OpenAI experienced a significant turnover in 2025, losing at least 12 key scientists and executives, primarily to Meta’s Superintelligence Lab. Among the departures were more than half a dozen researchers who joined Meta during the summer, marking a challenging year for the artificial intelligence firm.
The latest departures come in the wake of a restructuring period that began in 2024, which saw a wave of executive exits, including chief technology officer Mira Murati and chief research officer Bob McGrew. Currently, only Sam Altman, OpenAI’s CEO, remains from the original founding team of 11 members.
Key Departures to Meta
During July 2025, several prominent researchers transitioned to Meta’s Superintelligence Lab, which is backed by a significant financial investment aimed at enhancing its AI capabilities. Notable departures included:
Jason Wei, a research scientist who contributed to OpenAI’s advanced models, left for Meta’s Superintelligence Lab in July. Zhiqing Sun, another research scientist, followed suit, joining Wei in the transition.
Hyung Won Chung, part of a trio of researchers who moved to Meta, expressed excitement about the new opportunities on LinkedIn, stating that they were enjoying “building from a clean slate with a truly talent-dense team.”
Shengjia Zhao, who co-created ChatGPT and the GPT-4 model at OpenAI, became the chief scientist at Meta’s lab in July. He is reportedly collaborating directly with Meta’s leadership, including Mark Zuckerberg and chief AI officer Alexandr Wang.
Other researchers who made the shift include Jiahui Yu, known for leading the development of the perception capabilities of large language models, and Hongyu Ren, who was crucial to the advancement of OpenAI’s GPT-4o model. Shuchao Bi, who worked on multimodal and reinforcement learning at OpenAI, also joined Meta to further explore AI agents and post-training enhancements.
Additional Executive Exits
The turnover at OpenAI was not limited to researchers alone. Several key executives have also departed the organization. Larry Summers, former Treasury Secretary and president of Harvard University, resigned from the OpenAI board in November 2025. His resignation followed the release of emails that linked him to Jeffrey Epstein, who faced serious legal charges.
In August, Julia Villagra, who had recently been appointed as chief people officer, announced her resignation. Another executive, Liam Fedus, who served as vice president of research and post-training, left in March to co-found an AI startup named Periodic Labs, which aims to innovate in the development of AI scientists.
Tom Cunningham, previously OpenAI’s data scientist and economic researcher, departed in November to join Model Evaluation and Threat Research, a non-profit organization focused on evaluating AI models. Finally, Hannah Wong, the chief communications officer, shared her plans to embark on her “next chapter” just weeks before the end of 2025.
The loss of these prominent figures highlights the ongoing challenges and changes within OpenAI as it navigates a competitive landscape in artificial intelligence development. Each departure not only impacts OpenAI’s internal dynamics but also raises questions about the future direction of the organization amid increasing competition from industry giants like Meta.
