Before the bell rings

Part 3

OPENAI’S BEST HOPE to get rid of Elon Musk’s lawsuit was to convince Judge Yvonne Gonzalez Rogers that trial was not necessary.

In federal litigation, there are three distinct phases of a lawsuit. In the first phase, the court determines if the plaintiff’s complaint lays out a plausible collection of facts, which, if proven, would entitle the plaintiff to relief under applicable law. If the plaintiff can’t do that, the case is dismissed.

If the plaintiff makes it over that obstacle, the case moves to the next phase — where the parties exchange documents and take depositions of each other. Discovery is expensive and can drag on for years. At the close of discovery, the parties are entitled to ask the court to issue “summary judgment” in their favor.

Summary judgment is the legal term for a court order that disposes of a lawsuit after the parties have completed discovery but without holding a trial. The idea is that if no key facts are in dispute after discovery is complete, the court can rule whether the law allows or forecloses relief on the basis of those undisputed facts.

If a party survives summary judgment, it means that the case must go to the third phase where the court holds a trial to determine which of the disputed facts are true. In other words, the “trier of fact” — either a jury or the judge, depending on the plaintiff’s choice at the start of the case — decides who is telling the truth and who isn’t.

In the world of high dollar federal litigation, every plaintiff wants to get to the trial phase. The discovery phase is expensive and sucks up months and years of lawyer time. Trial will be expensive too, but trial is where a plaintiff can score big, particularly with a jury. As all veteran trial lawyers know, once a case goes to the jury — a jury made up of live people with emotions and empathy and their own sense of justice — anything can happen.

Judgment will not be summary

On June 5, 2025, OpenAI and Microsoft filed motions for summary judgment against Musk. Their lawyers laid out erudite legal arguments even as they continued to poison the waters by focusing on all of Musk’s many conflicts, eccentricities and vulnerabilities. They appeared to assume that if they could convince Gonzalez Rogers that Musk had an ulterior motive (e.g., to hamstring Open AI to advantage x.AI), she would get rid of the case.

They miscalculated.

Gonzalez Rogers found several things in her review of the record that made her think that Musk’s claim had factual support. She noted that in 2017 Musk, Sam Altman, another co-founder Greg Brockman and others discussed ways to obtain capital for the development of AI beyond what they could get from donations. Altman and Brockman wanted to create a for-profit affiliate that could raise the money by offering equity. Musk proposed somehow attaching it to Tesla as a “cash cow,” though Altman was not keen on that solution.

Gonzalez Rogers highlighted that Musk insisted OpenAI remain philanthropic and on Sept. 20, 2017, sent an email that said, “Guys, I’ve had enough. This is the final straw. Either go do something on your own or continue with OpenAI as a nonprofit. I will no longer fund OpenAI until you have made a firm commitment to stay or I’m just being a fool who is essentially providing free funding for you to create a startup. Discussions are over.”

Altman responded, “I remain enthusiastic about the nonprofit structure!”

Brockman said he would “like to continue with the nonprofit structure.”

Brockman’s diary-ish personal files were the subject of discovery. One revealed that in the same month as Musk’s ultimatum he mused, “This is the only chance we have to get out from Elon. Is he the ‘glorious leader’ that I would pick? We truly have a chance to make this happen. Financially, what will take me to $1 billion? … Accepting Elon’s terms nukes two things: our ability to choose (though maybe we could overrule him) and the economics.”

Sometime after the commitment was made to Musk, Brockman wrote, “Another realization from [this meeting] is that it’d be wrong to steal the nonprofit from him to convert to a b-corp (a for-profit business) without him. That’d be pretty morally bankrupt, and he’s really not an idiot.”

The judge also referred to another Brockman writing that appeared to acknowledge that if OpenAI told Musk they were on board with the nonprofit structure but three months later created a for-profit corporation “then it was a lie.”

On Jan. 15, 2026, Gonzalez Rogers largely denied OpenAI’s motion, clearing the way for the case to proceed to trial. She found that there were disputed facts on many of the key issues including whether there was a charitable trust, whether defendants had breached their fiduciary duties, and whether Musk reasonably relied on misstatements by Altman and Brockman.

Microsoft won a part of its summary judgment motion, but the claim that it had aided and abetted OpenAI’s breach of fiduciary duty survived.

It wasn’t a pretty path, but despite the setbacks since he first filed his case in San Francisco state court in February 2024, Musk appeared to have gotten clearance to tell a jury that Altman had betrayed the charitable mission of OpenAI and, together with Microsoft, defiled his noble quest to insure that super machine intelligence would benefit all of humanity.

And that wasn’t all.

Big testimony

At a further hearing in March, the judge made two rulings that had important implications for what the jury would hear when testimony got under way.

First, she decided to allow most of the testimony expected from Musk’s witness Dr. Stuart Russell, a professor at the University of California, Berkeley. Russell has a reputation as one of the leading authorities on the development of artificial intelligence and the pursuit of AGI. Russell has a deep concern about the safety of AGI.

According to his expert report, Russell will tell the jury that AI is a powerful and transformational technology, but without safeguards and a better understanding of how it works, the technology presents a substantial risk of misuse. He will also testify that AI companies (and their governments) “have very strong incentives to pursue artificial general intelligence, despite the safety risks.”

The judge said Russell satisfied the requirements to testify as an expert on those subjects and she would let the jury hear what he had to say. (She would not let him testify that those risks could be catastrophic for humankind, because she felt that Russell had not quantified that risk and it was unduly prejudicial.)

An even more consequential decision related to testimony by Dr. C. Paul Wazzan, the damage expert for Musk. In a 59-page report, Wazzan calculated the money that Musk would be entitled to recover from OpenAI if it were found to be liable for breach of charitable trust and from Microsoft if it was found to have aided and abetted OpenAI.

The expert said that if the charitable trust was breached, Musk would be entitled not just to the return of his contributions (the technical term was “disgorgement” of his contributions), but also the value that those contributions created. In other words, Musk would be entitled to recover the value of OpenAI that he had contributed and also the value attributable to those contributions.

To do that Wazzan quantified Musk’s charitable contributions and other non-financial support to OpenAI in its early days as a startup and then calculated the percentage of the value of the enterprise that those contributions represented at that time (50 to 75%).

After some adjustment, he then applied the modified percentage to the value of OpenAI — no longer a startup but now a multi-billion dollar enterprise — to calculate the amount of wrongful gains that Musk’s contributions created for OpenAI and Microsoft.

He gave a range with the upper end being $134 billion, divided between the gains of OpenAI (up to $109 billion) and Microsoft (up to $25 billion).

To put those amounts in context, the largest civil jury verdict in U.S. history is generally considered to be the $145 billion award in a Florida tobacco trial, but it was overturned on appeal. Moreover, the tobacco award was largely punitive damages — not damages suffered by the plaintiff but additional amounts to punish the wrongdoers. Wazzan’s calculation did not include punitive damages.

OpenAI argued that there was no science to support Wazzan’s approach and he did not follow an accepted methodology to reach his number.

The judge said that Wazzan had the academic credentials and experience to testify on the issue of valuation. She found that he had adequate support for his analysis. She found the defendants’ complaint about the lack of an accepted methodology “unpersuasive in light of the unique factors at play in valuing Silicon Valley startups and the relatively novel factual circumstances underlying Musk’s claims.”

She said, “the issue before the Court is not whether the Court agrees with Dr. Wazzan but whether in the context of the dispute, the methodology is unreliable or based on insufficient facts and data.” She concluded that he met that test.

With those rulings, it looked like Musk was within weeks of getting his claims in front of a federal jury with as much as $134 billion in the balance.

The pot is stirred again

But before Musk’s team could get comfortable, on March 31, Gonzalez Rogers dropped two more rulings that changed the playing field again.

In the first, she addressed the issue of punitive damages.

Musk planned to ask for “punitives,” as lawyers say, in addition to the disgorgement of the wrongful gains that Wazzan’s testimony quantified. Punitive damages are amounts assessed against a defendant — on top of whatever damages are awarded to compensate the plaintiff for its injury — to punish it for egregious conduct.

The possibility of hitting the defendant with a hefty punitive damage award is often a significant leverage point for a plaintiff, though with $134 billion already on the board, its weight may have been diminished here. But it didn’t matter, because Gonzalez Rogers found that Musk was not legally entitled to ask for them. The reason?

Turns out that Musk’s choice of remedy — disgorgement of ill-gotten gains — was technically under the court’s “equity” jurisdiction, and punitive damages cannot be awarded in equity, so the judge barred Musk from making any claim for punitives.

Moreover, under equity jurisdiction the decision on the amount of disgorgement was allocated to the judge, not the jury. That meant that it would be Gonzalez Rogers — not the jury — who would be deciding the amount of recovery.

That was not a good thing for Musk because even though Gonzalez Rogers cleared Wazzan’s testimony for presentation to the trier of fact, she had commented that she had not been particularly impressed by his opinion.

That wasn’t all.

The statute of limitations

Gonzalez also made a key ruling on the statute of limitations that favored OpenAI.

Statutes of limitations are rules that bar old and stale claims. They serve the purpose of encouraging lawsuits to be brought when witnesses are available and memories fresh but do so with the blunderbuss weapon of barring claims brought after the limitations period has run. The claims that Musk asserts have limitation periods of two, three or four years, depending on which claim. The longest period — four years — applies to the breach of charitable trust claim.

Musk brought the lawsuit on Aug. 5, 2024, which meant that claims that occurred before Aug. 5, 2020, were potentially barred by the statute of limitations.

However, calculating the period is not so cut and dried. Generally, the limitations period does not start running until the plaintiff knew (or should have known) that the claim existed. The idea is that if the plaintiff doesn’t know that they have a claim, it isn’t fair to count that time against them.

The statute of limitation issue had serious potential to affect Musk’s case, because many of the key events took place outside of even the four-year limitations period.

Some important events occurred before Aug. 5, 2020.

• Musk resigned from the nonprofit OpenAI Inc. board on Feb. 21, 2018
• The for-profit subsidiary, OpenAI, L.P., was formed on Sept. 19, 2018
• The nonprofit transferred substantially all of its intellectual property to the for-profit subsidiary in late 2018 and early 2019.
• Early investors invested $133 million (target redemption amount $13.3 Billion) on March 1, 2019.
• Altman became the CEO of OpenAI, L.P. on April 1, 2019.
• Microsoft invested $1 billion (target redemption $20 Billion) on July 2, 2019, and at the same time entered its technology-sharing agreement.
• OpenAI Inc. amended its charter on April 23, 2020, to drop the idea that its technology would be “open source” though the “for humanity” idea remained.

Those events were before the cutoff for the four-year limitation, and there are more significant events that occurred before the two- and three-year cutoffs. Altman’s team will argue that those events — or some of them — both evidence a breach of the charitable trust and that Musk knew or should have known about the issue at the time. If the jury agrees, the claim — along with others with even shorter deadlines — could be dismissed for being filed too late.

To avoid the risk that some or all of the claims would be completely barred, Musk asked Gonzalez Rogers to tell the jury that his claims didn’t just arise at a single time but continued to arise with each additional act that was improper. So, for example, if the claim for breach of charitable trust first arose back in late 2018 when the nonprofit’s intellectual property was transferred to the for-profit sub, and therefore arguably barred by limitations, another claim would arise when the next act that breached the trust occurred.

This was Musk’s version of the song from Annie: there would always be a new claim tomorrow.

But Gonzalez Rogers did not agree. She said she wouldn’t give the jury that instruction.

The upshot?

While how this will play out at trial remains to be seen, it seems reasonable to expect that OpenAI will bore down on the key questions of what Musk knew and when he knew it.

It is always hard to prove a negative, and given his notorious intelligence, resources and passion for the project, it may be hard for Musk to prove that he was kept in the dark while key events occurred. And even if he shows that he did not know, there will still be the question of whether he should have known.

The ruling gave OpenAI a lot to work with.

In Part Four we will discuss one more last-minute rejiggering of the plans seemingly locked in place a month ago.