The Microsoft-OpenAI Partnership: A Decade of Strategic Evolution

When Sam Altman and his colleagues announced the founding of OpenAI in December 2015, Microsoft CEO Satya Nadella immediately noticed something concerning: his company wasn’t invited to the party. “Did we get called to participate?” Nadella wrote to his team, noting that rival Amazon Web Services had “sneaked in there” as a donor. The absence sparked concern that this new AI nonprofit might become tethered exclusively to Amazon’s cloud. For Nadella, who had been CEO for less than two years and was working to expand Microsoft’s Azure cloud platform against AWS’s market dominance, this was a strategic gap that needed addressing. This moment of realization would set in motion one of the most consequential technological partnerships of the decade—one that would eventually involve billions of dollars, reshape both companies, and help launch the generative AI revolution that transformed the tech landscape.

Within a remarkably short time, Microsoft began aggressively courting OpenAI. By August 2016, Altman was emailing OpenAI co-founder Elon Musk about a potential partnership: “I have negotiated a $50 million compute donation from them over the next 3 years!” When asked for his opinion on switching from Amazon to Microsoft, Musk offered his characteristically blunt assessment: “I think Jeff [Bezos] is a bit of a tool and Satya is not, so I slightly prefer Microsoft, but I hate their marketing dept.” Altman explained that Amazon had been “dicking us around” on terms and conditions, particularly regarding marketing commitments, adding that their technical offering wasn’t particularly strong either. By November 2016, Microsoft and OpenAI announced their first partnership, with OpenAI shifting to Azure as its primary cloud platform. Internally, Microsoft viewed this as more than just flipping a high-profile customer from AWS—it saw opportunities for “thought leadership” in AI, a “halo effect” for Azure’s GPU capabilities, and access to new developer audiences. The relationship, however, would face significant challenges as OpenAI’s ambitions and computing needs grew increasingly expensive.

Despite early momentum, the partnership showed signs of strain by 2017-2018. When OpenAI sought Microsoft’s support for a massive computational project to train AI systems to beat human players at Dota 2 (requiring 10,000 servers with the latest Nvidia GPUs), Microsoft’s cloud executives balked at the cost. “I won’t take a complete bath to do it,” wrote Microsoft executive Jason Zander to Nadella. OpenAI contracted with Google instead, and by January 2018, the broader relationship appeared to be fraying. Microsoft’s CTO Kevin Scott complained that OpenAI was treating Microsoft “like a bucket of undifferentiated GPUs, which isn’t interesting for us at all,” while other executives questioned whether OpenAI was making sufficient progress toward artificial general intelligence (AGI). The situation changed dramatically in 2019 when OpenAI restructured to create a “capped profit” entity that could raise capital from investors while theoretically limiting their returns. Microsoft agreed to invest $1 billion with an option for a second billion, securing exclusive cloud computing rights and a commercial license to OpenAI’s technology. This pivotal investment was driven partly by Microsoft’s concerns about Google’s AI progress, as Scott had written to Nadella about “capability gaps” between the companies. The deal also gave Microsoft effective veto power over major decisions at OpenAI, as it controlled 85% of the total investment.

As the partnership deepened, the pressure to commercialize OpenAI’s technology increased. By 2022, according to deposition testimony from Mira Murati (then OpenAI’s VP of applied AI), the company had set a $100 million revenue target that Altman said needed to be hit to justify Microsoft’s next investment. Then came the watershed moment: On November 30, 2022, OpenAI launched ChatGPT, which became the fastest-growing consumer application in history, reaching 100 million users within two months. Suddenly, Microsoft’s multibillion-dollar bet looked prescient. In mid-January 2023, Nadella texted Altman asking when they planned to activate paid subscriptions, as ChatGPT was experiencing overwhelming demand—6 million daily active users with 50 million turned away due to capacity limitations. ChatGPT Plus launched at $20 monthly on February 1, 2023, just after Microsoft had made its landmark $10 billion investment in OpenAI. Ten months later, the partnership faced its greatest test when OpenAI’s nonprofit board abruptly fired Altman as CEO, citing concerns that he had not been “consistently candid in his communications.”

The November 2023 leadership crisis revealed the true depth of Microsoft’s influence over OpenAI’s future, despite having no formal board seat. When Altman was ousted, Microsoft responded with remarkable speed, creating a new subsidiary called Microsoft RAI Inc. within hours—ready to hire Altman and key team members if needed, with $25 billion allocated to match compensation packages. Though publicly maintaining distance from OpenAI’s governance, newly revealed text messages show Nadella, Scott, and Microsoft President Brad Smith actively vetting potential board candidates during the crisis, effectively exercising shadow veto power over who would oversee the nonprofit’s mission. After days of intense negotiations, Altman returned as CEO with a reconstituted board. The relationship would be formalized further in late 2024 when OpenAI announced it would unwind its capped-profit structure in a project internally called “Watershed.” By October 2025, the transaction closed with Microsoft securing a 27% stake in OpenAI at a $500 billion valuation and a commitment from OpenAI for $250 billion in future Azure purchases. In exchange, Microsoft gave up some exclusivity rights, and just six days later, OpenAI signed a seven-year, $38 billion infrastructure deal with AWS—bringing the partnership full circle to where it began.

The evolution of Microsoft and OpenAI’s relationship over this decade illustrates how strategic corporate partnerships can fundamentally reshape technological trajectories. What began with Nadella’s concern about being left out of a new AI lab’s founding ended with Microsoft as the dominant investor in what became one of the most valuable and influential technology companies in the world. OpenAI transformed from a research-focused nonprofit dependent on Microsoft’s cloud to a platform company generating over $20 billion in annual revenue with the leverage to diversify its cloud providers. As OpenAI CFO Sarah Friar noted in late 2025, “Three years ago, we relied on a single compute provider. Today, we are working with providers across a diversified ecosystem.” The partnership that redefined artificial intelligence had itself been redefined, evolving from dependency to interdependence, with both companies fundamentally transformed in the process. What started with Nadella asking if Microsoft had “been called to participate” concluded with nothing at OpenAI happening without the tech giant’s involvement—even as OpenAI secured the independence to ensure it would no longer be Microsoft’s alone.