I’ll be honest. I wasn’t expecting Sam Altman’s testimony to actually deliver. Most high-profile tech trials end up being a few hours of lawyers reading emails out loud while the witness says “I don’t recall.” This wasn’t that.
Tuesday and Wednesday at the federal courthouse in Oakland, Altman spent close to ten hours on the stand answering questions about how OpenAI became OpenAI. Elon Musk’s lawyer Steve Molo was on him from the jump. The very first question out of the gate was sharp. “Are you completely trustworthy?” Altman said yes. He called himself “an honest and trustworthy businessperson.” The courtroom didn’t laugh, but a few reporters in the back row exchanged looks.
Look. This case matters. It’s not just two billionaires fighting over an old email chain. The judge here is being asked to decide whether the company that built ChatGPT, the company that triggered the whole 2023 AI boom, is allowed to keep operating in its current shape. Or whether a court can force it back to its 2015 nonprofit roots.
What Came Out of Sam Altman Testimony
Here’s the bombshell that’s already running on every tech podcast this week.
Altman testified that back in the early years, before he was running the company, Musk wanted to personally own 90 percent of OpenAI. When the board said no, Musk floated a Plan B. Just fold the whole project into Tesla. Make it a Tesla subsidiary. The board said no to that too.
Musk walked away from OpenAI in 2018.
That detail reframes the entire lawsuit. Musk’s public version of events is that he was a co-founder who got betrayed when his nonprofit got hijacked into a profit machine. Altman’s version, backed by emails the prosecution has already submitted, is that Musk wanted control of the thing and lost a board fight. Two very different stories. Same set of facts.
The jury watched all of this. Hard to know what they’re thinking. But one beat reporter from NPR noted that two jurors took notes during the 90 percent revelation. Nobody took notes during the part where Musk’s lawyer was reading old tweets.
The “Stole a Charity” Line
Musk’s lawyers keep using that phrase. “Stole a charity.” It’s catchy. It’s been the headline of half the coverage. And when Molo dropped it on Altman directly, the OpenAI CEO didn’t dodge.
“It feels difficult to even wrap my head around that framing,” he said.
His defense, the same one his lawyers have been making since opening arguments, is straightforward. Building safe and powerful AI costs tens of billions of dollars. Donations don’t cover that. Grants don’t cover that. The only way to raise the money OpenAI needed was to create a capped-profit structure on top of the nonprofit. The nonprofit still controls the for-profit. The mission is still the same. The legal wrapper just changed.
Whether the jury buys this is the whole game.
Why Musk Is Actually Doing This
Okay let’s be real about this part for a second.
Musk runs xAI now. XAI is in a death match with OpenAI for compute, for talent, for chip allocations from Nvidia. Tesla also has its own AI compute business (we covered Musk’s Terafab AI chip play last month, and that context matters here). Musk is not exactly a neutral party in this lawsuit. He’s a competitor who’s suing the market leader.
That doesn’t mean he’s wrong. He might be right on the merits. The original 2015 OpenAI documents do read like a nonprofit charter. Reasonable people can argue the pivot crossed a line.
But the outcome Musk is asking for is telling. He wants Altman removed from leadership. He wants the for-profit structure unwound. If a federal judge granted both of those, OpenAI overnight loses its ability to outbid xAI for anything. The race resets. Musk wins by default.
The jury isn’t blind to this. Altman’s lawyer made sure of that on redirect.
What’s at Stake Beyond OpenAI
This part doesn’t get covered enough so let me spell it out.
If OpenAI loses, every American AI startup that took early nonprofit-style funding and later pivoted to commercial has to ask whether they’re next on the docket. Anthropic has a similar structure. So does the Allen Institute. So do half the university AI labs that spun out into companies. The whole 2026 AI ecosystem (already rattled by the Vercel security breach earlier this year) is watching this verdict closely.
If OpenAI wins, the playbook of “start as a research org, pivot to revenue once you have something to sell” becomes the official template for the next decade of AI companies. Investors are watching this trial with the same intensity as the developers.
And then there’s Microsoft.
Redmond has poured roughly $13 billion into OpenAI. That money is the foundation of Copilot, of Azure AI, of basically Microsoft’s entire 2026 product roadmap. CEO Satya Nadella testified earlier in the trial that Musk never once raised his concerns to him directly. That’s a quiet but devastating piece of evidence. If you really believed your former co-founder was stealing a charity, you’d call the guy writing the checks. Musk didn’t.
The Moment That Stuck With Me
About six hours into Wednesday’s session, Molo asked Altman whether he ever felt OpenAI was being run for the benefit of the public, the way the founding documents promised.
Altman paused. Longer than he’d paused for any other question.
Then he said: “I think we’re trying. I think anyone running a company this consequential is going to look back and find things they wish they’d done differently. But yes. I believe the mission is alive.”
It was the most human moment of the whole testimony. No corporate polish. No talking point. Just a guy admitting he’s not sure he’s gotten everything right but he’s trying anyway.
The court reporter caught it word for word. Whether the jury catches the same vibe, we’ll see.
Why This Matters
For American tech workers, investors, and anyone who uses ChatGPT, Claude, or Gemini on a daily basis, the verdict in this trial is one of the most consequential regulatory-shape events of the 2026 AI economy. We’re not talking about a fine or a slap on the wrist. We’re talking about whether a federal court can force one of the most valuable AI companies in the world to restructure on the spot.
The downstream effects show up fast. Model release timelines slow if OpenAI has to defend appeals. Enterprise contracts get re-priced if the corporate structure changes. Compute deals with Nvidia and Oracle get renegotiated. Every one of those moves hits the AI products you and I use within months.
What Happens Next
Closing arguments are expected next week. Most of the legal observers I follow expect a verdict before the end of May. Both sides have already signaled they’ll appeal whatever comes down, which means the actual final answer is probably 18 months and a Ninth Circuit ruling away.
In the meantime, OpenAI keeps shipping. There’s a rumored GPT-5.5 release in July. Altman, when he’s not on the witness stand, is reportedly flying to Beijing this week as part of the Trump business delegation visiting Xi Jinping. That detail tells you something about where Altman thinks the next regulatory headache is going to land. Hint: not Oakland.
USABlaze Takeaway
This isn’t really about Musk versus Altman. Not at its core. It’s about whether American AI gets built inside a structure that can compete globally, or whether a courtroom decides the 2015 promise still binds in 2026. We’ll be on closing arguments next week and breaking down the verdict the day it lands. The follow-on questions, what this means for Anthropic, for Meta’s open-source bet, for the AI Safety Institute, those are the ones we’ll dig into over the next month.
If you want our take when the verdict drops, this is the page to bookmark.
Sources: CNBC, NPR, Axios, Fortune, Al Jazeera.
By The USABlaze Editorial Desk