"We have each other's phone numbers," Nadella told jurors, a shorthand for how close Microsoft and OpenAI had become. He argued Microsoft had outsourced much of the core IP development, taken a costly dependency on OpenAI and accepted commercial risk to secure access to its models and computing capacity, tying commitments to milestones from early support in 2016 to revenue sharing in 2019 and later capital talks. The trial has moved into its final stretch as both sides call some of their last witnesses.
"We have each other's phone numbers," Nadella told jurors, underscoring the personal and corporate closeness he described between Microsoft and OpenAI. His testimony sought to explain why Microsoft shifted from a more traditional vendor relationship to something deeper, and why the company accepted both commercial risk and narrow strategic choices to secure access to AI models and computing capacity.
Timeline and strategic trade-offs
Nadella laid out a sequence of decisions he said were responses to technical and competitive realities. Microsoft provided early support in 2016, he said, and the court saw an August 2017 email from Elon Musk thanking Microsoft for that financial and computing help. The companies announced a public revenue-sharing arrangement in July 2019, and Nadella testified about further capital commitments discussed in 2022 and 2023. Collectively, he argued, those steps reflected a drive to ensure Microsoft and its customers could deploy advanced models on Azure rather than cede that layer to others.
He described one choice in particular as a "one-way door," saying Microsoft couldn't realistically build parallel supercomputing stacks and therefore had to divert scarce compute resources toward OpenAI. That, he told the court, was a deliberate trade-off: Microsoft sought contractual access to resulting intellectual property while continuing to develop its own AI capabilities, a hedge against becoming the next IBM in a shifting industry architecture.
The stakes reach beyond Microsoft and OpenAI. Nadella told the court the relationship affects software and cloud customers, enterprise users who rely on Azure for AI deployment, technology investors and a broader AI industry that now depends on a small set of large compute providers. Internal emails entered at trial showed Nadella probing candidates for an OpenAI board seat on competitive grounds and pushing for choices he judged less conflicted. One of his preferred candidates, Sue Desmond-Hellman, was ultimately appointed.
Nadella also described the commercial risk Microsoft accepted. He said the company absorbed roughly $15 million by letting OpenAI use Azure at a deep discount while early models were being developed.
That early operating loss was central to Microsoft’s account that it had put meaningful skin in the game to enable OpenAI’s work, including products like ChatGPT and Copilot that later reached millions of users.
Elon Musk is a central figure in the litigation. The complaint in his 2024 lawsuit alleges Microsoft aided a breach of the charitable trust that governed OpenAI’s original nonprofit and that funds Musk supplied early on, estimated in court filings at roughly $38 million to $44 million, were misused. Defendants at trial have argued Musk only began contesting the partnership after he filed that suit and after launching a rival project.
Microsoft’s defence, as Nadella presented it, rests on both the company’s early risk and the absence of contemporaneous objections. He told jurors he never received complaints from Musk even as Microsoft’s investment and licensing ties to OpenAI expanded, and he displayed the August 2017 thank-you email from Musk as evidence. The company also emphasised the commercial logic of its choices: without contractual terms and preferential access, Microsoft argued, it could have been shut out of the platform layer that now powers major AI services.
The trial record broadened after Nadella’s testimony. Witnesses with direct ties to OpenAI’s technical and governance history also took the stand, including cofounder Ilya Sutskever and board chairman Bret Taylor, who testified about the company’s public commitment to its founding mission. Those accounts fed into competing narratives at the core of the case: whether Microsoft enabled a private capture of a public-minded initiative, or whether Microsoft and OpenAI entered a commercial relationship built on mutual risk and shared technical needs.
For Microsoft, Nadella’s testimony tried to thread a narrow needle: acknowledge heavy dependence on OpenAI’s models while arguing the company had no practical alternative and had paid a real cost to secure access. For Musk and his legal team, the same timeline is evidence of a shift away from OpenAI’s original governance and purpose. The jury is left to weigh those competing interpretations as the trial nears its close.
Related Articles
- IREN stock jumps 24.7% after $9.7B Microsoft deal
- IPO rebound raises market-top risks after 1,035 peak
- Tariff threat: USTR proposes 25% duty on Brazilian imports
With both sides calling final witnesses, the jury will soon weigh competing narratives about whether the partnership was a commercial deal built on mutual risk or a departure from OpenAI’s original governance.
This article was created with AI assistance.
