OpenAI has entered into an agreement with Amazon Web Services (AWS) to market its AI solutions to the U.S. government for both classified and unclassified projects.

AWS has verified the agreement to TechCrunch. The Information was the initial outlet to announce.

This collaboration follows OpenAI’s arrangement with the Pentagon, permitting the military to utilize its AI models within its classified systems — a notable achievement amid tensions between Anthropic and the Department of Defense (DOD). Anthropic has subsequently been identified as a supply-chain threat by the DOD after it maintained its stance against deploying its technology for extensive surveillance of Americans and the development of fully autonomous weapons. In turn, Anthropic has filed a lawsuit against the Pentagon.

OpenAI’s partnership with AWS sees the AI leader entering Anthropic’s primary domain. Amazon has poured at least $4 billion into Anthropic, which relies on AWS as its primary cloud provider. Claude models are incorporated into Amazon Bedrock, AWS’s platform for AI in enterprise and government sectors, with Claude being one of the most thoroughly integrated frontier models in AWS GovCloud for public domain applications.

The collaboration not only enables OpenAI to assist the Pentagon with its new agreement, but it also broadens the AI firm’s reach within the federal sector, allowing it to cater to various government agencies through AWS’s established cloud framework. An AWS representative informed TechCrunch that AWS, a principal cloud service provider to U.S. agencies, has committed to distributing OpenAI offerings across its public-sector clientele, which encompasses Amazon Bedrock in government cloud settings such as AWS GovCloud and AWS Classified Regions for Secret and Top Secret operations, according to an OpenAI representative.

The OpenAI representative shared with TechCrunch that while its models will be accessible via AWS, OpenAI will maintain oversight of its technology by determining which models are accessible. AWS is also required to give prior notice before enabling particularly sensitive government entities, including intelligence agencies. OpenAI will liaise directly with clients regarding deployment terms, security stipulations, and operational conditions, and may mandate additional precautions for specific deployments.

This agreement could pave the way for more enterprise contracts, as organizations frequently regard government contracts as evidence of trustworthiness and dependability.

This article has been revised to include remarks and additional details from OpenAI and AWS.