🔥 What Happened

Satya Nadella walked into a federal courthouse in Oakland, California this morning to testify in the trial that's become Silicon Valley's main event. The Microsoft CEO is on the witness stand to explain a trove of internal emails showing his company had serious doubts about backing OpenAI — and only pulled the trigger because they were terrified the startup would run off to Amazon.

Elon Musk's legal team is trying to prove that Microsoft knowingly bankrolled OpenAI's transformation from a nonprofit research lab into a for-profit juggernaut worth an estimated $850 billion. Musk claims the whole thing was a bait-and-switch that violated his founding vision of AI for humanity.

The stakes? Musk is seeking $150 billion in damages from OpenAI and Microsoft.

🧠 Why This Matters

This isn't just billionaire drama (though there's plenty of that). The trial could reshape the entire AI industry structure:

If Musk wins, the court could unwind OpenAI's for-profit restructuring — effectively halting the company's IPO trajectory toward a nearly $1 trillion valuation. If he loses, it sets a precedent that nonprofit-to-for-profit pivots in AI are fair game, greenlighting similar moves across the industry.

The outcome also impacts Microsoft directly. The company has invested tens of billions into OpenAI and restructured its entire business around AI. A ruling against them would force a painful re-evaluation.

Meanwhile, Musk's own AI company xAI — now folded into SpaceX — is targeting a $1.75 trillion valuation for its IPO as early as June.

📊 Deep Dive

The emails entered as evidence paint a picture that's far from the standard "blindingly confident tech giant" narrative. Here's what really happened:

In 2016, Microsoft gave OpenAI $60 million worth of Azure credits at a steep discount. OpenAI burned through it twice as fast as expected. When they came back asking for more — about $300 million in compute — Microsoft's top brass balked.

Nadella himself wrote in a 2017 email: "Overall I can't tell what research they are doing and how if shared with us it could help us get ahead."

Microsoft CTO Kevin Scott was even more blunt: "I visited OpenAI about a year ago, and was not able to see any immediate breakthrough in AGI." He also wrote that his fear was of them "storming off to Amazon in a huff and shit-talking us and Azure on the way out."

Jason Zander, then Azure chief, said: "My worst case scenario is having them ditch Azure for AWS… and then land with some big new innovation that is shared with our competition."

Nadella forwarded the chain to 15 executives asking for their take. The majority was skeptical. An internal cost analysis showed Microsoft stood to lose $150 million over several years on the deal.

But here's the kicker: In 2019 — just 18 months after those doubts — Microsoft announced a $1 billion investment. And after ChatGPT launched in November 2022, that number ballooned into tens of billions.

The irony is thick. In April 2026, OpenAI announced a massive deal with Amazon worth $138 billion in cloud spending, plus Amazon agreed to invest $15–50 billion in OpenAI. The exact scenario Microsoft feared in 2018 is now playing out.

⚠️ The Catch

This trial has produced one jaw-dropping revelation after another, and both sides are leaving the courtroom bruised.

Greg Brockman, OpenAI's president and a key witness last week, testified that it was actually Musk who pushed for OpenAI to go for-profit in 2017, wanting "absolute control" and a majority equity stake. When the co-founders refused to give him unilateral control, Musk allegedly stormed out of a meeting with a painting of a Tesla.

Brockman's own journal was read in court — he wrote in 2017 about wanting to become a billionaire. Today, his stake in OpenAI is worth close to $30 billion.

And in a twist out of a movie: Shivon Zilis, a former OpenAI board member who also happens to be the mother of four of Musk's children, testified that Musk tried to poach Sam Altman to lead an AI team at Tesla.

🎯 What Happens Next

Nadella's testimony today is the trial's final act before closing arguments. Sam Altman is expected to take the stand Tuesday or Wednesday, and Ilya Sutskever — OpenAI's former chief scientist — is also scheduled to testify.

The jury will then decide:

Whether Musk's claims that OpenAI "betrayed" its nonprofit mission hold up

Whether Microsoft was a knowing accomplice in that shift

Whether to unwind OpenAI's for-profit restructuring

Any damages — Musk wants $150 billion and OpenAI's CEO removed

The trial has been ongoing for three weeks. A verdict is expected by late May.

🧩 Bigger Picture

This case highlights a fundamental tension that nobody in AI has figured out: how do you build super-intelligent AI for "all of humanity" when it costs billions of dollars in compute?

OpenAI started as a nonprofit precisely because its founders believed profit motives would corrupt AGI development. But the math didn't work — training frontier models requires datacenter-level resources. So they pivoted to for-profit, took Microsoft's money, and now we're here.

Musk's argument that this was a betrayal ignores the inconvenient truth that he himself wanted to turn it into a for-profit company when he was in charge. Both sides look messy.

What's really on trial here isn't just OpenAI — it's the entire idea that you can build AGI for the public good inside a capitalist system. The answer so far seems to be: you can't, unless you find a trillion-dollar corporation willing to foot the bill.

The jury in Oakland will answer the legal question. But the philosophical one is going to haunt the industry for years.