Unredacted messages show Microsoft’s fingerprints on OpenAI’s board reboot
The OpenAI governance crisis in November 2023 unfolded largely in public, but the real power struggle happened in private text threads. Newly unredacted court filings from the Musk v. Altman trial show Microsoft executives actively vetting and shaping the slate for a reconstituted OpenAI board. Microsoft CTO Kevin Scott went name by name through potential directors with Satya Nadella, Sam Altman, and Brad Smith, offering enthusiastic support for some and firm resistance to others. Former Airbnb COO Belinda Johnson was labeled “great,” while former Gates Foundation CEO Sue Desmond-Hellmann was Nadella’s own proposal and ultimately joined the OpenAI board. The exchanges reveal Microsoft as more than a passive strategic partner. Even as Nadella later testified that OpenAI’s board was free to ignore his input, the conversation shows Microsoft acting as a de facto gatekeeper at a moment when OpenAI’s leadership and governance were in flux.
Who Microsoft favored — and who drew a ‘strong, strong no’
The unsealed conversation exposes how concerns over competitive alignment drove Microsoft’s stance on individual candidates. Diane Greene, former Google Cloud CEO, triggered a “strong, strong no” from Kevin Scott, a response Nadella backed on the stand by citing her ties to a direct AI competitor. William “Bing” Gordon, a longtime Amazon board member, met similar resistance. By contrast, candidates viewed as strategically neutral drew praise. Brad Smith promoted Anne Sweeney as “solid, thoughtful, calm,” and Leslie Kilgore as “incredibly smart, firm, practical.” Scott ran through an extended list of possibilities, even briefly floating himself—an idea Nadella dismissed. Altman’s emergent framework of Bret Taylor, Larry Summers, and Adam D’Angelo as a three-person board showed OpenAI trying to reclaim control, but Microsoft’s preferences had already narrowed the field in ways aligned with its own AI ambitions.
Satya Nadella’s IBM warning and Microsoft’s dependency anxiety
Behind Microsoft’s intense interest in the OpenAI board sat a deeper strategic unease. In court, Satya Nadella’s earlier internal email surfaced, comparing the OpenAI partnership to Microsoft’s historic PC relationship with IBM. He warned he did not want Microsoft to become “the next IBM” while OpenAI became “the next Microsoft.” The comment underscored how the company viewed its USD 10 billion (approx. RM46 billion) 2023 investment as a “one-way door”: Microsoft could not build twin AI supercomputers for itself and OpenAI, and would instead “outsource” critical IP development to its partner. Nadella testified that this massive dependency meant Microsoft had to secure access to OpenAI’s intellectual property while still building its own capabilities. That calculus helps explain why board composition became a strategic battleground, with Microsoft seeking directors unlikely to tilt OpenAI toward rival ecosystems or reduce its leverage.
Balancing OpenAI’s independence with Microsoft’s strategic control
The trial record highlights a central tension: safeguarding OpenAI’s independence while protecting Microsoft’s interests. Nadella insisted that board discussions were initiated by Altman and other OpenAI insiders, and that the board could have disregarded his advice. Yet the messaging thread shows Microsoft executives deeply involved in evaluating personalities, corporate histories, and potential conflicts, from skepticism about Larry Summers’ “mercurial” nature to enthusiasm for candidates with governance experience but limited AI competitive overlap. One of Nadella’s preferred candidates, Sue Desmond-Hellmann, now sits on the OpenAI Foundation board alongside Bret Taylor, Adam D’Angelo, and others, indicating that Microsoft’s input translated into durable governance outcomes. As OpenAI’s nonprofit arm now holds a minority stake in the for‑profit entity, the episode illustrates how a key investor can shape governance structures during a leadership crisis—without ever formally holding a board seat.
Power, mission, and the future of AI governance
The Musk v. Altman case frames Microsoft’s behavior as evidence of overreach. Musk’s lawsuit argues that by safeguarding its investment and steering OpenAI’s direction, Microsoft helped divert the nonprofit from its original mission to build AI for the benefit of humanity, and that his early funding was misused. On the stand, Nadella countered that Microsoft assumed enormous risk to back an unfashionable AI lab and helped create “one of the largest nonprofits in the world,” enabling tools like ChatGPT and Copilot. Under cross-examination, however, he acknowledged that he was unaware of any full-time staff or tangible research output from the OpenAI nonprofit before March 2026. That contrast—between lofty mission and investor-driven control—captures the core of the OpenAI governance crisis, and raises broader questions about how corporate board influence will shape the trajectory of frontier AI labs.