Strip away the damages figure, the nonprofit-to-for-profit grievance, the Microsoft entanglement, and what's left in the OpenAI courtroom is something weirder and more revealing. Elon Musk, according to Sam Altman's sworn testimony, wanted to will artificial general intelligence to his children. Not sell it. Not steward it. Bequeath it.
That's the through-line the press coverage keeps treating as a colorful aside. It isn't. It's the entire story.
Altman testified on May 12, 2026, in the civil trial brought by Musk over the company the two co-founded in 2015. He described a "particularly hair-raising moment" from nearly a decade ago, when his cofounders pressed Musk on what would happen to OpenAI if Musk had control and then died. Musk's answer, as Altman recalled it: "I haven't thought about it a ton, but, you know, maybe it should just, the control should pass to my children". Altman told the jury he "didn't feel comfortable with that".
The dynastic frame nobody is using
Read the testimony carefully. Altman says Musk believed he "only trusted himself to make non-obvious decisions," and had decided to spend the rest of his career working only on companies he could control. That's the standard Musk operating principle. Tesla, SpaceX, X — same playbook.
But OpenAI was supposed to be the exception. Altman testified that part of the reason they started the company was that they did not think AGI should be under the control of any one person. Musk's children clause breaks that premise in a way the broader Musk-vs.-Altman narrative has mostly glossed over. According to Altman's account, it's not just a demand for control — it's a claim that his bloodline should inherit it.
Treat that seriously for a second. The most powerful technology of the century, allegedly proposed as a hereditary asset. Not a board seat. Not a foundation. An estate plan.
Why this lands differently than the lawsuit's official theory
Musk's 2024 complaint frames the case as betrayal — Altman and Greg Brockman accused of "looting" the charity through the Microsoft deal, abandoning the nonprofit mission for profit. He's seeking roughly $150 billion in damages from OpenAI and Microsoft. OpenAI's countertake, posted on X before jury selection, called the case a "baseless and jealous bid to derail a competitor".
Both framings are about competition. About money. About mission drift. Familiar tech-industry stuff.
The inheritance detail is something else. It suggests Musk's quarrel with OpenAI isn't really that it became a company. It's that it became a company he can't pass down. That reframes everything — the xAI launch, the Grok evangelism, the lawsuit itself — as the work of a founder who lost a thing he believed was his to keep, and to gift.
Musk's attorney, Steven Molo, opened his cross-examination of Altman by asking, "Are you completely trustworthy?". It's a great courtroom line. It's also the wrong question. The interesting question is what Musk thought he was building, and to whom he thought he was leaving it.
The succession problem AI hasn't admitted it has
Every frontier lab is run by a small number of people who behave like founders of nation-states. Altman at OpenAI. Dario Amodei at Anthropic. Demis Hassabis at DeepMind. Musk at xAI. They give interviews about timelines to superintelligence and then go home.
Nobody talks about what happens when they die. Nobody talks about whether their equity, their voting shares, their hand-picked boards default to spouses, trusts, children. The Musk testimony is uncomfortable because it makes the implicit explicit. Of course the question of inheritance applies. These are private companies held by mortal people.
Altman's discomfort, by his own account, was instinctive. Good. But the structural problem he flagged nearly a decade ago hasn't been solved by anyone. OpenAI itself is now a capped-profit entity with a complicated governance stack that has wobbled visibly — recall the November 2023 board episode. Anthropic has its Long-Term Benefit Trust. These are gestures at the question, not answers to it.
The Oakland jury is going to decide a contract and fiduciary case. Fine. The more useful trial, the one nobody is running, is the one where someone asks every AI CEO under oath: who inherits your shares?
Musk, to his credit, answered that question a decade ago. The answer was just the wrong one.
