OpenAI Inks Classified Deal with Pentagon
OpenAI has signed a classified agreement to deploy its advanced AI models on secure military networks for the Pentagon's Office of Warfighting (DOW). The deal is described as "guardrail-focused," signaling a deeper integration of generative AI into sensitive national security and intelligence operations.
This agreement is part of a much larger push, as the Pentagon's fiscal year 2026 budget request for AI and autonomy has reached a record $13.4 billion, the first time these technologies have been given their own dedicated budget line. Of that total, $9.4 billion is designated for aerial drones and $1.7 billion for maritime autonomous platforms. The deal is situated within a competitive landscape among top AI labs for lucrative government contracts. Just before the OpenAI agreement, rival company Anthropic was designated a "supply chain risk" by the administration after it refused to remove safeguards against its AI being used for mass surveillance or autonomous weapons. In response to public scrutiny, OpenAI published what it calls its "red lines," stating its technology will not be used for mass domestic surveillance, to direct autonomous weapons systems, or for high-stakes automated decisions. The company claims its agreement has more protective guardrails than any previous classified AI deployment. This partnership reflects a significant shift from just a few years ago when leading AI companies, including OpenAI, publicly resisted military applications of their technology. In January 2024, OpenAI quietly removed language from its usage policy that had prohibited "military and warfare" applications. The contract, valued at $200 million and running through July 2026, is with the Pentagon's Chief Digital and Artificial Intelligence Office (CDAO). The work will focus on developing "prototype frontier AI capabilities" for both warfighting and enterprise functions like improving military healthcare and cyber defense. This initiative is part of the Pentagon's broader AI Acceleration Strategy, which aims to make the department an "AI-first' warfighting force." The strategy outlines seven "Pace-Setting Projects" to fast-track AI integration in areas like AI-enabled battle management and autonomous swarms, with initial demonstrations due by July 2026. The deployment will be cloud-only, with OpenAI retaining control over its safety features and having cleared personnel involved to ensure compliance. The agreement also led to the creation of an "OpenAI for Government" initiative to coordinate its work with various government agencies.
