The Openness Paradox: Why the Man Who Built Closed Systems Is Suing for Open AI

When Elon Musk and Sam Altman face a jury in Oakland on Monday, the official case is about whether OpenAI betrayed its founding mission as a nonprofit. The real story is simpler and stranger: the man who built the most successful closed-system companies in modern technology is suing to define what open AI means, at the precise moment he needs the courts to help him compete. [The Guardian]

Musk has spent years telling anyone who will listen that OpenAI abandoned its founding promise. OpenAI says Musk approved the for-profit pivot while he was still on the board. The emails will come out. The diary entries will come out. The private texts will come out. But the backdrop to all of this is a $1.25 trillion IPO on one side and an $850 billion to $1 trillion IPO on the other, both scheduled for this year. [CNBC]

The openness contradiction runs through everything Musk does in AI. Tesla runs on proprietary software. SpaceX holds its rocket technology close. When Musk sued OpenAI, he positioned himself as a champion of open, democratic AI development. His own AI company xAI released Grok, a model that is not open source. The weights are not freely available. The license restricts commercial use. This is not an open AI company. It is a closed one, built on the same vertical integration that made Tesla and SpaceX formidable.

This is the puzzle at the center of the case. Musk argues that OpenAI betrayed the world by pivoting to profit. He is simultaneously building a profit-driven AI company that does not share the characteristics he claims to care about. One interpretation is cynicism: open is a weapon to hurt a competitor. Another is that the principle matters but the practice does not. Neither interpretation flatters him.

The legal case has narrowed significantly. On April 25, Musk dropped his fraud claims to streamline the proceedings to two remaining theories: unjust enrichment and breach of charitable trust. [Fortune] Professor Sam Brunson of Loyola University Chicago told The Verge that the case only reached trial because Musk can pay his attorneys to argue a losing one. If this were contingency, he said, he would assume he would not be paid. [The Verge]

The damages figure of $134 billion is not backed by any clear legal theory. [Fortune] The remedies Musk seeks are unusual. He wants Altman and Brockman removed from OpenAI. He wants the for-profit restructuring unwound. And he wants any monetary award directed to OpenAI itself, not to him personally. The case is as much about strangling a competitor as it is about principle.

What makes the trial significant for the AI industry is not the personal feud but what a jury verdict could establish. A finding that OpenAI breached its charitable trust would create legal precedent for forcing AI company restructurings. Courts would have a new tool to evaluate whether companies named their nonprofit status accurately. For an industry that has used nonprofit framing as a liability shield while building billion-dollar commercial enterprises, that is a genuinely important question.

The evidence that will emerge in court is real. The emails between Musk and Altman from 2015 show a genuine shared concern about Google and a belief that someone needed to build AI outside the corporate world. Altman wrote to Musk that if AI was going to develop anyway, it should be done by someone other than Google first. That is the founding spirit. The question the trial will answer is whether it was ever more than a pitch. [The Guardian]

Jury selection begins Monday in federal court in Oakland, with Judge Yvonne Gonzalez Rogers overseeing a trial expected to last two to three weeks. [The Guardian] The jury will issue an advisory verdict. Gonzalez Rogers will then decide whether Musk proved his claims and what remedies, if any, follow. May 18 is the scheduled remedies phase. [Fortune] Both sides will be calling witnesses. Satya Nadella and Kevin Scott from Microsoft are expected to testify. Former OpenAI executives including Ilya Sutskever and Mira Murati may be called. [The Verge]

The competing IPO timelines are the structural pressure that makes this case matter beyond the courtroom drama. OpenAI is targeting a fourth-quarter public offering. xAI and SpaceX are targeting June. [CNBC] Whatever happens in Oakland this spring will affect both deals. That timing is not an accident. It is the point.