On Monday, Microsoft and OpenAI announced that they have, once again, renegotiated the deal binding the two companies. Despite some opinions on X that frame it as a victory for the ChatGPT maker over the Windows giant, both sides are walking away winners.

Most important, the new terms solve an issue that was hanging over OpenAI’s head since it signed its up-to-$50 billion dollar deal with Amazon.

With this new deal, instead of Microsoft having exclusive access to all of OpenAI’s products and IP until the magical day when OpenAI produces AGI, it’s partnership has a definitive timeline. This contract gives Microsoft a non-exclusive license to OpenAI IP for models and products through 2032.

The two companies are still calling Microsoft OpenAI’s “primary cloud partner,” meaning that the bulk of OpenAI’s cloud will likely be served by Azure for the six years this deal covers, even as OpenAI rushes to build its own data centers with other partners. In October, OpenAI agreed to buy an additional $250 billion worth of Microsoft’s cloud. This line is a message to Microsoft shareholders that OpenAI will still be an enormous Azure customer.

OpenAI products will ship “first on Azure, unless Microsoft cannot and chooses not to support the necessary capabilities,” the companies say. But, critically, “OpenAI can now serve all its products to customers across any cloud provider.”

Again, “first” is not defined clearly in this announcement, whether that means exclusive on Azure only for some time period or just that Microsoft will also be among the vendors carrying OpenAI’s latest products.

But the most important part of this term: it solves the possibility that Microsoft could sue OpenAI over the AI lab’s deal with Amazon.

To recap that messiness: In February, OpenAI announced that Amazon was investing up to $50 billion in the model maker, comprised of a $15 billion initial investment and another $35 billion “in the coming months when certain conditions are met,” the companies said, without specifying what those conditions were.

In exchange, OpenAI agreed to co-develop a “stateful runtime technology” on AWS Bedrock (the AWS service that serves up various AI models and services). Stateful runtime is the tech that supports AI agents, allowing them to remember tasks and contexts for long periods of time.

OpenAI also promised that AWS would have exclusive rights to serve up OpenAI’s new agent-making tool Frontier. And there’s the rub.

OpenAI’s initial agreement with Microsoft prevented OpenAI from selling Frontier exclusively on AWS, and possibly prevented AWS from selling it at all.

While Microsoft had previously agreed to let OpenAI run certain select products, like consumer ChatGPT, on other cloud providers, it retained exclusive rights to any OpenAI product accessed through an API, such as Frontier.

In fact, the same day that OpenAI announced its AWS deal, Microsoft publicly refuted the AWS exclusively terms, writing (emphasis Microsoft’s):

“Microsoft maintains its exclusive license and access to intellectual property across OpenAI models and products. … Azure remains the exclusive cloud provider of stateless OpenAI APIs. … Any stateless API calls to OpenAI models that result from a collaboration between OpenAI and any third party – including Amazon – would be hosted on Azure. … OpenAI’s first party products, including Frontier, will continue to be hosted on Azure.“

Microsoft also emphasized that its terms were in effect until OpenAI achieved AGI. The Financial Times reported that Microsoft even contemplated legal action if it had to enforce these contract terms.

So, the new agreement eliminates Microsoft’s exclusive rights and solves the AWS legal peril. In a post on X, Amazon CEO Andy Jassy celebrated the deal, adding that it meant OpenAI’s models would become available to customers on AWS Bedrock.

While this deal is good for OpenAI, Microsoft walked with some wins, too. The new deal now allows Microsoft to stop paying a revenue share to OpenAI, while OpenAI will continue to pay revenue share to Microsoft through 2030, although this is now subject to a cap.

Exactly how much cash will flow to Microsoft is hard to tell, but it’s likely in the billions. Last quarter, Microsoft reported that it made $7.5 billion in a single quarter from its investment in OpenAI.

The kicker is that Microsoft remains a major shareholder in OpenAI, owning about 27 percent of the for-profit entity, it said in October. It financially benefits from OpenAI’s growth, even the sales it makes on AWS.

The downside, of course, is that Microsoft loses out on any extra cloud services it would be able to sell as a result of an exclusive deal with OpenAI.

That may not matter much. Just like OpenAI has been courting Microsoft’s biggest competitors, Microsoft has a new, cozy relationship with OpenAI rival Anthropic for cloud giant to use its Claude AI to power agentic products.

The biggest winners here are enterprises, who get to choose their models and their clouds while the giants compete with each other to serve them.

Here’s a timeline of the recent changes in Microsoft’s relationship with OpenAI.

In October, Microsoft and OpenAI announce a new agreement to help OpenAI fend off the lawsuit from Elon Musk about its corporate structure that gives OpenAI the ability to run non-API-accessed products on other clouds.

In November, OpenAI and Amazon sign their first multi-year agreement, in which OpenAI contracts for $38 billion worth of AWS cloud.

In February, Amazon announces an up to $50 billion investment in OpenAI, pending “certain conditions” including the exclusive tech development and hosting deal for Frontier and stateful tech. On the same day, Microsoft refutes that AWS will have that tech exclusively.

In March, FT publishes that Microsoft is considering legal action.

In April, OpenAI and Microsoft announce a new deal, that includes a calendar end date for their exclusive partnership and allowing OpenAI to run all of its products on other clouds. Microsoft no longer has to pay OpenAI revenue share. Microsoft remains a major shareholder in OpenAI.