According to Business Insider, OpenAI saw a significant exodus of at least 12 executives and researchers in 2025, with the majority—including at least seven over the summer—departing for Meta’s new Superintelligence Lab. Key losses include research scientists like Jason Wei, Zhiqing Sun, and Hyung Won Chung in July, followed by prominent figures like GPT-4 co-creator Shengjia Zhao, who became Meta’s chief scientist. The brain drain started earlier with executives like former CTO Mira Murati leaving in 2024 and continued through November with board member Larry Summers resigning. This leaves CEO Sam Altman as one of just two remaining active members from the original 11-person founding team, marking a dramatic shift in the company’s core leadership and research talent.
Meta is Building a Superteam
Here’s the thing: this isn’t just normal job hopping in tech. This is a targeted, strategic raid. Meta, with its “billion-dollar effort” for AI, is basically assembling an all-star team using OpenAI‘s own playbook. When someone like Shengjia Zhao, a ChatGPT and GPT-4 co-creator, leaves to become Chief Scientist at a rival lab and work directly with Mark Zuckerberg? That’s a statement. And the LinkedIn post from Hyung Won Chung about having “so much fun building from a clean slate with a truly talent-dense team” is a not-so-subtle dig. It suggests these researchers felt constrained or bogged down at OpenAI and see Meta as a fresh, ambitious playground. Meta isn’t just hiring AI talent; it’s buying the institutional knowledge of OpenAI’s biggest breakthroughs.
What’s Left at OpenAI?
So what does this mean for OpenAI? Losing this many core contributors to key models like GPT-4o and the o1 research line is a huge hit to institutional momentum. Think about it. Jiahui Yu led the “Perception” team for multimodal senses. Hongyu Ren was a core contributor to GPT-4o. These aren’t backbench engineers. This level of departure creates a knowledge gap that’s incredibly hard to plug, no matter how much funding you have. With Altman now one of the last original founders standing, the company’s culture and direction are almost certainly shifting under his sole, concentrated vision. The question is, can they rebuild their research mojo while under this intense competitive and internal pressure?
The Broader AI Talent War
But it’s not just a two-horse race between OpenAI and Meta. The other departures show the talent is spreading out, seeding the entire ecosystem. Liam Fedus, a former VP of research, left in March to co-found Periodic Labs, which aims to create an “AI scientist.” That’s a bold, specific bet on a new direction. Tom Cunningham went to a non-profit focused on AI model evaluation and safety. This fragmentation is healthy for the field but challenging for any single company trying to maintain a monopoly on breakthrough ideas. The era of one or two labs holding all the top minds is clearly over. The gold rush is on, and the prospectors are setting up their own claims.
The Altman Factor
All of this circles back to Sam Altman. He’s the constant. After the board drama of 2023 and the executive exodus of 2024, the 2025 research exodus feels like the next phase of a prolonged shakeout. The company is now unmistakably *his* company. That brings focus, but it also raises risks. Does a more monolithic culture drive away the kind of brilliant, independent thinkers who made OpenAI famous in the first place? The departures suggest maybe it does. Going forward, OpenAI’s biggest challenge might not be technical. It might be cultural. Can they keep the next generation of Jason Weis and Shengjia Zhaos from getting that call from Meta? The future of the AI race depends on it.
