Satya Nadella took the witness stand Monday morning in San Francisco federal court, and the Microsoft CEO had to answer the $13 billion question: How did his company go from calling OpenAI "worthless" to betting everything on it?

The answer, according to testimony in Elon Musk's lawsuit against OpenAI, comes down to one thing: Microsoft was terrified of losing OpenAI to Amazon.

Nadella's testimony followed damaging revelations from Week 2 of the trial, where internal Microsoft emails showed his entire executive team was deeply skeptical of OpenAI in 2017 and 2018. CTO Kevin Scott called OpenAI "a bucket of undifferentiated GPUs" and was "highly skeptical of imminent AGI breakthrough." The AI team said there was "no value in engaging." Jason Zander, Executive VP for Azure, worried about the "worst case" scenario of OpenAI "ditching Azure for AWS." Even Microsoft's research team thought their own work was "more advanced."

But by 2019, Microsoft had invested $1 billion. Then $13 billion more. Now that stake is valued at $228 billion — 17 times the initial investment.

The cosmic irony? OpenAI committed $138 billion to AWS anyway. The exact Amazon scenario Microsoft feared most came to pass despite their massive investment.

Nadella's questioning likely focused on newly disclosed emails from January 2018 showing Microsoft executives debating whether to give OpenAI a discount on Azure computing power. "From what Elon is telling everyone... he feels Open AI is at verge of some big AGI (artificial general intelligence) breakthroughs," Nadella wrote at the time.

But skepticism dominated. Scott feared OpenAI might "storm off to Amazon in a huff." When Sam Altman asked for $300 million in Azure credits, Microsoft couldn't even commit $35-50 million from the Xbox budget.

Then cash-strapped OpenAI established a for-profit subsidiary to attract investments rather than relying on donations. Eighteen months after turning its back on the startup, Microsoft finally opened its checkbook.

Musk's lawyers are using Nadella's testimony to argue that Microsoft knew it was helping divert a nonprofit from its original mission. Musk is seeking $134-150 billion in damages and wants OpenAI forced back to nonprofit status — a move that would crater its position in the AI race against Anthropic, Google, and China.

Ilya Sutskever, OpenAI's former chief scientist who briefly ousted Altman in November 2023, testifies after Nadella. Sam Altman takes the stand Tuesday or Wednesday in what's expected to be one of the final stages of the trial.

The advisory jury will reach a verdict by May 18, with Judge Yvonne Gonzalez Rogers making the final ruling on liability and remedies. If she sides with Musk, OpenAI's planned IPO could be in jeopardy.

The trial has already revealed the internal chaos at OpenAI through testimony from four former leaders — Mira Murati, Jan Leike, Helen Toner, and Tasha McCauley — who all testified against Altman. Greg Brockman testified that Musk physically threatened him in 2017 over control of the company.

But Nadella's appearance marks the first time a sitting Big Tech CEO has testified in what's become the most consequential AI trial in history. His answers about Microsoft's strategic whiplash — from skeptical to all-in — could determine whether Musk succeeds in unwinding one of the largest tech partnerships ever created.