Musk Tried to Recruit Altman to Tesla AI Lab

Leaked Messages Reveal Musk's Secret Plan to Control AI Through Tesla

Newly surfaced messages between Shivon Zilis and Tesla executives reveal that Elon Musk explored a dramatic contingency plan in 2017: recruiting Sam Altman — or even Demis Hassabis — to lead a rival artificial intelligence laboratory, potentially housed under Tesla. The revelations shed new light on Musk's years-long obsession with controlling the trajectory of advanced AI development, a saga that has since spiraled into one of Silicon Valley's most bitter corporate feuds.

The communications, which emerged as part of ongoing legal and media scrutiny surrounding the OpenAI-Musk dispute, paint a picture of a billionaire who saw the future of AI as too important to leave in anyone else's hands. They also underscore just how intertwined the personal ambitions of tech's biggest players have been in shaping the AI landscape we see today.

Key Takeaways From the Revealed Communications

• Musk's inner circle actively discussed launching a competing AI lab as early as 2017, well before his public falling-out with OpenAI.
• Sam Altman was considered a potential leader for the rival venture — the same person Musk would later sue over OpenAI's direction.
• Demis Hassabis, co-founder and CEO of Google DeepMind, was also floated as a candidate to helm the effort.
• Shivon Zilis, a venture capitalist and later Neuralink executive, served as a key intermediary in these discussions.
• Tesla was the proposed home for the new AI initiative, which would have given Musk direct corporate control.
• The plans ultimately did not materialize in their original form, but Musk later launched xAI in 2023 with similar ambitions.

The 2017 Power Play: Why Musk Wanted His Own AI Lab

By 2017, tensions between Musk and OpenAI's leadership were already simmering beneath the surface. Musk had co-founded OpenAI in 2015 with a $1 billion pledge, positioning the organization as a nonprofit counterweight to Google's growing dominance in AI through its acquisition of DeepMind.

But Musk grew increasingly frustrated with what he perceived as OpenAI's slow progress and its inability to compete with Google's vast computational resources. The messages suggest he began exploring alternatives — not as a backup plan, but as a serious strategic initiative.

The idea of housing an AI lab under Tesla made strategic sense from Musk's perspective. Tesla already had access to massive datasets from its autonomous driving program, significant GPU infrastructure, and the engineering talent needed to scale AI systems. A Tesla-based AI lab would have given Musk something he could never achieve at OpenAI: unilateral control.

Altman as Recruit: From Potential Ally to Bitter Rival

Perhaps the most striking revelation is that Musk's team considered Sam Altman himself as a potential leader for the rival lab. At the time, Altman was serving as president of Y Combinator and was deeply involved in OpenAI's governance, though he had not yet assumed the CEO role he holds today.

The irony is difficult to overstate. Musk's camp apparently viewed Altman as someone who shared his concerns about AI safety and could be persuaded to join a venture where those concerns would take center stage. Fast forward to 2024 and 2025, and the two men are locked in a legal battle over OpenAI's transition from a nonprofit to a for-profit structure — a move Musk has called a betrayal of the organization's founding mission.

• Altman ultimately became OpenAI's CEO in 2019, steering the company toward its blockbuster partnership with Microsoft, which has invested over $13 billion in the venture.
• Musk departed OpenAI's board in 2018, citing potential conflicts of interest with Tesla's own AI work.
• By 2023, Musk launched xAI and its chatbot Grok, finally realizing his vision of a personally controlled AI company.
• Musk filed a lawsuit against OpenAI in early 2024, later dropped it, then refiled — alleging the company abandoned its open-source, nonprofit roots.

The trajectory from 'potential recruit' to 'courtroom adversary' captures the volatile nature of relationships in the AI industry, where billions of dollars and existential technology converge.

The Hassabis Factor: Courting Google's AI Chief

Equally remarkable is the mention of Demis Hassabis as a candidate. Hassabis co-founded DeepMind in 2010, which Google acquired for approximately $500 million in 2014. By 2017, he was already one of the most respected figures in AI research, having led the team behind AlphaGo's historic victory over world champion Go player Lee Sedol.

Recruiting Hassabis would have been an extraordinary coup — and an extraordinarily difficult one. Hassabis was deeply embedded within Google's ecosystem, and any attempt to poach him would have triggered significant legal and corporate complications.

The fact that Musk's team even floated the idea speaks to the scale of ambition involved. This was not about building a modest research group. It was about assembling the single most capable AI organization on the planet, led by one of two people Musk considered capable of the task.

Hassabis, for his part, has remained at Google, where he now leads Google DeepMind — the merged entity combining DeepMind and Google Brain. Under his leadership, the division has produced Gemini, Google's flagship family of large language models that directly competes with OpenAI's GPT-4 and its successors.

Shivon Zilis: The Hidden Connector in Musk's AI Network

Shivon Zilis has long operated as a quiet but influential figure in Musk's orbit. A former partner at Bloomberg Beta and a board member at OpenAI, Zilis later joined Neuralink, Musk's brain-computer interface company, as director of operations and special projects.

Her role as an intermediary in these 2017 discussions highlights how Musk's various ventures — Tesla, Neuralink, OpenAI, and eventually xAI — have never operated in true isolation. The same network of advisors and executives has circulated through multiple Musk entities, creating a web of influence that extends across the AI industry.

• Zilis served on OpenAI's board while simultaneously working within Musk's corporate ecosystem.
• Her dual positioning gave Musk's team unique insight into OpenAI's internal dynamics and strategic direction.
• The messages suggest she played a coordinating role, connecting Tesla's executive team with Musk's broader AI ambitions.
• Her continued presence at Neuralink underscores the lasting nature of these professional relationships.

Industry Context: A Pattern of AI Power Consolidation

The 2017 messages fit into a broader pattern that has defined the AI industry over the past decade: the concentration of AI power among a handful of tech billionaires and megacorporations.

Today, the most advanced AI systems in the world are controlled by a remarkably small group. OpenAI (backed by Microsoft), Google DeepMind (owned by Alphabet), Anthropic (backed by Amazon and Google), Meta AI (led by Mark Zuckerberg), and xAI (owned by Musk) account for the vast majority of frontier AI research and deployment.

Compared to the early days of the internet — which saw thousands of startups competing on relatively equal footing — the AI revolution is being driven by entities with access to billions in capital and massive computational infrastructure. The barriers to entry have never been higher, with training runs for frontier models now costing upward of $100 million and requiring tens of thousands of specialized GPUs.

Musk's 2017 plan was, in many ways, an early recognition of this dynamic. He understood that whoever controlled the compute, the talent, and the organizational structure would control the future of AI.

What This Means for the AI Industry Today

These revelations matter for several reasons beyond historical curiosity.

First, they provide important context for Musk's ongoing legal battle against OpenAI. His lawsuit alleges that OpenAI's leadership — particularly Altman — betrayed the organization's founding principles. But the 2017 messages suggest Musk was already planning to circumvent OpenAI years before the alleged betrayal occurred, potentially undermining his legal narrative.

Second, the messages illuminate the personal dynamics that continue to shape AI policy and development. The decisions being made about AI safety, deployment, and governance are not purely technical or institutional — they are deeply personal, driven by rivalries, ambitions, and relationships that date back nearly a decade.

Third, for developers, startups, and businesses building on AI platforms, these revelations are a reminder that the platforms themselves are subject to the whims and conflicts of their controlling figures. The stability of any AI ecosystem depends, in part, on the stability of the relationships at the top.

Looking Ahead: The Musk-Altman Rivalry Intensifies

The Musk-Altman conflict shows no signs of cooling. OpenAI is in the process of restructuring into a for-profit public benefit corporation, a move that could value the company at over $300 billion. Musk has offered to acquire OpenAI's nonprofit assets for $97.4 billion — an offer the company has rejected as a PR stunt.

Meanwhile, xAI raised $6 billion in late 2024 and is rapidly scaling its Colossus supercomputer cluster in Memphis, Tennessee. Musk has positioned Grok as a direct competitor to ChatGPT, integrating it into the X (formerly Twitter) platform and offering it to Tesla vehicle owners.

The next 12 months will likely bring further legal proceedings, competing product launches, and continued public sparring. For the broader AI community, the key question remains: will this rivalry accelerate innovation, or will it distract from the collaborative work needed to ensure AI develops safely and equitably?

What started as a 2017 plan to recruit a potential ally has become one of the defining conflicts of the AI era. The messages from Shivon Zilis and Tesla executives are not just historical artifacts — they are a roadmap to understanding the personal forces that continue to shape the most transformative technology of our time.