OpenAI Revises Microsoft Partnership, Ends Revenue Sharing Arrangement

Reports say OpenAI is bringing its generative AI models to Amazon Web Services after revising its Microsoft agreement, widening cloud access while Azure remains the first launch venue for OpenAI products.

Fact Check
The claim is strongly supported by Reuters and an official OpenAI source. Reuters' April 27, 2026 report states that Microsoft will no longer receive/pay revenue sharing under the revised arrangement, that its access/license is no longer exclusive, and that OpenAI products will generally launch on Azure first unless Microsoft cannot or does not provide them. The official OpenAI page 'Joint Statement from OpenAI and Microsoft' corroborates the revenue-sharing change and Azure-first/exclusive API cloud terms in the search snippet. The only limitation is that the OpenAI page could not be fetched directly due to a WAF, but the Reuters article independently confirms the same core facts.
    Reference123
Summary

OpenAI is expanding availability of its generative AI models to Amazon Web Services, according to reports, shortly after revising its partnership with Microsoft to end revenue sharing and make Microsoft’s license to OpenAI technology non-exclusive. OpenAI said the changes are intended to simplify the relationship, while Microsoft said it will continue deepening the partnership and that OpenAI products will still launch first on Azure. The Information previously reported that the companies also agreed to exclude technology OpenAI gains through acquisitions after negotiations between Satya Nadella and Sam Altman tied to disagreements over OpenAI’s AWS deal.

Terms & Concepts
  • Non-exclusive license: A licensing arrangement that allows the owner to grant similar rights to multiple parties rather than limiting access to one partner.
  • Amazon Web Services: Amazon’s cloud computing platform, which provides infrastructure and services for hosting software, data, and AI workloads.
  • Azure: Microsoft’s cloud computing platform, used to host applications, data, and artificial intelligence services at scale.