OpenAI Lands Deal with US Department of Defense to Develop AI

OpenAI formalizes an agreement with the US Department of Defense (DoD). The contract, with a maximum value of US$200 million, focuses on the development and prototyping of advanced AI capabilities to address national security challenges.

The award, which is managed by the Office of the Chief Digital and AI Officer (CDAO) at the DoD, has the goal "to prototype frontier AI capabilities to address critical national security challenges across combat and business domains." OpenAI says that the contract aims to "help the Department of Defense identify and prototype how frontier AI can transform its administrative operations."

This agreement falls within OpenAI's new initiative, "OpenAI for Government." This program was created to consolidate and expand the company’s collaborations with US federal, state, and local agencies. The initiative aims to provide public sector institutions with access to OpenAI’s most advanced AI models to optimize processes and enhance service delivery.

Before this agreement with the DoD, OpenAI had already collaborated with various government entities, including the US National Laboratories, the Air Force Research Laboratory, NASA, the National Institutes of Health (NIH), and the Department of the Treasury, reports The Verge. The formal launch of the "OpenAI for Government" program consolidates these efforts and establishes a structured path for future partnerships. The Pentagon contract is the first to be awarded under this framework, positioning the CDAO as a key stakeholder in the adoption of AI within the defense infrastructure.

The agreement, registered under code HQ0883-25-9-0012, is a fixed-amount prototype transaction with an estimated end date of July 2026. An initial US$2.0 million has been obligated from fiscal year 2025 research, development, test, and evaluation funds.

According to OpenAI, the AI capabilities will be used to identify prototypes that can transform administrative operations, improve healthcare access for service members and their families, optimize data analysis for acquisition programs, and support "proactive cyber defense." The DoD also mentions applications in "warfighting" domains, a point that contrasts with OpenAI’s usage policy, which explicitly prohibits the use of its services to "develop or use weapons" or "cause harm or destroy property." OpenAI says that all use cases must comply with its policies and guidelines.

As part of the "OpenAI for Government" initiative, the company offers government agencies access to its most advanced models through platforms such as ChatGPT Enterprise and ChatGPT Gov. These environments are designed to meet security and regulatory compliance requirements. In addition, limited customized models will be offered for national security applications, along with direct technical support and early access to future technological developments to support strategic planning.