xAI loses two cofounders in 48 hours

Elon Musk's AI venture just lost half of its founding brain trust, and the exodus shows no signs of slowing down.

xAI co-founders Tony Wu and Jimmy Ba resigned within 48 hours of each other, leaving only six of the original 12 co-founders at the company. Alongside the high-profile exits, at least 11 engineers have publicly announced their departures over the past week, with several stating they plan to launch new ventures together.

Post-merger turbulence

The shakeup follows a massive corporate restructuring. Last week, Musk announced the merger of xAI, social network X, and rocket manufacturer SpaceX at a staggering valuation between $1.25 trillion and $1.5 trillion, ahead of a planned IPO later this year.

Following the deal, Musk reorganized xAI into four divisions: Grok Main and Voice, Coding, Imagine (image and video), and Macrohard, a project aimed at digitally emulating entire enterprises. Musk also pitched grand visions of space-based AI data centers, AI satellite factories, and lunar mass drivers.

Yet internal tensions had been mounting. Ba, an assistant professor at the University of Toronto who studied under AI pioneer Geoffrey Hinton, previously oversaw key teams and over a thousand AI data tutors, but saw his responsibilities gradually reassigned. Shortly before Wu resigned, his Slack account was deactivated. Musk addressed the moves publicly, stating the company reorganized to "improve speed of execution" and that scaling "unfortunately required parting ways with some people."

At the same time, xAI is grappling with severe regulatory fallout after Grok generated nonconsensual explicit deepfakes, which recently prompted French authorities to raid X's offices.

Why it matters

For researchers and engineers, xAI has fundamentally pivoted. It is no longer an agile frontier research collective guided by elite academics; it has become an industrialized product engine engineered to support a massive multi-trillion-dollar IPO. Those looking for pure exploratory research are already looking toward smaller, independent teams.

Across the broader landscape, losing half of your founding team within a year is a glaring red flag in the war for scarce AI talent. While Musk frames the exits as trimming fat for late-stage execution, the departures of foundational figures like Ba and Wu hand a clear recruiting advantage to OpenAI, Anthropic, and Google.

In our view, Musk's rush to tie xAI to interplanetary infrastructure and speculative moon projects looks like an attempt to paper over real operational friction. Scaling frontier models requires intense scientific rigor and research stability; turning an AI lab into a high-pressure corporate satellite might accelerate near-term product shipping, but it risks bleeding the very talent that made the models capable in the first place.