How Musk Tried to Bring Altman to Tesla Before Leaving OpenAI

Court documents from the Musk v. Altman trial have shown that Elon Musk attempted to recruit Sam Altman to lead a major AI research lab inside Tesla, just months before leaving OpenAI's board in February 2018. The pitch included offering Altman a seat on Tesla's board, according to emails and testimony presented in federal court.

These newly revealed communications expose a deeper story: a clash over how OpenAI should be structured and who would control it. Messages from 2017 show Musk and his associates were planning a rival AI lab that might be led by either Altman or Demis Hassabis, a co-founder of DeepMind. Shivon Zilis, a Tesla executive, acted as the go-between in these discussions.

The Fight Over OpenAI's Future

At its heart, this case is about control. Musk alleges that Altman and Greg Brockman, OpenAI's president, "effectively stole a nonprofit organization." But the court evidence cuts both ways: internal messages show that Musk himself wanted to absorb OpenAI into Tesla and take it over entirely.

Both sides agreed in 2017 that OpenAI needed to shift from a nonprofit to a for-profit structure. Here's where they diverged. In September 2017, Musk created a for-profit version of OpenAI and asked for a controlling stake—roughly 50 percent—plus complete decision-making power. OpenAI's leadership refused. By January 2018, Musk told them the organization faced "certain failure" without merging into Tesla. A month later, he resigned from the board and pivoted to building AI at Tesla instead.

The Money Behind the Fight

The financial stakes have ballooned since OpenAI began as a nonprofit research group. Musk says he put in $38 million before OpenAI went private. Today, OpenAI is valued at over $800 billion—some reports cite $852 billion—as it moves toward going public.

That $800 billion figure dwarfs earlier compensation battles Musk has fought. A Delaware court previously rejected a $55.8 billion pay package for Musk at Tesla, calling it "250 times larger" than what other company executives typically received. Tesla has since moved its headquarters to Texas, which means Texas courts will now handle future disputes rather than Delaware courts.

A Pattern From an Earlier Era

The recruitment efforts and corporate maneuvering reflect a larger story about competition for top AI talent and organizational control, particularly between 2017 and 2018 when the field was taking shape commercially. This is not unfamiliar territory in technology. When cloud computing emerged in the mid-2000s, we saw a similar dynamic play out at Amazon, Microsoft, and Google—each building their own cloud divisions while raiding the same pool of technical talent and arguing over whether research should remain independent or become tightly tied to corporate strategy.

The trial materials suggest that Musk and OpenAI's founders held fundamentally different visions. Musk wanted to integrate AI research directly into Tesla's automotive and energy business—a vertical approach that could move discoveries from lab to production faster. OpenAI's leadership wanted to keep research separate and raise money from outside partners, a model closer to academic independence.

The Legal Process Unfolds

The case is being heard in Oakland federal court, with jury selection beginning in late April. Musk sought a settlement before trial, though no terms were made public. The judge expects to have a jury verdict by late May on whether OpenAI broke promises it made to Musk.

Other details have emerged in court filings: both sides have been asked to stop using ephemeral messaging apps like Signal and XChat for work communications. This reflects a broader issue in high-stakes technology cases—companies often need to preserve evidence, and disappearing messages make that harder.

The legal questions go beyond simple contract disputes. They touch on how nonprofits should be governed and what duties their leaders owe to founders and investors. Altman has filed statements defending OpenAI's actions, arguing that Musk's departure was about his failure to seize control rather than any organizational betrayal.

What This Tells Us About AI's Future

The trial reveals how engineers and executives were thinking about large-scale AI development back in 2017 and 2018. A Tesla-integrated AI lab, as Musk envisioned it, could have compressed the cycle from a research breakthrough to actual use in cars and power systems. That kind of speed matters in competitive industries.

But there were real tradeoffs. Tightly binding research to a single company's needs can limit what researchers explore and can make it harder to recruit and keep the best talent. OpenAI's looser, partnership-based approach ultimately led to the GPT models and a different competitive trajectory than Musk's vertical integration strategy would have produced.

The case also raises broader questions about how organizational structure shapes what gets built. In AI especially, where talent and capital are concentrated among a few organizations, these decisions—nonprofit versus for-profit, independent versus vertically integrated, open partnerships versus proprietary control—carry real weight. The trial outcome could set a legal precedent for how courts evaluate expectations and promises when mission-driven research organizations transition into commercial enterprises in fast-moving fields.

For now, the jury will decide whether OpenAI broke faith with Musk. But the larger lesson is already clear: the shape of OpenAI's organization, and the choice to resist absorption into Tesla, redirected the entire trajectory of how advanced AI research moved from laboratory to the world.