• OpenAI will prototype “frontier AI” solutions aimed at both combat and enterprise purposes, under a one-year, $200 million contract .
• The deal aligns with the launch of OpenAI for Government and coincides with the company’s broader pivot towards defence, including policy updates and prior partnerships.
What happened: US taps OpenAI for frontier AI solutions
On 17 June 2025, the US Department of Defense revealed a one-year, $200 million contract with OpenAI to co-create advanced “frontier AI” prototypes addressing national security challenges across both war-fighting and administrative spheres. Tasks include enhancing cyber-defence, streamlining service-member healthcare, optimising logistics, and refining acquisition processes, all to be completed by July 2026 in the Washington, DC area.
The contract emerged from a competitive bidding process with twelve bids, demonstrating OpenAI’s growing stature amongst national-security contractors. Concurrently, OpenAI launched “OpenAI for Government” to consolidate its expanding public-sector engagements, from NASA to the NIH. importantly, OpenAI has clarified that its internal usage policies continue to bar any development of weapons or intent to inflict harm.
Also read: OpenAI appoints ex-Facebook VP as new applications CEO
Also read: OpenAI and NTT Data launch AI solutions in Japan
Why it’s important
This marks OpenAI’s first direct engagement with the Pentagon, signalling a momentous shift from its earlier stance of avoiding military use . The policy change in 2024—removing prohibitions on military applications—has paved the way for such collaborations, including a separate tie-up with Anduril on anti-drone systems.
The scale of the award, one of the largest ever DoD deals for a software company, highlights OpenAI’s meteoric rise as the organisation reports a $10 billion annualised runrate and aggressive fundraising, potentially reaching a $300 billion valuation supported by SoftBank funding.
Analysts view this collaboration as part of a broader push by the US government to integrate commercial AI into defence, alongside similar partnerships involving Anthropic, Google, and Meta . It’s also likely to pit OpenAI against governmental tech stalwarts like Palantir—this contract alone