Elon Musk and OpenAI are locked in an AI arms race? Sam Altman’s lawyers think so
Elon Musk wrapped up his testimony at the Oakland federal courthouse last Thursday, and what should have been just another chapter in a corporate feud turned into a full-on spectacle about who actually wants to save the world from artificial intelligence — and who just wants to win it.
Sam Altman‘s defense went straight for the jugular during cross-examination: lawyers argued that Musk didn’t file this lawsuit out of altruism, but because he lost the internal power struggle within OpenAI and decided to compete from the outside, with his own AI company in his back pocket.
And therein lies the irony that even had the judge raising her eyebrows 🤨
The man who claims he wants to protect humanity from the risks of artificial intelligence is the same one who founded xAI — a company created, by his own admission, to pursue exactly the kind of technology he says is too dangerous in the wrong hands.
What sits at the heart of this dispute goes way beyond an allegedly broken promise. We are talking about billions of dollars, the future of AGI, and a question nobody in that courtroom could answer clearly: Does Musk really want to protect the world, or does he just want to lead the race? 🚀
What brought all of this to court
To understand the scale of this battle, we need to rewind a bit. OpenAI was co-founded in 2015 as a nonprofit organization by Elon Musk, Sam Altman, and Greg Brockman, the company’s current president. The mission was pretty clear: develop artificial intelligence safely and openly, ensuring its benefits would be shared with all of humanity, rather than concentrated in the hands of a small group of people or corporations.
Musk stated in court that he co-founded the organization altruistically, motivated by fear of the dangers AI could pose. During its early years, OpenAI operated as a research organization focused on developing open-source models, and everything seemed aligned with that original vision.
The trouble started showing up in 2017, when company executives — including Altman, Musk, Brockman, and Ilya Sutskever, one of the most important computer scientists in the organization since its founding — began discussions about creating a for-profit subsidiary. The rationale was practical: OpenAI needed more money to compete in a field that was growing rapidly, especially given the ambitious goal of developing artificial general intelligence, or AGI.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
For Musk, this shift represented a direct betrayal of the organization’s founding principles. The relationship between co-founders deteriorated, and Musk left OpenAI in early 2018. The gap widened over the following years, especially after Microsoft became the company’s largest investor, pouring in billions of dollars and securing an exclusive licensing deal for its technology.
When ChatGPT exploded in popularity in late 2022 and OpenAI became practically synonymous with artificial intelligence for the general public, tensions hit a new level. In November 2024, Musk finally filed the lawsuit against OpenAI and Microsoft, alleging that his former colleagues had violated the founding agreements by abandoning the nonprofit mission. The company is now valued at a staggering 852 billion dollars.
The defense strikes back: control, not protection
If Musk’s narrative revolved around safety and principles, Sam Altman‘s and Microsoft‘s defense didn’t hold back. The lawyers painted a completely different picture of events, suggesting that the real motivation behind the lawsuit was Musk’s frustration at failing to take control of OpenAI.
According to William Savitt, Altman’s lead attorney, Musk tried to seize control of the company at least twice during the discussions about creating the for-profit subsidiary. In his first attempt, he reportedly insisted on holding a majority stake in any for-profit entity, controlling the board of directors, and serving as CEO. In the second, he proposed that OpenAI be folded into Tesla, where he was already the top executive.
Savitt also stated that when these proposals were rejected, Musk began withholding quarterly contributions of five million dollars he had committed to making, pressuring the company to accept his terms. And when that strategy didn’t work either, Musk simply walked away from the organization.
But not without causing damage first, according to the defense. Savitt accused Musk of recruiting OpenAI employees on his way out in early 2018, including Andrej Karpathy, one of the organization’s founding members, who was brought over to Tesla. This kind of accusation is always a sore spot in the tech startup world, where talent is the most valuable asset a company can have.
The message to Mark Zuckerberg
One of the most revealing moments of the trial came when Savitt directed the jury’s attention to a message Musk sent to Mark Zuckerberg in February 2025. In it, Musk asked whether the Meta CEO would be open to the idea of making a joint bid for OpenAI‘s intellectual property.
This message adds another layer of complexity to Musk’s narrative. If the genuine concern was AI safety, why would the answer be trying to buy the company’s technology alongside another tech billionaire? The defense used this evidence to reinforce their argument that Musk’s real interest was always commercial and strategic, not philanthropic.
The distillation of OpenAI models
Another point that drew significant attention during cross-examination was the revelation that xAI had partially engaged in so-called model distillation from OpenAI. In simple terms, distillation is a technique where the output of a more powerful AI model is used to train or refine a smaller one. This practice violates OpenAI‘s terms of service, and the fact that Musk’s company was involved in this kind of activity considerably weakens the image of someone who positions himself as a champion of ethics and transparency in AI development.
AGI: the trophy everyone wants, but nobody can properly define
One of the most interesting points that emerged during the trial was precisely how difficult it is to define what AGI actually is. Artificial General Intelligence is, roughly speaking, the concept of an AI capable of performing any cognitive task a human can — learning, reasoning, solving problems in entirely new contexts, adapting to unexpected situations. It is the kind of technology that moves beyond specialized tools and enters what many people call, perhaps prematurely, true intelligence.
Both OpenAI and xAI are racing toward this goal. And this race is not cheap. We are talking about massive computing infrastructure, research teams with the best talent on the planet, and strategic partnerships with governments and major corporations.
Musk, during his testimony, tried to downplay xAI‘s competitiveness compared to OpenAI. He described his company as having only a few hundred employees and a small market share. In his words, xAI is technically competitive, but much smaller than OpenAI.
However, that strategic humility was quickly challenged by a post Musk himself made on the X platform. In March 2026, he wrote that Tesla would be one of the companies to develop AGI and probably the first to do so in humanoid and atomic manipulation form. Musk also stated in court that Tesla is not pursuing AGI, but that post told a different story to the jury. The contradiction was laid bare pretty clearly. 🤔
Tense moments in the courtroom
Musk’s three days of testimony were not exactly smooth sailing. The atmosphere got particularly heated during cross-examination led by William Savitt on Wednesday afternoon, when both the lawyer and Judge Yvonne Gonzalez Rogers had to repeatedly ask Musk to actually answer the questions being posed. At one point, Musk accused Savitt of intentionally trying to trick him.
But the most heated moment of the trial so far happened before the jury was even called into the room, on Thursday morning. During a discussion about what AI safety expert Stuart Russell — a computer science professor at UC Berkeley and co-author of the standard textbook Artificial Intelligence: A Modern Approach — would be willing to testify about, Musk’s lawyers argued that Russell should be allowed to discuss the climate risks associated with AI, claiming that everyone could die.
Judge Gonzalez Rogers’ response was spot-on and pretty much sums up the central irony of this whole case. She observed that it is ironic that the lawyers’ client, despite claiming these risks, is building a company in the exact same space. She added that she suspects there are people who do not want to put the future in Mr. Musk’s hands.
What is at stake beyond the money
It is easy to reduce this case to a billionaire fight over control of one of the most valuable companies in the world. But what is being debated in this courtroom has implications that go far beyond the numbers. OpenAI is in the process of transitioning to a broader for-profit corporate model, which involves a significant restructuring of its governance. If Musk can slow down or complicate this process through litigation, the impact on the organization’s plans could be substantial — and could directly affect how fast artificial intelligence-based technologies reach the market and end users.
There is also the question of legal precedent. If the court decides that co-founders of nonprofit organizations have legal rights over how the original mission should be preserved during future structural transformations, that significantly changes the landscape for other tech organizations operating under hybrid structures or considering similar transitions. The Musk versus OpenAI case could become an important legal reference for the AI startup and research organization ecosystem in the United States and, by extension, influence how these discussions are handled in other countries.
On top of that, there is the symbolic impact. OpenAI is, today, the most recognizable name in the world when it comes to artificial intelligence for the general public. Any hit to its credibility or operational structure affects how everyday people perceive how these technologies are developed and by whom. And at a time when governments around the world are trying to figure out how to regulate AI, having one of its original creators publicly pointing out that the mission was betrayed is not exactly great news for the industry as a whole.
The race nobody can stop
What this trial in Oakland is making increasingly clear is that the race for artificial general intelligence has reached a stage where the lines between commercial competition, genuine safety concerns, and personal ambition blur in ways that are nearly impossible to untangle. Musk, Altman, Microsoft, Meta, and dozens of other companies and researchers are all running in the same direction, with absurd amounts of money and growing pressure to deliver results before the competition does.
Microsoft, which until recently held exclusive rights to license and sell OpenAI‘s technology, remains the organization’s primary investor and strategic partner. Attorney Russell Cohen, representing Microsoft, was particularly effective in pointing out the timeline of events. He highlighted that Musk only filed the lawsuit in November 2024 — after he had already formed xAI. The sequence of events suggests the lawsuit may have been driven more by competition than by founding principles.
Meanwhile, expert Stuart Russell took the witness stand on Thursday afternoon to discuss the real risks of artificial intelligence. Russell is one of the most respected voices in academia when it comes to AI safety, and his participation in the trial added a technical perspective that goes beyond the corporate disputes between the parties involved.
The trial has not ended yet, and new hearings are expected in the coming weeks. But regardless of the legal outcome, one thing is already clear: the race for AGI is so intense, so full of money and ego, that not even its own architects can agree on what they are building — or for whom.
