A Boardroom Crisis Exposes Hidden Power in the Microsoft OpenAI Board Relationship
When Sam Altman was briefly ousted from OpenAI in November 2023, attention focused on the drama inside the research lab. Newly unredacted court documents and Satya Nadella’s testimony now show how deeply Microsoft was embedded in the search for a new OpenAI board. Text threads between Nadella, Microsoft CTO Kevin Scott, Microsoft president Brad Smith, and Altman reveal a rapid-fire effort to rebuild governance while restoring stability to a partnership that underpins products like ChatGPT and Copilot. OpenAI insiders invited Microsoft’s leadership to weigh in on who should sit on the reconstituted board. While Nadella emphasized that the board was free to ignore his advice, the exchanges make clear that Microsoft’s views carried significant weight. The Sam Altman crisis thus became a live test of AI partnership governance, with Microsoft acting as both strategic investor and de facto stabilizer of OpenAI’s leadership structure.
Who Microsoft Backed—and Blocked—for OpenAI’s Board
The unsealed text thread details names Microsoft wanted near OpenAI’s helm—and those it rejected outright. Kevin Scott floated former Google Cloud CEO Diane Greene, only to label her a “strong, strong no” after Nadella objected due to her ties to a major AI competitor. Nadella similarly pushed back on William “Bing” Gordon, citing his long-standing Amazon connections. In contrast, Belinda Johnson, former Airbnb COO, drew praise from Scott, and Nadella endorsed the idea. Nadella himself proposed Sue Desmond-Hellmann, former CEO of the Gates Foundation, who later joined the OpenAI board. Other suggestions included Ursula Burns, Anne Sweeney, Leslie Kilgore, and a slate of seasoned executives spanning tech, finance, and media. Scott even half-jokingly suggested serving himself, an idea Nadella dismissed. Ultimately, Altman proposed Bret Taylor, Larry Summers, and Adam D’Angelo as a three-person board, with Microsoft scrutinising each name for risk to its strategic position.
Satya Nadella’s IBM Warning and Microsoft’s Dependency Risk
Nadella’s testimony adds crucial context to why Microsoft cared so much about OpenAI’s governance. In an internal email from April 2022, he warned against repeating history: he did not want Microsoft to become “the next IBM” while OpenAI turned into “the next Microsoft.” The concern was structural. Microsoft was preparing to invest USD 10 billion (approx. RM46 billion) more into OpenAI, diverting scarce computing resources toward a partner rather than duplicating supercomputers for its own teams. Nadella called this a “one-way door,” acknowledging that Microsoft was effectively outsourcing key intellectual property and taking a massive dependency on OpenAI’s research. His priority was to ensure access to OpenAI’s IP while building internal expertise. That strategic vulnerability explains why Microsoft pushed back on board candidates with strong ties to rival AI players and advocated for directors it viewed as steady, independent, and non-threatening to its long-term position.
AI Partnership Governance as a Tool for Risk Mitigation
The Musk v. Altman lawsuit frames Microsoft’s influence as a betrayal of OpenAI’s original nonprofit mission, alleging that the company helped redirect a charitable trust into a commercial powerhouse. Nadella countered that Microsoft assumed enormous risk to fund an AI lab few others would back, enabling OpenAI to become one of the largest nonprofits and to ship widely used tools. Yet his cross-examination highlighted gaps: Nadella was unaware of full-time nonprofit staff or major open-sourced research before 2026. Regardless of who prevails in court, the case illustrates how AI partnership governance doubles as risk management. With a USD 10 billion (approx. RM46 billion) commitment and deep product integration, Microsoft treated board composition as an extension of its own corporate oversight. The episode underscores a broader reality: in modern AI alliances, capital, infrastructure, and governance are tightly intertwined levers of control.