Friday afternoon. Three days before the trial of the year. Elon Musk's lawyers filed paperwork to drop 24 of his 26 claims against OpenAI and Sam Altman. The fraud charges. The constructive fraud charges. Every spicy theory. Gone.

Jury selection starts Monday in Oakland. There's still $150 billion on the line.

The case Judge Yvonne Gonzalez Rogers will hear has been narrowed to two arguments: breach of charitable trust and unjust enrichment. Both rest on the same theory — that when OpenAI converted from a nonprofit to a for-profit in 2019, it took assets "irrevocably dedicated to charitable purposes" and turned them into a private company now valued at over $500 billion.

Musk filed the original complaint in November 2024 with 26 separate claims. By April 25, 2026, two remained. That's a 92% reduction in legal theory, 72 hours before opening arguments.

Musk's team called the move "streamlining." Altman's lawyers almost certainly called it something else.

Here's what's still at stake. Musk wants up to $150 billion in damages, paid to OpenAI's charitable arm rather than to him personally. He wants Sam Altman and Greg Brockman removed from leadership. He wants OpenAI's nonprofit status restored. That last one would functionally unwind the corporate structure OpenAI uses to raise capital from Microsoft, SoftBank, Amazon, and now Google's $40 billion compute commitment from last week.

The trial unfolds in two phases. First, a jury hears arguments and returns an "advisory verdict" — non-binding, just for the judge's reference. Then Gonzalez Rogers rules on the actual remedies herself. So the jury is essentially a $150 billion focus group with subpoena power. Useful, theatrical, but not decisive. The judge has the final word, and that word will arrive months from now.

Witness list watchers should know that Greg Brockman's personal journal entries from 2017 to 2019 are already in the record. So are emails between Musk and Altman from the same window. The discovery phase produced 700,000 pages of documents. Most of those pages have not been seen by anyone outside the case. They will be next week.

My Opinion

I think the fraud claims got dropped because they were the weakest part of the case, not the strongest. Fraud requires proving someone lied at a specific moment with specific intent to deceive. That's a steep hill in any courtroom. Charitable trust violation is much simpler — just show that money raised for one purpose got used for another. OpenAI's 2019 structural pirouette makes that the easier story to tell twelve jurors in Oakland.

What bugs me is the framing. Musk says he dropped the fraud claims to "focus jurors on whether OpenAI serves humanity." Sure. He's also asking for $150 billion from a company he tried to take over and lost in 2018. The "OpenAI must serve humanity" pitch is the moral high ground he's using to extract a financial verdict he can call selfless because the money technically goes to OpenAI's charity. It's clever positioning. It's still litigation, and the litigant is the second-richest man on Earth.

The real story isn't whether Musk wins this round. It's what happens to every AI company built on a converted nonprofit if a federal judge rules that OpenAI's 2019 restructuring was a breach of charitable trust. That ruling would put a permanent asterisk next to every dollar OpenAI has raised since 2019, and put the whole "convert your nonprofit research lab into a $500 billion hyperscaler" playbook on a watch list. Anthropic was never a nonprofit. xAI was never a nonprofit. Only OpenAI did the dance. On Monday, twelve people in Oakland get to weigh whether that dance was legal.

Author: Yahor Kamarou (Mark) / www.humai.blog / 26 Apr 2026