In a major shift that redefines the AI landscape, OpenAI has ended its seven-year cloud exclusivity deal with Microsoft. As of late April, the partnership transitioned from an exclusive to a non-exclusive arrangement, granting OpenAI the ability to offer its full suite of AI models on competing platforms like Amazon Web Services and Google Cloud.
The restructuring, announced jointly by both firms, effectively resolves a brewing legal dispute that emerged after OpenAI secured a massive $50 billion investment from Amazon in February. That deal had given AWS exclusive third-party cloud distribution for Frontier, OpenAI’s enterprise agent platform, which conflicted with the prior Microsoft agreement.
Under the new terms, Microsoft will hold a non-exclusive license to OpenAI’s intellectual property through 2032. OpenAI must still deliver new models to Azure first, but now it can also provide them via AWS Bedrock and eventually Google Cloud. Amazon CEO Andy Jassy confirmed that OpenAI models will be available on AWS Bedrock within weeks.
Financially, Microsoft will no longer receive a revenue share from OpenAI, while OpenAI will continue paying Microsoft until 2030, subject to an undisclosed cap. Microsoft retains its roughly 27% stake in OpenAI’s for-profit entity, which generated $7.5 billion in revenue last quarter.
OpenAI’s chief revenue officer, Denise Dresser, noted that the previous exclusivity limited the company’s ability to meet enterprise demand. AWS CEO Matt Garman echoed this, stating that customers have long requested access to OpenAI models on AWS. Google Cloud is currently reviewing the new terms to explore possible partnerships.
This strategic pivot underscores the growing tension between the two tech giants, as their product lines increasingly overlap—from GitHub Copilot versus OpenAI’s Windsurf to competing proprietary LLMs. The new flexibility is expected to benefit enterprises that previously had to rely solely on Azure for OpenAI access.