Breaking News

Half of xAI's Founding Team Has Left: What It Means for the $1.25 Trillion IPO

February 11, 2026 10 min read

Six of xAI's original 12 cofounders have now departed the company, including two in the past 48 hours. With the SpaceX-xAI IPO targeted for June 2026, the talent hemorrhage raises serious questions about execution capability and investor confidence.

6 of 12
Cofounders Gone
$1.25T
Combined Valuation
2
Left This Week
June '26
Planned IPO

The Exodus Timeline

xAI was founded in July 2023 with a team of 12 cofounders, many recruited from elite AI labs including Google DeepMind, OpenAI, Microsoft Research, and the University of Toronto. Less than three years later, half are gone.

Mid-2024
Kyle Kosic departs
Oversaw infrastructure at xAI. Joined OpenAI - the company Musk has publicly battled in court.
February 2025
Christian Szegedy departs
Former Google engineer and influential AI researcher. Left after roughly 18 months at xAI.
August 2025
Igor Babuschkin departs
Left to launch his own venture capital firm. Had previously worked at DeepMind.
January 2026
Greg Yang departs
Former Microsoft researcher. Cited health-related reasons for departure.
February 10, 2026
Tony Wu departs
Led xAI's reasoning team. Announced his departure on Monday evening.
February 11, 2026
Jimmy Ba departs
Reported directly to Musk. Credited with key research behind Grok 4. Left after internal tensions over model performance demands.

The Jimmy Ba Departure Is Significant

Ba is a University of Toronto professor whose research directly influenced xAI's Grok 4 models. He reported to Musk personally. The Financial Times reported his resignation was triggered by tensions within the technical team over demands to improve Grok's performance amid growing competition from OpenAI and Anthropic.

It's Not Just Cofounders

The departures extend well beyond the founding team:

Why Are They Leaving?

No single factor explains the exodus, but several pressures have converged:

1. Grok's Product Problems

xAI's flagship chatbot has been a liability as often as an asset:

For research-minded cofounders, building systems that generate harmful content isn't what they signed up for. Safety-conscious departures (like Jimmy Ba, who led research and safety) suggest deep disagreements about how fast to ship versus how carefully to test.

2. The Musk Management Style

Musk is known for demanding extreme pace and dedication. At xAI, that reportedly translated into pressure to match or beat OpenAI and Anthropic on benchmarks, sometimes at the expense of safety testing. The FT reported that internal tensions over model performance demands were a direct factor in Ba's departure.

3. Competition for Talent

The AI talent market is the most competitive in tech history. OpenAI, Anthropic, Google DeepMind, and Meta AI are all aggressively recruiting. Kyle Kosic going directly to OpenAI - the company Musk is suing - illustrates how easily talent can walk to a competitor.

4. Better Opportunities

Igor Babuschkin left to start a VC firm. Others may be eyeing their own startups. With AI funding at record levels, top researchers can raise capital easily.

The IPO Problem

This is where the cofounder exodus goes from embarrassing to potentially damaging.

IPO Timeline: ~4 Months Away

Musk announced the SpaceX-xAI merger in early February 2026, creating a combined entity valued at roughly $1.25 trillion. He aims to take the company public as early as June 2026. Losing half your founding team four months before an IPO is not the narrative you want in your S-1 filing.

Here's why investors should pay attention:

The Remaining Cofounders

Six of the original 12 remain. While their identities haven't all been publicly confirmed, the remaining team is under pressure to demonstrate stability. Musk himself is technically a cofounder, and his personal involvement in xAI's direction has intensified since the SpaceX merger.

Historical Precedent

OpenAI also experienced significant cofounder departures - Elon Musk himself left the board in 2018, and several other early leaders departed over the years. OpenAI survived and thrived. But OpenAI had more time to rebuild before going to market. xAI doesn't have that luxury with a June IPO target.

What This Means for xAI's Competitors

Factor xAI OpenAI Anthropic
Founding Team Retention 50% gone Multiple departures over time Core team intact
Latest Model Grok 4 GPT-5.3-Codex Claude Opus 4.6
Team Size ~500 ~3,000+ ~1,000+
Funding/Valuation $250B (xAI portion) $340B ~$60B
Product Controversies Major (imagery, bias) Moderate Minimal

For OpenAI and Anthropic, xAI's talent losses are a recruiting opportunity. For Meta, Google, and Amazon, it's a reminder that building and retaining an AI research team is one of the hardest problems in tech.

What Founders Should Take Away

1. Founder Retention Is a Signal

If you're evaluating whether to build on xAI's APIs (Grok) or partner with them, the cofounder departures are a data point. A company losing half its founding team in under three years suggests internal instability that could affect product reliability and roadmap execution.

2. The Talent War Is Real

If xAI - backed by the world's richest person and a $250B valuation - can't retain its cofounders, the AI talent market is even tighter than it appears. For founders competing for engineering talent, this means:

3. Speed vs. Safety Trade-offs Have Consequences

xAI's aggressive shipping pace produced harmful outputs (the Grok imagery scandal, the antisemitic outputs). The departures suggest at least some cofounders disagreed with the approach. For founders building AI products: the technical staff you most want to keep are often the ones who care most about doing things right.

4. IPO Timing Matters

Musk is pushing for a June 2026 IPO. The cofounder departures could force a delay, or at minimum, create headwinds in the roadshow. If you're tracking AI company valuations (as an investor or a potential acquisition target), watch how the market reacts to this story.

What to Watch Next

  1. More departures: If a seventh or eighth cofounder leaves, the narrative shifts from "talent churn" to "collapse." Watch LinkedIn and X for announcements
  2. IPO filing: The S-1 will reveal how xAI frames the departures and what retention packages are in place for remaining team
  3. Grok 5 progress: xAI needs a strong model release to counter the narrative. Can the remaining team deliver?
  4. Regulatory fallout: The Grok imagery probes in Europe and Asia could add legal costs and distraction
  5. Musk's attention: With SpaceX, Tesla, X, Neuralink, and now DOGE commitments, how much focus can Musk actually give xAI?

Bottom Line

Losing half your founding team is never a good look. Losing them four months before the largest tech IPO in history is a genuine problem.

xAI has resources that most AI companies can only dream of - SpaceX infrastructure, Starlink data, and Musk's ability to raise capital. But resources without the right people to deploy them are just expensive overhead.

The question isn't whether xAI will survive. It almost certainly will. The question is whether it can compete at the frontier of AI research with a depleted founding team, a controversial product, and the immense distraction of an IPO - all at the same time.

For founders: don't build your strategy around any single AI provider's stability. Diversify your model dependencies, and pay close attention to the talent signals. The companies that retain their best researchers are the ones most likely to deliver reliable, improving products over time.

Track AI Industry Power Shifts

Get weekly analysis on who's winning and losing in the AI race - and what it means for founders.

Welcome! You'll get our next issue.
Something went wrong. Please try again.