OpenAI (Unlisted:OPAI) has signed a new agreement with Amazon Web Services (AWS) to offer its AI models to US federal agencies, including the Department of Defense, for both classified and unclassified work, the company said Tuesday.
The deal follows a Pentagon contract OpenAI secured late last month, after the department ended its relationship with Anthropic, a previous AI provider.
Anthropic had worked with Palantir and AWS to deploy its Claude models for military and intelligence systems but fell out of favor after restricting military applications, including domestic surveillance and autonomous weapons. The Pentagon subsequently labeled Anthropic a “supply chain risk.”
Under the new arrangement, OpenAI’s AI models, such as those powering ChatGPT, will run on AWS infrastructure, including Amazon’s proprietary AI chips, Trainium.
OpenAI has committed to substantial multi-year compute purchases, and AWS will serve as the exclusive third-party cloud distributor for OpenAI’s Frontier enterprise platform.
The partnership allows OpenAI to make its AI tools available to approximately 3 million Defense Department employees and expands access across civilian agencies by leveraging AWS’s existing government contracts and security clearances. The deal is also seen as enhancing OpenAI’s credibility with large corporate clients.
Shares of Amazon.com Inc (NASDAQ:AMZN) traded up 1% at $214 following the announcement.