The Great Walkout: How Elon Musk Lost Every xAI Co-Founder and 80 of His Best Engineers

When SpaceX acquired xAI in February 2026, Elon Musk described it as a unification — two of his most important companies merging into a single, vertically integrated force that could outpace every AI lab on the planet. Three months later, all 11 of xAI's original co-founders are gone. More than 80 researchers and engineers have followed. The supercomputer in Memphis is still humming, Grok is still shipping updates, and Musk is still posting about AI dominance — but the people who actually built the thing have quietly walked out the door.

The Merger That Broke the Culture

xAI was founded in 2023 by a cohort of researchers who mostly came from OpenAI, DeepMind, and Google Brain. The founding culture reflected that lineage: deliberate, research-oriented, comfortable with long cycles between breakthrough and product. SpaceX is the opposite. It is famous — and genuinely admired in some circles — for moving faster than any aerospace organization in history, often by discarding procedures that most engineers treat as load-bearing. When the two companies merged, those two cultures did not harmonize. They collided.

Sources familiar with the departures describe a pattern that repeated across teams: Musk set aggressive timelines for new Grok capabilities, engineers flagged that the timelines required cutting corners on safety testing and evaluation infrastructure, and the corners got cut anyway. The pre-training team — which handles the foundational model training that everything else depends on — saw its lead, Juntang Zhuang, depart, a loss that multiple sources describe as particularly damaging. Key leaders in world models, coding, and Grok voice followed. By late March 2026, Manuel Kroiss and Ross Nordeen — the last two remaining co-founders — left as well, completing a cofounder exodus that had no precedent at any comparably valued AI lab.

Where They're Going

The departing talent did not disappear quietly into retirement. At least 11 xAI engineers joined Meta's AI research division, where the recruiting pitch was straightforward: Yann LeCun's lab, academic research culture, competitive compensation, and no mandatory tweets. At least seven went to Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati. Murati raised $2 billion in early 2026 and has been deliberately targeting researchers with strong pre-training and alignment backgrounds — exactly the profile xAI alumni carry.

The pattern matters because it represents a direct transfer of institutional knowledge away from one of the most heavily capitalized AI projects in the world. These are not junior engineers frustrated with bureaucracy. They are the people who trained Grok 1 through Grok 4, who designed the Colossus supercomputer's training infrastructure, who built the evaluation pipelines that determine whether a model is ready to ship. Their departure does not make SpaceXAI incompetent, but it does mean the lab is rebuilding institutional knowledge at exactly the moment it is trying to accelerate.

What's Actually at Stake

The stakes here extend beyond hurt feelings and LinkedIn updates. SpaceXAI is currently developing what it describes as the next major Grok release, competing directly with GPT-5 and Gemini Ultra. It operates the Colossus supercomputer — at 200,000 Nvidia H100s, one of the largest AI training clusters ever assembled. And SpaceX has been signaling for months that a public offering is being prepared, one that may value the combined entity north of $350 billion.

A talent exodus at this scale introduces real risk into each of those threads. Model quality depends on the people who understand what went wrong in previous training runs. Training infrastructure at Colossus scale requires engineers who built and debugged it from the start. And institutional investors doing IPO diligence tend to notice when a company's entire founding research team has left in a single quarter. Musk publicly acknowledged that xAI "wasn't built right" in a March post, framing the SpaceX integration as a rebuild rather than a continuation — a characterization that may be honest but is not exactly reassuring to anyone evaluating the company's technical trajectory.

The Harder Question

The SpaceXAI situation is unusual in its scale, but it points to a tension that shows up across the AI industry: the best AI researchers tend to come from environments that value intellectual autonomy and long research horizons, while the companies best positioned to productize their work tend to operate with the kind of velocity that compresses both. OpenAI navigated this — imperfectly, with significant drama — for years before the culture tipped decisively toward product. Google DeepMind has been managing it since the 2023 merger that joined DeepMind's academic culture with Google Brain's engineering culture. The consensus is that it takes years to find an equilibrium, and that the equilibrium often involves losing people in the process.

Musk has never been interested in that kind of patience. The question is whether, at the frontier of AI capability, that impatience costs more than it buys.