OpenAI and Microsoft announced a sweeping overhaul of their seven-year partnership on April 27, ending the exclusive cloud licensing deal that has shaped AI infrastructure distribution since 2019. Under the amended agreement, OpenAI can now serve all of its products to customers across any cloud provider, including Amazon Web Services and Google Cloud.
Revenue Share, Capped
Microsoft retains a 20% revenue share on OpenAI’s products, including every ChatGPT subscription purchase, according to CNBC. The total payout is now subject to an unspecified cap and guaranteed only through 2030. One structural change flips the existing economics: Microsoft will no longer pay revenue share to OpenAI when Azure customers access OpenAI models, a payment flow that previously ran in the opposite direction.
The revenue share is now “independent of OpenAI’s technology progress,” per Ars Technica, a reference to the original “AGI clause” that would have scrapped exclusivity if OpenAI achieved artificial general intelligence. That clause is effectively dead.
Non-Exclusive License Through 2032
Microsoft keeps its license to OpenAI’s intellectual property and models through 2032, but the license is no longer exclusive. Azure remains the “primary cloud partner,” and OpenAI products will ship first on Azure unless Microsoft decides otherwise. The practical effect: OpenAI is now a multi-cloud vendor.
The shift follows months of visible strain. In an internal memo reported by CNBC on April 13, OpenAI revenue chief Denise Dresser wrote that the Microsoft partnership had “limited our ability to meet enterprises where they are.”
AWS Moves Fast
The timing is not coincidental. OpenAI and Amazon formed a major strategic partnership in February, with Amazon investing up to $50 billion and OpenAI committing to expand its AWS agreement by $100 billion over eight years. AWS already serves as the exclusive third-party cloud distribution provider for OpenAI’s enterprise platform Frontier.
Amazon CEO Andy Jassy posted on X Monday that AWS would begin providing OpenAI models through its Bedrock service “in the next few weeks,” with details coming at a San Francisco event on Tuesday.
The Multi-Cloud Calculus
For enterprise buyers running agentic workloads, the deal removes a meaningful constraint. Organizations locked into AWS or Google Cloud infrastructure no longer need Azure as a prerequisite for OpenAI model access. As agent deployments scale from minutes-long tasks to hours-long autonomous sessions, the cloud layer they run on becomes a cost and latency decision rather than a vendor-dictated one.
Microsoft shares were unchanged in Monday’s trading. The company’s investment in OpenAI’s for-profit arm was last valued at $135 billion, roughly 27% of the company on an as-converted diluted basis, according to CNBC.