OpenAI has finalized a multi-year agreement with the U.S. Department of Defense to deploy its advanced AI models across military operations, just hours after Anthropic was barred from federal contracts under new executive directives. The deal marks a pivotal moment in the national security use of artificial intelligence.
- OpenAI secured a $120 million, three-year contract with the U.S. Department of Defense
- The agreement enables deployment of GPT-5 and specialized variants for intelligence and logistics operations
- Exec Order 2026-04 suspended Anthropic from federal contracts due to national security concerns
- OpenAI will operate a secure enclave within DoD cloud infrastructure with human-in-the-loop oversight
- Contract includes quarterly audits by DISA and strict access protocols for sensitive data
- Lockheed Martin and Raytheon are pursuing integration of OpenAI models into defense systems
OpenAI has officially entered into a strategic agreement with the U.S. Department of Defense to integrate its latest generative AI models into defense systems, according to a public statement by CEO Sam Altman. The contract, valued at $120 million over three years, enables secure deployment of OpenAI’s flagship models—specifically GPT-5 and its specialized variants—for tasks including intelligence analysis, logistics optimization, and real-time battlefield simulation. The timing of the deal is significant, occurring just hours after the White House issued Executive Order 2026-04, which suspended federal contracts with Anthropic pending a security review. The order cited concerns over data sovereignty, model transparency, and potential foreign influence risks in AI systems used by national security agencies. While no specific data breach was reported, the directive reflects a broader effort to align AI procurement with U.S. strategic interests. Under the new agreement, OpenAI will establish a dedicated secure enclave within the Department of Defense’s cloud infrastructure, allowing for on-premise processing of sensitive military data. The system will be subject to quarterly audits by the Defense Information Systems Agency (DISA) and include strict access controls, with all model outputs undergoing human-in-the-loop verification before operational use. The move has drawn attention from defense contractors, with Lockheed Martin and Raytheon already initiating talks to integrate OpenAI’s models into their autonomous systems. Meanwhile, European defense firms, including Airbus and MBDA, have expressed interest in similar partnerships, though access to U.S.-developed AI remains restricted under current export control regulations.