OpenAI rewrites Microsoft partnership terms

The big thing here is cloud distribution — and who controls access to OpenAI’s models. For years, Microsoft had the privileged seat. Azure was the exclusive cloud, Microsoft had an exclusive IP license, and the whole partnership looked less like a vendor deal and more like a fused business structure. That changed on April 27, when both companies publicly rewrote the terms. Microsoft is still central. But the exclusivity is gone. ### What actually changed? OpenAI can now serve all of its products across any cloud provider. That is the cleanest line in the new deal, and it is the one that matters most. It means OpenAI is no longer boxed into Azure-only distribution and can sell through rival infrastructure, including AWS and Google Cloud. Microsoft, meanwhile, keeps a license to OpenAI IP for models and products through 2032, but that license is now non-exclusive. ### Why is “non-exclusive” such a big deal? Because exclusivity was the moat. Microsoft’s old position gave it a special hold over how OpenAI technology reached customers. A non-exclusive license means Microsoft still has rights, but not rights nobody else can get. Basically, Microsoft remains a preferred partner, not the sole gatekeeper. That is a very different power balance in enterprise AI. ### Does Microsoft still get anything special? Yes — quite a lot. Microsoft remains OpenAI’s primary cloud partner, and OpenAI products are still supposed to ship first on Azure unless Microsoft cannot or chooses not to provide the needed capabilities. So this is not a breakup. It is more like OpenAI cutting open escape hatches while keeping Microsoft in the inside lane. ### What changed on the money side? Microsoft said it will no longer pay a revenue share to OpenAI. That is one of the stranger but important pieces, because the old arrangement had money moving both ways under a tangle of commercial terms. The reset simplifies that part of the relationship and suggests both companies wanted fewer structural knots as OpenAI scales beyond a single-cloud channel. ### Why would OpenAI want this now? Scale. OpenAI needs more distribution, more compute options, and more negotiating leverage. If your products are in huge demand, being tied too tightly to one cloud starts to look like a bottleneck. Multi-cloud access lets OpenAI reach customers where they already run workloads instead of forcing everything through Azure. That matters even more as enterprise buyers try to avoid depending too heavily on one vendor. ### Why would Microsoft agree? Because keeping most of the relationship is better than forcing a fight it might lose. Microsoft still has deep commercial ties, first-shipping priority on Azure, and IP rights through 2032. It also avoids being seen as the company choking off OpenAI’s growth just as the AI infrastructure race gets more crowded. Turns out a looser alliance may preserve more value than a brittle exclusive one. ### What does this mean for AWS and Google? It means they can now compete to host and distribute OpenAI products directly. That does not automatically make them equal to Azure overnight, but it changes the conversation from “can they participate?” to “how much business can they win?” Forbes noted OpenAI moved onto AWS Bedrock almost immediately after the exclusivity era ended, which shows this was not a symbolic clause change. ### So what is the bottom line? The old Microsoft-OpenAI deal helped define the first phase of the generative AI boom. The new one defines the next phase — less lock-in, more channels, and more room for OpenAI to behave like a platform company instead of a captive asset. Microsoft is still in the room. It just no longer owns the whole room.