The Musk v. Altman trial unfolding in federal court has peeled back layers of Technology history that most people assumed was settled. Turns out, it wasn’t. What we’re learning through emails, text messages, and testimony isn’t just a story about a lawsuit. It’s a story about what happens when powerful people in AI disagree about ownership and control.
The core claim is straightforward: Elon Musk says Sam Altman and Greg Brockman stole a nonprofit organization, using $38 million of Musk’s money to build something worth over $800 billion today. That’s a serious accusation. But the evidence being presented suggests something more complicated and arguably more revealing about how OpenAI actually came to be.
The Tesla Recruitment Gambit
Before Musk stepped down from OpenAI’s board in February 2018, he was trying to recruit Altman to run an AI lab inside Tesla. Not as a side project. As the main thing. The emails presented in court show Musk even dangled a Tesla board seat as incentive.
This wasn’t vague interest either. There were detailed plans. A draft FAQ for a Tesla AI event at the NeurIPS conference described the lab’s purpose: “share that Tesla is building a world leading AI lab which will rival the likes of Google / DeepMind and Facebook AI Research.” Someone had already listed potential leaders. Andrej Karpathy’s name appeared. So did Sam Altman’s, with two question marks next to it.
The question marks are doing a lot of work in that document. They suggest uncertainty. Maybe it meant Altman hadn’t committed yet. Maybe it meant Musk wasn’t sure if Altman would bite. Either way, the implication was clear: absorb OpenAI’s talent and brainpower into Tesla, where Musk could control it directly.
Shivon Zilis, a former OpenAI adviser and board member who testified on Wednesday, kept the communication flowing between Musk and Altman during this period. In one text, she asked Altman directly: “Did you think through a B Corp subsidiary of Tesla?” In an email to Musk months earlier, she wrote that OpenAI’s cofounders hadn’t “internalized the advantages of burying this in Tesla for stealth advantage.” When OpenAI’s lawyers questioned what “burying” meant, Zilis clarified it didn’t mean going closed-source. It meant, she said, being “a small fish in a big pond.”
That phrase captures something important. Being absorbed into Tesla wouldn’t have killed OpenAI as a research organization. It would have just subordinated it to Tesla’s interests and Musk’s vision. There’s a difference between being acquired and being buried.
The Conflict of Interest Nobody Mentioned
Here’s where things get genuinely uncomfortable. In January 2020, Zilis was appointed to OpenAI’s board. This was years after her recruitment efforts for Tesla had fizzled. But here’s the part that matters for understanding potential conflicts of interest: during her time on the board, Zilis became pregnant with four of Musk’s children through IVF. She didn’t tell the other board members, citing a confidentiality agreement she’d signed with Musk.
That’s a staggering omission from a governance perspective. She’s representing Musk’s interests on one board while bearing his children and having signed confidentiality agreements with him. When Business Insider reached out in 2022 planning to report on the children, Zilis called Altman first. Not her lawyer. Altman.
She remained on OpenAI’s board until February 2023, months before Musk launched xAI, his own competing AI lab. The timing is what makes this story interesting. Zilis resigned from OpenAI shortly after what she says was a call from Altman informing her about Musk’s effort. Yet text messages suggest she already knew about xAI before that call even happened.
“Have to resign from the openai board btw. [Musk’s] effort has become well known,” she told a friend. “When the father of your babies starts a competitive effort and will recruit out of openai there is nothing to be done.”
What OpenAI’s Defense Actually Reveals
OpenAI’s legal team has countered that Musk’s real motivation was wounded ego. They argue he’s had “sour grapes” since failing to take control of OpenAI in 2017, and that he’s since launched xAI as a rival operation. Their strategy during the trial has been to paint Musk as someone who wanted to corrupt and absorb OpenAI into Tesla, not a visionary investor betrayed by nonprofit founders.
There’s something to this. Musk’s pattern of attempting to recruit Altman and other key researchers does look like an effort to consolidate AI research under Tesla’s umbrella. And the fact that he eventually started xAI after those efforts failed suggests he was serious about being in control of AI development, one way or another.
But here’s where the defense gets weird. OpenAI’s lawyer, William Savitt, described Musk’s recruitment attempts as part of an effort to “corrupt OpenAI and absorb it into Tesla.” That’s one way to frame it. Another way is to say Musk believed his money and vision deserved a seat at the decision-making table, and when he didn’t get it, he left to build his own thing.
Neither framing is wrong, exactly. Both are probably partially true.
The Deeper Question Nobody’s Asking Clearly
The trial has surfaced something that deserves more scrutiny than it’s probably getting: nonprofit governance in AI is broken, or at least radically underspecified. OpenAI was founded as a nonprofit, received tens of millions in investment, and then somehow became a for-profit Business worth $800 billion. That transition happened, and now we’re in court arguing about who owns what.
Musk’s claim that he invested $38 million in a nonprofit that became a megavalued private company is technically correct. What’s less clear is whether he’s owed something for that, or whether he’s just mad he doesn’t control where that value ended up. Those are different things, legally and morally.
Zilis’ presence on the board while pregnant with Musk’s children and bound by secrecy agreements is a governance failure regardless of who wins this lawsuit. How does that happen? How does a board not know? And if they did know or suspect, why didn’t they address it?
The trial continues with more witnesses expected, including a former Columbia Law School dean who specializes in nonprofit law. That choice feels deliberate. OpenAI’s structure, its transition from nonprofit to for-profit, and how it was governed are apparently going to be central to determining what, if anything, Musk is owed.
The emails and text messages paint a picture of people who wanted very different things from AI research in the late 2010s. Musk wanted control and integration with his other companies. Altman wanted independence and growth. Those desires collided, and the organization’s nonprofit structure became a battleground for how that collision would be resolved.
What nobody predicted, apparently, is that the collision would end up in federal court with a jury staring at text messages about whether someone is the “father of your babies” and whether that changes fiduciary obligations.
