’s desire for rapid innovation and the need to address potential safety concerns associated with increasingly powerful AI models. The Department is attempting to establish a single, centrally managed innovation system led by the Under Secretary of War for Research & Engineering, aiming for “wartime speed” in AI deployment. This approach, though, appears to clash with Anthropic’s more cautious stance.
Details of the specific guardrails Anthropic is seeking remain confidential, but sources suggest they relate to limitations on the use of AI in autonomous weapons systems and safeguards against unintended consequences. The company’s demands reflect a broader debate within the AI community about the ethical and security implications of advanced AI technologies.
The Department of War has not publicly commented on the negotiations with Anthropic, and representatives for the company declined to provide specifics. As of February 23, 2026, the two parties remain at odds, with no immediate resolution in sight. The Pentagon’s AI strategy, unveiled earlier this month, mandates data access and aims to remove barriers to AI adoption, potentially creating further friction with companies prioritizing safety considerations.
