April 27, 2026
BlogMicrosoft, OpenAI Amend Their Agreement Again
Microsoft and OpenAI have amended their partnership agreement yet again. While Microsoft will remain OpenAI’s primary cloud partner and OpenAI products will ship first on Azure unless Microsoft is unable or unwilling to supply necessary capabilities, OpenAI now has Microsoft’s official permission to provide all of its services to customers across any cloud provider.
OpenAI already had deals in place with AWS and Google Cloud, so the new deal, announced April 27, seems to give what was already happening Microsoft’s official blessing.
Last summer, Google and OpenAI struck a deal via which OpenAI would use Google’s cloud for running its services. In February 2026, AWS and OpenAI announced that AWS would be the exclusive third-party cloud distribution provider for OpenAI Frontier, which is OpenAI’s platform for building, deploying and managing teams of AI agents. And OpenAI said it planned to buy 2 gigawatts of Trainium (Amazon’s purpose-built chip for AI) capacity for certain workloads. In return, Amazon said it planned to invest $50 billion in OpenAI. Microsoft made some noises about disputing the AWS-OpenAI deal, but didn’t take legal action.
Under the latest arrangement, Microsoft continues to have a license to OpenAI IP for models and products through 2032. But Microsoft’s license is no longer exclusive, and it no longer has to pay a revenue share to OpenAI. However, OpenAI does have to continue to make capped revenue-share payments at the same percentage to Microsoft through 2030, “independent of OpenAI’s technology progress,” (which means the “until AGI has achieved” clause is gone).
Late last year, Microsoft and OpenAI had amended their partnership so that OpenAI could turn itself into a for-profit business. Microsoft and OpenAI agreed Microsoft would no longer have right of first refusal for cloud workloads, even though OpenAI said it would spend $250 billion on Azure services. Microsoft also had exclusive access to OpenAI’s IP and models until OpenAI and Microsoft deemed that OpenAI hit the loosely defined artificial general intelligence (AGI) milestone.
A Good or Bad Deal for Microsoft and Customers? It Depends.
Microsoft used OpenAI to achieve a strong AI presence, riding the popularity of OpenAI’s GPT model by incorporating it into its own products, such as Azure OpenAI. Lately, however, its OpenAI partnership benefits have seemed a lot less clear-cut.
The latest update to the pair’s agreement “should help distract regulators who have been concerned about Microsoft pushing classic lock-in up to this point,” said Directions on Microsoft analyst Wes Miller. “I also think it’s also a bit of mutual acknowledgement (like it or not) that neither Microsoft nor OpenAI are actually in a clear leadership position anymore.”
Directions’ analyst David Berry had a somewhat different take.
“There is a general shift to multi-model/multi-vendor to reduce lock-in risks. While both benefit in that regard, Microsoft can pick the best of the best models to offer while OpenAI has to be the best of the best to attract cloud providers,” Berry said.
“Microsoft is reducing its exposure to OpenAI and is working with Anthropic, which arguably is the industry leader now,” he added.
To date, Microsoft has invested $13 billion in OpenAI. It had been serving as its exclusive cloud provider for the past several years.
During its second quarter FY’26 earnings call, Microsoft officials disclosed that its backlog of cloud orders more than doubled in Q2, to a whopping $625 billion. Of this total, 45 percent was attributable to OpenAI’s Azure commitments over multiple years, which worried some market watchers who have been cautious about OpenAI’s plans to become profitable.
