When Elon Musk launched xAI in 2023, he put together a team of 12 cofounders from top AI labs like Google, OpenAI, and DeepMind. The goal was to create AI that helps humanity grasp the universe better, a bold pitch that drew in experts eager to push boundaries. Tony Wu, a former Google research scientist, joined that group with high hopes for what they could build together. Now Wu has become the fourth cofounder to leave in about a year, sharing a short post on X to mark the moment.
Wu kept his farewell positive and vague. He called his time at xAI the ride of a lifetime, thanked Musk and the team, and said it was time for his next chapter. No hints of conflict or big plans ahead, just a nod to small AI powered teams achieving huge things. That tone matches what others have said before him, leaving business watchers to wonder what lies beneath the polite words. For those new to AI startups, this mix of thanks and silence often signals personal shifts rather than blowups.
Look back over the past year, and the pattern emerges clearly. Igor Babuschkin exited in mid 2025 to start his own venture firm, Babuschkin Ventures, aimed at AI safety and early-stage companies. He spoke warmly about long chats with Musk that shaped his ideas on AI’s role in humanity’s future. Greg Yang stepped into an advisory role last month after health issues sidelined him from full time work. Each case shows a cofounder chasing something different, from investment plays to recovery time.
Earlier shifts set the stage too. Kyle Kosic, one of the original engineers, left in 2024 and went back to OpenAI, where he had worked before joining xAI’s startup push. Christian Szegedy, a key researcher, took a chief scientist spot at Morph Labs around the same period. These moves highlight how fluid the top AI talent pool remains, with people circling between labs based on fit, resources, or fresh challenges. No one burned bridges publicly; they just moved on.
Public comments from the leavers focus on growth and opportunity. Wu hinted at wanting more autonomy in an era where AI lets small groups punch above their weight. Babuschkin saw his firm as a way to back multiple AI paths, building on xAI lessons without being tied to one company. Yang laid out health as the straightforward cause. Industry talk fills in the rest, with some online claims of 50% staff turnover over a year sparking debate. Musk countered that real losses are low and xAI outpaces rivals in speed.Â
xAI keeps raising big money despite the changes, pulling in tens of billions in funding at valuations over $200 billion. That cash fuels data centers, models like Grok, and aggressive hiring. Investors seem unphased by the cofounder flux, betting on Musk’s track record over stable leadership. Still, losing four founding voices in quick order tests any startup’s ability to hold its core vision steady.
Business minds track these exits because AI reshapes trillion dollar industries from chips to software. High churn at the top can spread expertise across the field, as ex cofounders launch firms or advise others. It also exposes risks in intense environments where burnout, health, or better offers pull people away. xAI attracts talent with its mission and resources, but retaining it means balancing Musk’s drive with room for individual paths.
