Here's why Meta's $1 million AI star scientist quit Superintelligence Lab just five months in

Meta’s grand experiment to build “superintelligence” is already facing cracks. Just five months after joining with a million-dollar salary, AI scientist Rishabh Agarwal has decided to leave the company’s newly created Superintelligence Lab, a unit that Mark Zuckerberg had pitched as central to Meta’s future.

Agarwal announced his decision in a post on X: “It was a tough decision not to continue with the new Superintelligence TBD lab, especially given the talent and compute density. But after 7.5 years across Google Brain, DeepMind, and Meta, I felt the pull to take on a different kind of risk.”

For Meta, the loss stings. Agarwal wasn’t just another hire. An alumnus of IIT Bombay and Mila–Quebec Artificial Intelligence Institute, his career has taken him through Google Brain, where his reinforcement learning research won acclaim, to DeepMind, where he worked on advancing large language models. When Meta managed to lure him in April, it was seen as a coup, proof that Zuckerberg’s promise of billions in investment and access to “unparalleled compute” could attract some of the brightest minds from OpenAI, Google, and xAI.

But Agarwal isn’t alone in heading for the door. Wired has reported that at least three researchers have walked away from the lab in recent weeks. Two of them, Avi Verma and Ethan Knight, have already returned to OpenAI, underscoring how difficult it may be to keep hold of elite researchers in today’s AI talent wars.

Zuckerberg’s vision of building superintelligence, unveiled just two months ago, was supposed to be Meta’s boldest bet yet. He even recruited Scale AI CEO Alexandr Wang to help lead the charge. The company offered multi-million-dollar packages in its aggressive hiring spree, pulling talent from competitors at a pace that rattled the industry.

Yet the early departures highlight a reality: salaries and compute alone may not be enough to keep top researchers in-house. The reasons are complex, from cultural fit to scientific freedom, but they point to turbulence ahead for Meta’s ambitions.

As for Agarwal, his next step remains unknown. Whether he joins a rival, launches a startup, or pursues academic research, his departure symbolises the volatility of the AI arms race and the difficulty even tech giants face in holding on to the people they need most.