Nadella on the Stand: Microsoft Feared Becoming 'the Next IBM' in Its OpenAI Bet

The most powerful technology executive not directly named in the lawsuit took the stand Monday morning in the Musk v. Altman trial, and the picture he painted of Microsoft’s OpenAI partnership was simultaneously triumphant and strategically terrified.

Satya Nadella, who has stewarded Microsoft’s transformation from a stagnating software company into an AI-first technology giant, spent several hours before the federal court in Oakland, California, testifying about the decisions that produced the most consequential corporate technology partnership of the decade. What emerged was a portrait of a CEO who knew he was making an enormous bet, feared exactly the wrong outcome, and made the investment anyway.

The IBM Email

The morning’s most explosive moment came courtesy of Elon Musk’s lead trial attorney, Steven Molo, who introduced an internal Microsoft email dated April 2022. In it, Nadella wrote a warning to his team as Microsoft was preparing to commit an additional $10 billion to OpenAI: he did not want Microsoft to become “the next IBM” while OpenAI became “the next Microsoft.”

The IBM parallel was deliberately chosen and surgically precise. IBM dominated the computing industry of the 1970s, then famously allowed Microsoft to retain the rights to MS-DOS when licensing it for IBM’s personal computers — a decision that handed Microsoft the platform leverage that made it the defining technology company of the following two decades. IBM built the hardware; Microsoft owned the software that ran on it. IBM became a vendor to its own strategic partner.

Nadella was describing exactly that fear: that Microsoft, by pouring capital and computing resources into OpenAI’s development, was potentially constructing the same trap for itself. OpenAI would own the intelligence layer — the model, the interfaces, the consumer and enterprise relationships — while Microsoft provided the infrastructure on which it all ran.

The email was not a rejection of the investment. Nadella made the investment. But it reveals the strategic anxiety operating beneath the surface of what Microsoft has publicly described as a seamless partnership built on shared vision.

A “One-Way Door”

Nadella reinforced that anxiety with testimony about the nature of the decision itself. He described the choice to deepen Microsoft’s commitment to OpenAI as a “one-way door” — a decision that could not be reversed without enormous cost once made.

The specific irreversibility he cited was hardware. Microsoft could not simultaneously build a supercomputer-scale infrastructure cluster for its own AI development and provision the equivalent compute that OpenAI required to train frontier models. The capital and physical infrastructure involved meant Microsoft had to choose — and in choosing to support OpenAI’s compute needs, it accepted a structural dependency on OpenAI’s success that it could not easily unwind.

“We were outsourcing essentially a lot of the core IP development and taking a massive dependency on OpenAI,” Nadella testified. He said he wanted to ensure Microsoft had access to the intellectual property generated by the partnership — which is precisely why the licensing terms that Musk’s legal team has been scrutinizing throughout the trial were so critical to Microsoft’s position.

Musk Never Called

One of the cleaner moments of Nadella’s testimony, from Microsoft’s perspective, was his account of Elon Musk’s silence. Nadella testified that despite his role as a Microsoft investor and longtime acquaintance, Musk never contacted Nadella to raise concerns that Microsoft’s investments in OpenAI were violating any commitments Musk believed OpenAI had made to remain a nonprofit.

This testimony directly undercuts a pillar of Musk’s case: that the Microsoft investment was part of a corrupting commercial transformation that violated the original charter of the organization Musk co-founded. If Musk had genuine concerns at the time, the argument goes, he would have raised them with the man writing the checks. He did not.

Musk’s attorneys have countered that Musk’s silence does not constitute approval — and that the alleged violation occurred through the restructuring process itself, not merely through the Microsoft investment. But Nadella’s testimony on this point landed cleanly enough that it is likely to feature prominently in OpenAI’s closing argument.

”Proud” of the Investment

Nadella was unequivocal about the outcome. Asked whether he viewed Microsoft’s investment in OpenAI as a success, he said he was “proud” of what the partnership had produced.

That pride is commercially legible. Microsoft’s more than $13 billion in cumulative investment — $1 billion in 2019, $2 billion in 2021, and $10 billion in 2023 — has been transformed by the AI boom into a strategic asset that analysts credit with much of Microsoft’s extraordinary market valuation over the past three years. Azure has become the cloud infrastructure backbone for OpenAI’s operations. Microsoft 365 Copilot, built on OpenAI models, is now a major enterprise product. The IBM scenario that Nadella feared has not materialized — at least not yet.

Whether that remains true depends significantly on how the partnership evolves under the revised terms the two companies agreed in April 2026. Under the amended agreement, Microsoft’s license to OpenAI technology is no longer exclusive, and OpenAI is now free to serve customers on other cloud providers. Nadella’s IBM fear may have receded; it has not fully disappeared.

The Trial’s Final Act

The Musk v. Altman trial is now in its third week, with an advisory jury verdict expected by the week of May 18. Sam Altman himself is expected to take the stand in the coming days — the testimony that will define the trial’s closing narrative.

The stakes are not merely reputational. Musk’s lawsuit seeks roughly $134 billion in damages and the structural unwinding of what he characterizes as the illegal commercial conversion of a nonprofit into a profit-seeking entity. OpenAI’s defense has focused on the factual record: that the nonprofit’s board authorized the conversion, that donors were aware of the possibility of commercialization, and that the mission — developing AI for humanity’s benefit — has not changed even as the commercial structure around it has.

Nadella’s testimony added texture to that defense without resolving its central tension. Microsoft invested in OpenAI because it believed the mission was real and the models would be transformative. It also invested because it feared what would happen to Microsoft if it did not. Those two motivations coexisted in April 2022, as they do now.

Whether a jury finds that coexistence compatible with OpenAI’s nonprofit obligations is the question the next two weeks will answer.