OpenAI and Elon Musk's "Century Lawsuit" Behind the Scenes: A Trial Concerning AI Governance Structures

Author: Deep Tide TechFlow

April 28, federal court in Oakland, California.

Elon Musk took the witness stand and was questioned by lawyers for nearly two hours. He traced his childhood in South Africa, the founding of SpaceX—from “The Terminator” to “Star Trek”—trying to convince the nine jurors that everything he has done in his life has been to save humanity.

Then he said one thing: “If the verdict is fine with looting charities, American charitable giving will be destroyed.”

On the surface, this case is a private grudge between two tech billionaires. Musk is seeking to remove Altman, restore OpenAI’s nonprofit status, claim $134 billion in damages, and state that all compensation will go to OpenAI’s charitable entity.

OpenAI’s lawyer Bill Savitt opened with a different version: “We’re here because Mr. Musk didn’t get his way at OpenAI. He stepped away, saying they would definitely fail. But my client had the courage to succeed without him.”

Two narratives, each with its own script. But what’s really worth dissecting isn’t who is lying.

The 38 million nuclear button

Between 2016 and 2020, Musk donated about $38 million to $44 million to OpenAI. Using the highest figure, that’s about 0.005% of OpenAI’s current valuation of $852 billion.

With this money, he now has the standing to ask the court to dismantle the structure of a trillion-dollar company, remove the CEO and president, unwind the partnership with Microsoft, and claw back hundreds of billions of dollars in “unjust enrichment.”

That couldn’t happen in a normal business world. If you only own 0.005% of a company, you can’t even get into the door of shareholders’ meetings. But OpenAI’s starting point is 501©(3)—a tax-exempt charitable organization under U.S. law. Musk’s money was a donation, which enjoyed tax deductions, and therefore gave the donor legal authority to pursue action for the charity’s deviation from its mission.

Many people think a donation is just giving money. But under U.S. charitable trust law, if you can prove that the organization has deviated from its founding mission, the donor has standing to sue. The amount doesn’t matter.

In other words, the $38 million gave Musk not equity, but a nuclear button.

And at OpenAI’s most critical moment, that button was pressed. OpenAI had just completed a $122 billion funding round, valued at $852 billion, and was preparing for an IPO in this year’s fourth quarter. On the Kalshi prediction platform, the probability that Musk would win is 47%.

In fact, OpenAI’s biggest risk is its legacy corporate structure. It grew a trillion-dollar body, but it has been wrapped in a 501©(3) cloak. That cloak can be torn off at any time—and the person doing the tearing might only need to pay a small price.

Silicon Valley’s open secret

OpenAI is not the only AI lab walking the tightrope between a “nonprofit mission” and “business ambition.”

This game has a template in Silicon Valley. First, establish as a nonprofit, using the banner of “benefiting humanity” to attract top talent and early funding; then, when it’s time to burn money, embed a for-profit subsidiary. The nonprofit shell keeps the mission narrative, while the for-profit entity handles making money and fundraising.

Mozilla has done this, and OpenAI is no exception. Founded as a nonprofit in 2015, it created a limited-profit subsidiary in 2019, and in 2025 split into a Public Benefit Corporation (PBC)—changing structures, raising funds, and growing along the way.

Anthropic chose another path. From the start, it registered directly as a Delaware public benefit corporation—going straight down the commercial-entity route—but added a governance body called a “Long-Term Benefit Trust” (LTBT) to constrain the company’s behavior. Anthropic’s co-founders likely looked at OpenAI’s governance dilemma and chose a structure from the beginning that carried no nonprofit baggage.

But the key question is: who do these structures constrain?

OpenAI’s nonprofit board tried to oust Altman in November 2023. That “palace intrigue” lasted less than a week. Altman returned with Microsoft’s support, and the board members who tried to remove him ended up getting removed. At the moment when the nonprofit governance structure was supposed to play a real role, it was crushed by business forces.

OpenAI’s lesson is that nonprofit structures are a shield in the early stage, decoration in the mid stage, and a loophole in the later stage. They can’t protect the founding mission, and they leave external attackers with a perfect entry point.

The real chess game outside the courtroom

After talking about the structural issues, let’s look at the people.

In court, Musk compares himself to a savior of humanity. But look at what condition his own AI company, xAI, is in now.

Founded in 2023, its valuation surged to $230 billion in 2025. The speed is astonishing. But by early 2026, things started to change. In February, SpaceX acquired xAI, followed by large-scale layoffs and restructuring. Co-founders left one after another. By the end of March, only Musk remained among the 11 co-founders. In April, the CFO left, and SpaceX’s Starlink Vice President was parachuted in as xAI’s new president.

After SpaceX took over, xAI basically became a department—not an independent company. The reason the founders left is simple: they joined an AI lab, not a SpaceX subsidiary.

And on the enterprise side? Grok claims 64 million monthly active users, but that’s because it’s embedded in X’s interface—when users open X, they’re counted. Revenue from pilots with Morgan Stanley and Palantir is “in the hundreds of thousands to millions of dollars” range. xAI’s independent annualized revenue (excluding X’s ads and subscriptions) was about $500 million at the end of 2025.

Meanwhile, OpenAI’s monthly revenue had already reached $2 billion by March 2026.

Musk, standing in court holding an AI company with the founding team largely gone, enterprise revenue nearly zero, and one that has already been absorbed by SpaceX, is demanding the dismantling of the world’s largest AI company.

He says it’s for humanity. OpenAI’s lawyer says it’s because xAI can’t beat OpenAI, so Musk is trying to do through legal means what he can’t accomplish commercially.

What is the real reason? Look at the timeline and it becomes clear. Musk filed the lawsuit in 2024, shortly after xAI was founded that same year. In 2025, xAI went on a frenzy of fundraising, trying to catch up with OpenAI in technology and scale. By 2026, xAI was collapsing internally, and around the same time, the case finally went to court.

Perhaps if xAI could compete with OpenAI on technology, Musk might never have taken this step toward court. Suing is Plan B after commercial failure.

The ruins of the winner

Now, zoom out and take in the full picture.

This lawsuit is expected to last three to four weeks. Judge Yvonne Gonzalez Rogers (the same judge who oversaw Epic v. Apple) will base her ruling on the jury’s recommendations, expected in mid-May.

Most legal analysts believe the most likely outcome is a mixed verdict. The court might find that OpenAI violated some fiduciary duties to donors, but it is unlikely to fully unwind the for-profit structure or remove management. Whoever wins or loses, the losing side will appeal to the Ninth Circuit Court, and the case could drag into 2027.

But no matter what the verdict is, this lawsuit has already changed a few things.

For OpenAI, it exposes a weakness. The world’s highest-valued private tech company’s legal structure can be shaken by a small donation from 10 years ago. IPO prospectuses will have to disclose this risk, and every future investor will ask: could there be other historical donors who might step forward?

a16z co-founder Marc Andreessen said, “Whatever the outcome, it has set a template for corporate governance for all cutting-edge AI labs going forward. The path from nonprofit startup to mid-course transformation needs to be reevaluated starting today.”

As for Musk himself, he told a story about saving humanity in court. But his own AI company is being hollowed out: the founding team has run off, and it has been absorbed by SpaceX, turning into a department. With a lawsuit, he’s using legal action to cover a building that’s already starting to collapse.

And Altman had already left the courtroom before Musk testified.

Who’s more panicked and who’s acting—maybe the trial record will provide the answer. Perhaps it won’t be until the Ninth Circuit Court’s appeal window in 2027 that this high-stakes gamble reaches its real showdown.