The following information is summarized from The next chapter of the Microsoft–OpenAI partnership
In October 2025, Microsoft and OpenAI redefined their 2019 partnership agreement as OpenAI switched from a non-profit to a public benefit corporation. Microsoft remains its largest investor (250 billion in Azure services in recompense.
Some of the details:
Microsoft retains exclusive IP rights for some IP and Azure API exclusivity through 2032 or AGI is achieved.
- Microsoft’s rights to OpenAI research IP (confidential development methods) last until either:
- AGI is verified by the panel, or
- 2030 — whichever comes first.
- The deal distinguishes between “research IP” and “non-research IP” but the terms are unclear in the blog post.
- “Research IP” is defined as “confidential methods used in the development of models and systems.”
- “Research IP” does not include model architectures, weights, inference or fine-tuning code, or datacenter hardware/software designs.
- But “Microsoft retains these non-Research IP rights.”
- So does Microsoft retain rights to both types of IP, just with different terms?
The agreement changes when AGI is achieved, but reaching AGI must “be verified by an independent expert panel.”
Microsoft and OpenAI continue to have revenue sharing.
OpenAI may co-develop products with third parties. API-based products must run on Azure (except in the case of U.S. national security applications), but non-API products can run on any cloud provider.
Microsoft may independently pursue AGI, either alone or with partners.
Microsoft no longer has a right of first refusal to be OpenAI’s compute provider.