OpenAI and Microsoft loosen the grip: the AI platform becomes less exclusive

OpenAI announced an amended agreement with Microsoft on April 27. This is not a breakup. It is more interesting than that: the partnership becomes more flexible, less exclusive, and better suited to an AI market where customers do not want to be locked into one cloud.

What changes

Microsoft remains OpenAI’s primary cloud partner, and OpenAI products will still ship first on Azure when Microsoft can support the required capabilities. But OpenAI can now serve all its products to customers across any cloud provider.

Microsoft keeps a license to OpenAI models and products through 2032, but that license is now non-exclusive. Microsoft will also no longer pay revenue share to OpenAI. OpenAI will continue revenue share payments to Microsoft through 2030, but with a total cap.

Why it matters

For CIOs, this is about vendor risk. OpenAI becomes less tightly bound to Azure. Microsoft still gets access to IP and remains a major shareholder, but it loses the exclusivity that made the partnership easier to read.

There are three practical consequences:

• Multi-cloud AI becomes more realistic, not just a procurement wish.
• Azure remains important, but not necessarily the only path to OpenAI capacity.
• Microsoft customers should separate Copilot strategy, Azure strategy, and OpenAI strategy.

If you are building AI architecture now, this is a signal to avoid a hard single-vendor lock-in. Use Microsoft where governance, identity, and the data layer fit. Use OpenAI directly where model access and product velocity matter most. And make sure your agent layer can swap model and cloud without rebuilding the whole stack.

The old AI deal was easy to explain: OpenAI plus Microsoft. The new one is more grown up: OpenAI, Microsoft, Azure first, but not Azure alone.