Microsoft has been investing in OpenAI for the past seven years, pouring $13.75 billion into the generative AI pioneer over that time as the two companies more tightly tied their AI futures to each other through various exclusivity deals.

However, those same deals that gave each one a running start in the booming field a few years later were holding them back as the AI industry has expanded, trillions of dollars have streamed into the market, the number of AI companies has grown, and trends like more advanced models, AI agents, and reasoning AI grow the boundaries of what the technology can do.

The AI space of today barely resembles what it looked like even two years ago, and exclusivity deals like that between Microsoft and OpenAI, which once promised a fast-track into AI’s future, now hold them back.

Given that, Microsoft and OpenAI announced an amended partnership that will still give them some of the benefits of the relationship but allow them work with other AI vendors and cloud providers. Essentially, they mutually agreed to remain friends, but now they can see others. There will be benefits to both.

“The rapid pace of innovation requires us to continue to evolve our partnership to benefit our customers and both companies,” both Microsoft and OpenAI wrote in identical statements, adding that the new deal will “simplify our partnership and the way we work together, grounded in flexibility, certainty and a focus on delivering the benefits of AI broadly. The greater predictability in the amended agreement strengthens our joint ability to build and operate AI platforms at scale while providing both companies the flexibility to pursue new opportunities.”

What Microsoft And OpenAI Get

That means that while Microsoft – which will continue as a major shareholder – will stay OpenAI’s primary cloud provider, OpenAI can now run all of its products with any of the others, from Amazon Web Services and Google Cloud to neocloud GPU-as-a-service providers like CoreWeave, Lambda, and Vast.ai. OpenAI offerings will ship first on Azure, unless Microsoft decides otherwise.

Through 2032, Microsoft will continue to have a license to OpenAI code for models and products, though the license will no longer be exclusive.

Here is where the money comes in: Microsoft will no longer have to pay a revenue share to OpenAI for when Microsoft’s customers access OpenAI models through Azure, and while OpenAI will continue pay Microsoft for revenue share payments at the current percentage – 20 percent – that will only continue through 2030, and will come with a cap on the total amount paid.

For Microsoft, the reconfigured OpenAI deal gives it more flexibility to pursue partnerships with other AI vendors. Key among those is vendors is Anthropic.

Courting Anthropic

In November, Microsoft and Anthropic announced a deal aimed at broadening the access users have to Anthropic’s Claude AI models that are running on Azure, with the AI vendor agreeing to buy $30 billion of compute capacity on Microsoft’s cloud infrastructure. Anthropic also agreed to contract more compute capacity, up to a gigawatt. The agreement expanded access to customers of Microsoft Foundry to Anthropic’s Sonnet, Opus, and Haiku frontier models. Microsoft also said it will continue enabling Claude access across the hyperscaler’s Copilot lineup, including, GitHub Copilot, Microsoft 365 Copilot, and Copilot Studio, and also agreed to invest up to $5 billion into Anthropic.

Microsoft also has partnerships with other model vendors, including Meta, Mistral AI, and Hugging Face, but the alliance with Anthropic is important because it has become the most significant competitor to OpenAI. It’s also established itself in such areas a software development with Claude Code and research and other knowledge work with Claude Cowork. With its high-profile Mythos Preview, the company also has grabbed onto the cybersecurity field, introducing a frontier model that is significantly better than others at detecting and identifying security vulnerabilities in software. It’s also very good at creating exploits for those vulnerabilities, which convinced Anthropic to limit the release of the model to particular organizations rather than giving with broad availability.

A week later, OpenAI unveiled GPT-5.4-Cyber, a variant of its latest frontier model that also for the cybersecurity community. Having more leeway – and money – to work with both model vendors in this and other areas will be important for Microsoft. Removing the exclusivity to OpenAI and not having to pay out a share of its revenue for reselling access to ChatGPT models puts Microsoft in a better position to compete with the likes of Google, which this week said it is investing up to $40 billion in Anthropic, a move that comes days after Amazon made a $5 billion investment.

The Forthcoming IPOs

It also gives Microsoft more running room to develop and run more of its own AI models, a time when rivals like Meta Platforms – with open source Llama and closed source Muse Spark– and Google via Gemini are aggressively pushing ahead with their AI roadmaps. Microsoft can be seen as something more than a gateway into OpenAI’s models.

In the background of all these moves are the ongoing plans for both OpenAI and Anthropic to go public, IPOs that are rumored to net each as much as a trillion or more dollars. Being able to work more closely with a broader rang of cloud providers and not be so tightly identified with Microsoft could help crank that number a bit more for OpenAI.

The Evolution Of The Partnership

Microsoft first invested in OpenAI in 2019, when it gave the startup $1 billion and became its exclusive cloud provider. Two months after OpenAI launched ChatGPT in November 2022, kicking off the generative AI era, Microsoft invested another $10 billion over multiple years, a move that brought its total investment to more than $13 billion.

In September 2025, Microsoft and OpenAI restructured its deal to allow the AI model builder to become a public benefit corporation, a move that had Microsoft holding a 27 percent share in OpenAI and letting the startup go public.

That same month saw both OpenAI and Microsoft continuing to pursue AI futures that expand even more beyond each other. OpenAI, as part of an effort to diversify as its compute needs grow, signed a massive $300 billion, five-year cloud computing deal with Oracle set to begin next year.

At the same time, Microsoft signed a $19 billion deal with Nebius Group, an AI infrastructure company in The Netherlands, that will deliver more AI computing capacity to Microsoft through 2031.

Now with the latest reconfiguring the of the deal, both Microsoft and OpenAI see much more wide-open AI fields to play in. We will have to wait to see where that freedom takes them.