Unredacted filings expose Microsoft’s hand in the OpenAI governance crisis
Newly unredacted documents from the Musk Altman trial shed rare light on how deeply Microsoft executives were involved in OpenAI’s governance crisis after Sam Altman’s brief removal as CEO. Text messages between Satya Nadella, Kevin Scott, Brad Smith and Sam Altman—once heavily blacked out—now reveal that Microsoft did far more than watch from the sidelines. While OpenAI insiders were scrambling to stabilize leadership, they actively sought Microsoft’s views on the composition of a reconstituted board. In response, Microsoft’s leadership evaluated candidates, signaled preferences and issued firm vetoes, effectively shaping the emerging structure of the Microsoft OpenAI board relationship. Nadella has testified that the board was free to ignore his suggestions, but the communications show OpenAI leadership treating Microsoft as a critical stakeholder whose buy-in was essential to restoring stability and investor confidence during the OpenAI governance crisis.
Satya Nadella’s ‘next IBM’ fear and the stakes of the OpenAI deal
Satya Nadella’s testimony in the Musk Altman trial underscores why Microsoft was so invested in OpenAI’s governance. In an internal email produced at trial, Nadella compared the OpenAI partnership to Microsoft’s early PC-era relationship with IBM. His concern: Microsoft might end up in IBM’s old position, while OpenAI became the next Microsoft. The company was preparing a massive new investment in OpenAI, treating it as a “one-way door” because it could not justify building two separate AI supercomputers. Nadella described the arrangement as effectively outsourcing key intellectual property development and taking a “massive dependency” on OpenAI. That strategic exposure explains why Microsoft insisted on robust access to OpenAI’s IP and why its leadership was so focused on who would oversee the organization. Governance was not abstract; it was directly tied to how Microsoft protected its technological and competitive position.
Vetoes and favorites: How Microsoft filtered OpenAI’s board candidates
The unsealed text thread shows Microsoft executives carefully screening OpenAI board candidates through a competitive lens. When former Google Cloud CEO Diane Greene surfaced as a possibility, Microsoft CTO Kevin Scott responded with a “strong, strong no,” and Nadella later testified he opposed her due to her ties to a major rival in AI. William “Bing” Gordon drew similar pushback because of his deep Amazon connections. Yet other names drew enthusiasm: Belinda Johnson, former Airbnb COO, was praised as “great,” and Brad Smith called former Disney-ABC president Anne Sweeney “solid, thoughtful, calm.” Nadella personally suggested former Gates Foundation CEO Sue Desmond-Hellmann, who ultimately joined the OpenAI Foundation board. Scott floated additional high-profile figures and even half-joked about serving himself, a suggestion Nadella disliked. These exchanges reveal Microsoft acting as an informal gatekeeper, encouraging allies and screening out perceived competitive risks.
Negotiating a new board while rescuing Altman’s leadership
As the crisis progressed, OpenAI leadership and Microsoft began converging on a compact, credibility-boosting board. Sam Altman proposed a three-person structure led by Bret Taylor, Larry Summers and Adam D’Angelo, with himself restored as CEO but kept off the board. Brad Smith quickly raised concerns about Summers, describing him as brilliant but “so mercurial” and therefore risky in a volatile moment. Altman admitted shared misgivings yet signaled he’d accept Summers to stabilize the situation, adding that he wanted to “save this” arrangement. Nadella’s response—asking if he could call Summers directly—shows Microsoft stepping in not just to opine on names, but to personally vet and influence crucial appointments. Summers ultimately joined the board and served until his later resignation, illustrating how Microsoft’s real-time interventions during the crisis translated into concrete, long-term governance outcomes.
Power, trust and the evolving Microsoft OpenAI partnership
The Musk Altman trial frames Microsoft’s actions as part of a broader power struggle over OpenAI’s mission and control. Elon Musk’s legal team argues that Microsoft’s efforts to secure its interests—through heavy investment, IP protections and influence over the Microsoft OpenAI board dynamic—diverted OpenAI from its original nonprofit purpose of benefiting humanity. Nadella counters that Microsoft took enormous risk to fund an organization others would not support, enabling products like ChatGPT and Copilot and helping create one of the world’s wealthiest nonprofits. Under cross-examination, however, he acknowledged he was unaware of any full-time nonprofit employees or tangible nonprofit output before 2026. The internal emails and text messages reveal a partnership defined by mutual dependence but also asymmetrical power: OpenAI needed Microsoft’s capital and infrastructure, while Microsoft, fearing becoming “the next IBM,” pushed hard to ensure governance aligned with its strategic needs.