OpenAI breaks Azure tie — its models are coming to AWS and (soon) Google Cloud
In a major reshuffle announced April 27, OpenAI and Microsoft retooled their landmark 2019 partnership, turning Microsoft’s cloud license from exclusive into a non‑exclusive deal. The change clears the way for OpenAI to offer its full model suite on Amazon Web Services and, potentially, Google Cloud — with Amazon CEO Andy Jassy confirming OpenAI models will land on AWS Bedrock “within weeks.”
What changed
- Microsoft’s license has been explicitly converted from exclusive to non‑exclusive through 2032.
- Microsoft still holds roughly 27% of OpenAI’s for‑profit entity and last quarter reported $7.5 billion in revenue tied to its OpenAI stake.
- Under the new terms Microsoft will no longer share revenue with OpenAI, while OpenAI will continue making payments to Microsoft through 2030.
- Google Cloud is reviewing the revised terms to determine what kind of partnership it can form under the new structure.
Immediate impact for enterprises
Cloud workloads that previously had to run on Azure to access OpenAI models can now run on AWS — and, in time, on Google Cloud. OpenAI’s models and its Codex agent are already available to AWS customers through Amazon Bedrock, and Amazon Bedrock Managed Agents powered by OpenAI let enterprises build autonomous AI agents within AWS infrastructure. OpenAI’s chief revenue officer Denise Dresser told staff that the Microsoft arrangement “has limited our ability to meet enterprises where they are,” adding that inbound demand for the AWS offering has been “frankly staggering.” AWS CEO Matt Garman echoed the sentiment at a San Francisco launch event, saying “This is what our customers have been asking us for for a really long time.”
Why the restructuring happened
The move resolves a growing commercial and legal tension between the companies. After OpenAI’s February deal with Amazon granted AWS exclusive third‑party cloud distribution for Frontier — OpenAI’s enterprise agent platform — Microsoft had considered legal action. Making Microsoft’s license non‑exclusive removes that legal overhang, TechCrunch reported, and reflects the reality that the two companies had been drifting toward a split since June 2025 as competing products (from GitHub Copilot versus OpenAI’s Windsurf to Microsoft’s proprietary LLMs) made strict exclusivity untenable.
What to watch next
- AWS: OpenAI models are rolling into Bedrock and enterprises can begin building on AWS immediately.
- Google Cloud: now evaluating the new terms; a partnership is possible but not finalized.
- Market effects: broader cloud distribution could accelerate enterprise adoption of OpenAI models and increase competition among cloud providers for AI workloads.
Bottom line
The April 27 agreement ends years of effective Azure exclusivity and opens OpenAI’s tech to a wider range of cloud platforms. For enterprises, it means greater choice in where they run AI workloads; for the cloud providers, it ushers in a new phase of competition for hosting next‑generation AI services.
Read more AI-generated news on: undefined/news
