OpenAI Microsoft exclusivity ends after seven years

Microsoft and OpenAI restructured their landmark 2019 partnership on April 27, converting Microsoft’s cloud license from exclusive to non-exclusive, allowing OpenAI to sell its full model suite on Amazon Web Services and Google Cloud for the first time, as Amazon CEO Andy Jassy confirmed OpenAI models will arrive on AWS Bedrock within weeks.

OpenAI Microsoft released a joint statement on April 27 announcing the partnership overhaul that ends seven years of effective Azure exclusivity. As crypto.news reported, OpenAI’s models and its Codex agent are already available to AWS customers through Amazon Bedrock, with Amazon Bedrock Managed Agents powered by OpenAI allowing enterprises to build autonomous AI agents within AWS infrastructure. OpenAI chief revenue officer Denise Dresser told staff in an internal memo that the Microsoft arrangement “has limited our ability to meet enterprises where they are,” and that inbound demand for the AWS offering has been “frankly staggering.” “This is what our customers have been asking us for for a really long time,” AWS CEO Matt Garman said at a San Francisco launch event on April 28. The revised deal converts Microsoft’s license from exclusive to non-exclusive through 2032. Microsoft no longer shares revenue with OpenAI, while OpenAI continues paying Microsoft through 2030. Google Cloud is now reviewing the terms to assess what partnership is possible under the new structure, according to Reuters.

The financial and legal mechanics of the restructuring are precise. As TechCrunch noted, Microsoft was considering legal action against OpenAI after the February Amazon deal granted AWS exclusive third-party cloud distribution for Frontier, OpenAI’s enterprise agent platform. The April 27 restructuring eliminates that legal overhang by making Microsoft’s license explicitly non-exclusive. Microsoft retains roughly 27% of OpenAI’s for-profit entity and last quarter reported $7.5 billion in revenue from its OpenAI stake. For enterprises, the practical change is immediate: cloud workloads that previously required Azure to access OpenAI models can now run on AWS or eventually Google Cloud. As crypto.news documented, the two companies had been building toward this restructuring since June 2025 when their increasingly competitive product lines, from GitHub Copilot versus OpenAI’s Windsurf to Microsoft’s own proprietary LLMs, made the exclusivity arrangement structurally untenable for both sides.