Microsoft and OpenAI loosen their ties

Silicon Valley’s most consequential partnership is officially loosening its tightest knots.

OpenAI announced an amended agreement with Microsoft designed to bring long-term predictability to their massive alliance, trading rigid exclusivity for operational flexibility.

Multi-cloud freedom meets non-exclusive IP

The biggest headline is that OpenAI is no longer locked exclusively to Azure.

Under the new terms, OpenAI gains the freedom to serve all of its products to customers across any cloud provider. Azure still keeps its throne as OpenAI’s primary cloud partner, meaning OpenAI products will continue to ship first on Microsoft’s platform, unless Microsoft cannot or chooses not to support the required capabilities.

In return, Microsoft extended its license to OpenAI intellectual property for models and products through 2032. The catch is that this license is now non-exclusive, giving OpenAI the legal room to work more openly with the rest of the tech ecosystem.

Reworking the revenue split

The two tech giants also untangled their complex financial revenue-sharing setup.

Microsoft will completely stop paying a revenue share to OpenAI. Meanwhile, OpenAI’s revenue share payments to Microsoft will continue through 2030 at the same percentage, completely independent of OpenAI’s technical milestones, but now capped at an agreed total limit.

Microsoft remains a major shareholder directly benefiting from OpenAI’s growth, and their physical infrastructure ties are not slowing down. Both companies confirmed plans to keep scaling gigawatts of datacenter capacity and co-developing next-generation AI silicon.

Why it matters

For enterprise buyers and developers, this amendment opens up much-needed breathing room. You will no longer be boxed strictly into Azure if you want to deploy OpenAI’s cutting-edge tooling at scale, which makes hybrid and multi-cloud strategies far more viable.

For the broader industry, it effectively redraws the cloud battlefield. AWS and Google Cloud can now realistically vie for OpenAI product workloads that were previously off-limits. At the same time, Microsoft locking down model IP through 2032 guarantees safety for Copilot without keeping OpenAI on a corporate leash.

We think this is a pragmatic reset rather than a breakup. The original deal was hammered out when OpenAI desperately needed chips and Microsoft desperately needed frontier models. Today, OpenAI wants to be a global enterprise utility, and Microsoft is diversifying its own AI portfolio. Giving up exclusivity prevents regulatory headaches and allows both companies to chase massive datacenter scale without stepping on each other’s toes.