The Pentagon has transferred at least two-thirds of its artificial intelligence workload from Anthropic to rival providers following a significant policy dispute. This vendor shake-up marks a pivotal moment in the Defense Department’s engagement with commercial AI technology.
The conflict escalated in early 2026 when defense officials requested unrestricted lawful uses for Claude, including autonomous weapons deployment. Anthropic CEO Dario Amodei declined the request, citing safety principles. Consequently, the department designated Anthropic a supply-chain risk and cancelled portions of a $200 million prototyping contract awarded in July 2025.
By June 2026, the DoD formalized new agreements with OpenAI, Google, and Microsoft to support classified networks. OpenAI capitalized on the opening by expanding service access, leveraging updated usage policies that permit military applications. Google and Microsoft also secured positions through existing federal cloud infrastructure relationships.
This strategic pivot carries substantial weight for investors tracking the AI sector. OpenAI has added significant government revenue to its profile while exploring liquidity options. Furthermore, legal challenges regarding Anthropic's blacklisting could establish critical precedents defining whether acceptable use policies constitute liability in future federal contract negotiations.