The man suing OpenAI for betraying its nonprofit roots once tried to poach its CEO with a Tesla board seat and a rival AI lab.
The Summary
- Messages from Shivon Zilis reveal 2017 plans for Musk to launch a competing AI lab, with Altman or DeepMind's Demis Hassabis as potential leaders
- Musk considered offering Altman a Tesla board seat to lure him away from OpenAI
- Trial evidence shows Musk was already positioning to control AI development outside OpenAI before the 2018 split he now claims blindsided him
The Signal
Internal messages from Shivon Zilis, now Neuralink's director of operations, surfaced in court this week. They show Tesla executives in 2017 discussing plans to build an AI lab that would compete directly with OpenAI. The proposed leadership shortlist: Sam Altman or Demis Hassabis, the founder of DeepMind.
This matters because Musk's entire lawsuit hinges on the claim that OpenAI betrayed its nonprofit mission when it took Microsoft's money and went commercial. But these messages show Musk was already planning his exit, already building the alternative. He even floated giving Altman a Tesla board seat to sweeten the recruitment pitch.
"The man claiming betrayal was simultaneously plotting to raid the organization he now says abandoned its principles."
The timeline is clarifying:
- 2017: Musk floats rival AI lab, tries to recruit Altman to Tesla
- 2018: Musk leaves OpenAI's board, citing conflicts with Tesla's AI work
- 2019: OpenAI creates capped-profit structure, partners with Microsoft
- 2023: Musk files lawsuit claiming OpenAI broke its nonprofit promises
The trial testimony reveals what Musk's lawyers hoped would stay buried. He wasn't a faithful steward pushed out by greed. He was already shopping for talent to build his own thing. The "betrayal" narrative only emerged after Altman said no and OpenAI figured out how to scale without Musk's money.
This is the messy truth about early-stage AI companies: everyone was hedging. Musk wanted control. Altman wanted compute. Microsoft wanted access. Nobody knew which bet would pay off, so smart players kept multiple irons in the fire. Musk's mistake was assuming he could recruit the CEO of the company he co-founded while simultaneously positioning that company as the enemy of his master plan.
The Implication
If you're building an AI company today, this trial is a masterclass in why clear governance and honest communication matter more than cap tables. Musk's real mistake wasn't leaving OpenAI. It was leaving without admitting he'd been planning to compete all along, then wrapping his exit in the language of principle.
Watch for more discovery in this case. If Zilis's messages made it to a jury, there are likely others that detail just how far Musk's 2017 recruiting effort went. The narrative of OpenAI as the villain gets harder to maintain when the plaintiff was actively trying to gut the organization years before the "betrayal" he's suing over.