Microsoft and OpenAI End Exclusive Cloud Deal — Microsoft Stops Paying Revenue Share, OpenAI Free to Sell on AWS and Google (April 2026)

Microsoft and OpenAI on April 27, 2026 announced an amended partnership agreement that ends Microsoft's exclusive right to host and resell OpenAI's models, lets OpenAI serve customers from any cloud provider including Amazon and Google, and stops Microsoft's revenue-share payments to OpenAI — while keeping OpenAI's revenue share to Microsoft running through 2030, subject to a newly added total cap.

What Happened

In a joint announcement published on the Official Microsoft Blog, the two companies described the deal as “an amended agreement to simplify our partnership and the way we work together.” The new terms, which Bloomberg reported earlier in the day, dismantle two of the most consequential clauses of the original 2019 Microsoft-OpenAI pact: cloud exclusivity and the cross-payment revenue share.

Microsoft Vice Chair Brad Smith and OpenAI CEO Sam Altman framed the changes as giving each side “flexibility, certainty and a focus on delivering the benefits of AI broadly.” Investors interpreted them differently. Microsoft (NASDAQ: MSFT) shares dropped roughly 3% in pre-market trading on the news, while Amazon climbed about 1% on expectations that OpenAI will move a meaningful chunk of its inference workload onto AWS, according to Seeking Alpha.

Key Details

• Cloud exclusivity ends: OpenAI can now serve all of its products to customers on any cloud provider. Microsoft remains the “primary cloud partner” and OpenAI products will still “ship first on Azure unless Microsoft cannot and chooses not to support the necessary capabilities.”
• License extended but non-exclusive: Microsoft retains a license to OpenAI's IP — both models and products — through 2032, but the license is no longer exclusive.
• Microsoft stops paying OpenAI a revenue share on OpenAI products it resells via Azure.
• OpenAI keeps paying Microsoft a revenue share through 2030, at the same percentage as before but “subject to a total cap,” independent of OpenAI's technology progress.
• Microsoft remains a major OpenAI shareholder and the two companies continue joint work on data-center capacity, next-generation silicon, and applying AI to cybersecurity.
• Legal pressure unblocked: per TechCrunch, the change clears the path for OpenAI's previously announced ~$50B compute deal with Amazon Web Services that Microsoft's exclusivity clause had jeopardized.

What Developers and Users Are Saying

The news lit up the front page of Hacker News, where the top thread reached 454 points and 379 comments within five hours. The dominant developer sentiment is mixed: many welcome the prospect of OpenAI APIs being available natively on AWS Bedrock and Google Vertex AI, simplifying multi-cloud architectures and reducing Azure lock-in. Others worry about the practical impact on Azure OpenAI Service customers who built around guarantees that Microsoft's exclusivity used to imply.

On X/Twitter, OpenAI president Greg Brockman described the deal as “a meaningful simplification.” Microsoft CEO Satya Nadella said in a memo cited by Windows Central that “the work we're doing together remains ambitious.” Industry analysts at The Register read the new terms as Microsoft hedging against OpenAI commoditization risk, while Motley Fool argued Microsoft's AI moat is now “measurably narrower.”

What This Means for Developers

For application developers, the immediate consequence is that OpenAI's models — ChatGPT, the GPT family, the o-series reasoning models, Sora, and the upcoming GPT-5.5 endpoints — will become available natively on AWS, Google Cloud, and other providers, not just Azure OpenAI Service. Teams that have been forced to either pay Azure premiums or proxy through OpenAI's first-party API will gain real choice. Latency-sensitive workloads should benefit from being able to co-locate OpenAI inference next to existing infrastructure on AWS or GCP regions.

For Azure OpenAI Service customers specifically, Microsoft has confirmed Azure will still get OpenAI products first, so the “day-one model availability” advantage is preserved. The bigger long-term question is whether Microsoft will lean harder into its in-house MAI models and Mistral / Anthropic partnerships now that revenue share to OpenAI is off the table. Enterprises with large Azure commits should expect Microsoft to push first-party Copilot Studio and MAI alternatives more aggressively than before.

What's Next

Both companies say the structural amendment is now in effect, but contractual transitions, cloud-region availability of OpenAI APIs on AWS and GCP, and the precise revenue-share cap figure remain to be disclosed. Watch for OpenAI's next blog post on multi-cloud availability and for AWS re:Invent 2026 in late November, where an official OpenAI-on-Bedrock launch is widely expected. Microsoft's earnings call on April 29, 2026 — just two days from the announcement — will be the first chance for analysts to press CFO Amy Hood on what the cap means for Azure margins.

Sources

• The Official Microsoft Blog — primary announcement of the amended Microsoft-OpenAI agreement
• Bloomberg — original report breaking the story
• CNBC — details on the revenue-share cap and exclusivity end
• TechCrunch — how the change unblocks OpenAI's $50B AWS deal
• The Register — analysis of the new partnership structure
• 9to5Google — implications for Google Cloud as a future OpenAI host
• Hacker News discussion — 454-point developer reaction thread