Rishabh Agarwal, a highly regarded AI researcher, left Meta’s newly formed Superintelligence Lab just five months after joining in April 2025, despite a reported million-dollar salary package.
Background and Career Highlights
- Rishabh Agarwal holds a computer science degree from IIT Bombay and a PhD from the Mila-Quebec AI Institute.
- His career includes key roles at Google, Brain, and DeepMind, specialising in reinforcement learning and large language models.
- At Meta, he worked on advancing “thinking” models by pushing reinforcement learning scalability and improving distillation techniques.
Reasons for Rishabh Agarwal’s Departure
- In a post on X (formerly Twitter), Rishabh Agarwal described the decision as “tough” but cited a desire to take on “a different kind of risk” after 7.5 years in AI research across Google and Meta.
- He mentioned following advice from Mark Zuckerberg: “In a world that’s changing so fast, the biggest risk you can take is not taking any risk.”
- Meta’s Superintelligence Lab reportedly boasts high talent and compute resources, but despite this, Rishabh Agarwal chose to depart.
- His next professional move remains unclear, and we have no idea whether he wants to join another company, start a venture, or academia.
Broader Context: Turmoil at Meta’s AI Lab
- Rishabh Agarwal is not the only high-profile exit; at least two other researchers have left Meta’s Superintelligence Lab recently, with some returning to OpenAI.
- These departures highlight challenges Meta faces retaining elite AI talent amid an intense competitive arms race fuelled by lucrative offers and shifting career priorities.
- Meta continues to restructure and ramp up investment in AI, including facilities with multi-gigawatt compute capacity aimed at advancing artificial general intelligence (AGI).
Meta’s Response and Industry Implications
- Meta downplayed the resignations as normal in a highly competitive hiring climate.
- The exits point to broader industry volatility and the difficulty of building sustainable teams focused on next-generation AI breakthroughs.
- Mark Zuckerberg’s vision for superintelligence, officially announced in mid-2025, remains ambitious, though turbulent in early execution.
This case provides a window into the complexity of the AI talent war and the human factors influencing the future of AI innovation.
What this means for Entrepreneurs
Agarwal’s decision to walk away from one of the most resource-rich AI labs in the world, even with a million-dollar package on the table, shows that innovation in AI isn’t just about compute power or paychecks.
It’s about risk-taking, timing, and vision.
For entrepreneurs, this is a reminder that AI talent is restless, mobile, and motivated by impact more than money.
The volatility in Big Tech’s AI labs creates windows of opportunity for startups to attract world-class researchers looking for more ownership and freedom.
Also read: Mark Zuckerberg is building the most powerful AI Team at Meta
