A Historic Parallel: Nadella’s IBM Warning
Satya Nadella’s testimony in the Musk v. Altman trial revealed a striking internal warning: he did not want Microsoft to become “IBM” while OpenAI became “the next Microsoft.” The comparison refers to IBM’s early dominance in personal computing, and how its reliance on partners helped others seize the strategic high ground. Nadella said this fear surfaced as Microsoft prepared a major expansion of its OpenAI investment in April 2022, describing the decision as a “one-way door.” The company could not realistically build two separate AI supercomputers—one for itself and one for OpenAI—without sacrificing scarce compute resources. That trade-off meant accepting a deep Microsoft OpenAI dependency, even as the company worried about repeating IBM’s mistake of enabling a partner to outgrow and outmaneuver it. The Microsoft IBM comparison now sits at the heart of the company’s AI strategy risks.
Outsourcing Core AI While Fighting Dependency
On the stand, Nadella described Microsoft’s OpenAI deal as simultaneously a breakthrough and a structural risk. By his account, Microsoft “outsourced essentially a lot of the core IP development,” taking a massive dependency on OpenAI’s research and models. The logic was pragmatic: OpenAI needed infrastructure and capital, while Microsoft needed a fast path to cutting-edge AI without duplicating efforts. Yet Nadella emphasized that Microsoft insisted on access to the intellectual property generated by the partnership and on continuing to build its own capabilities. The tension is obvious. Microsoft must align tightly with OpenAI to reap the benefits of models like those behind ChatGPT and Copilot, while avoiding a future where it is merely the infrastructure provider for a more powerful AI company. This push-and-pull defines its current AI strategy risks and underpins Nadella’s IBM analogy.
Boardroom Maneuvers During OpenAI’s Leadership Crisis
Newly unredacted messages, entered into evidence alongside Nadella’s testimony, show how deeply Microsoft engaged in OpenAI’s governance during the November 2023 crisis that briefly ousted Sam Altman. In text exchanges with Altman and senior Microsoft executives, CTO Kevin Scott floated a slate of possible OpenAI board members. Nadella pushed back on candidates such as former Google Cloud CEO Diane Greene and longtime Amazon-linked executive William “Bing” Gordon, objecting to their ties with companies competing directly with Microsoft in AI. At the same time, Nadella proposed figures like former Gates Foundation CEO Sue Desmond-Hellmann, who ultimately joined the board. The thread shows Microsoft trying to shape a board friendly to its strategic interests while insisting it was only responding to requests for input. These board considerations highlight how governance became another lever to manage Microsoft OpenAI dependency without formal control.
Balancing Partnership Gains with Strategic Vulnerability
Nadella portrayed Microsoft’s alliance with OpenAI as a high-risk, high-reward bet that enabled a “fledgling AI lab” to scale into one of the largest nonprofits in the world, powering products such as ChatGPT and Copilot. He argued that no other funder was ready to back OpenAI at that level and that Microsoft’s infrastructure and capital were decisive. Under cross-examination, however, he acknowledged that he knew of no full-time employees at the OpenAI nonprofit before March 2026, nor of grants or open-sourced technology from that entity, a point Musk’s legal team used to question whether the original nonprofit mission is still central. The result is a paradox: Microsoft has unlocked enormous commercial value while deepening reliance on a partner whose governance and mission are under scrutiny. Its future AI competitiveness may hinge on whether it can sustain the alliance without repeating IBM’s loss of strategic control.