Meta has officially named Shengjia Zhao, a former OpenAI researcher and co-author of the original ChatGPT paper, as Chief Scientist of its newly established Superintelligence AI group. Zhao, who joined Meta quietly in June 2025, played a pivotal role in the development of OpenAI’s early reasoning models, including “o1”—the model that sparked the now-widespread “chain-of-thought” approach later adopted by Google and DeepSeek. His appointment signals Meta’s growing ambition to position itself at the forefront of the AI race.
CEO Mark Zuckerberg announced Zhao’s promotion on Threads, stating, “our lead scientist from day one.” He added, “Now that our recruiting is going well, and our team is coming together, we have decided to formalise his leadership role.” Zhao will report directly to Alexandr Wang, Meta’s recently appointed Chief AI Officer and former CEO of Scale AI, who has been tasked with steering Meta toward artificial general intelligence (AGI)—a long-term goal focused on building systems that can reason and think at human or superhuman levels.
Meta’s Superintelligence Lab, launched in June 2025, is central to this vision. It operates independently from the longstanding Facebook AI Research (FAIR) unit, which remains under the leadership of AI pioneer Yann LeCun. In a significant restructuring, LeCun now reports to Wang, giving the new Chief AI Officer oversight of Meta’s dual-track AI research efforts.
In recent months, Meta has aggressively recruited top AI talent from major competitors like OpenAI, Google, Apple, and Anthropic. Among the notable hires are Jiahui Yu, Shuchao Bi, and Hongyu Ren—all formerly of OpenAI—alongside Trapit Bansal and engineers from OpenAI’s Zurich team specializing in multimodal AI. This hiring spree has reportedly included lavish compensation offers, sometimes reaching into eight or nine figures, and fast-moving “exploding offers” to secure top researchers quickly.
Zuckerberg has personally led much of this recruitment effort, even inviting candidates to his Lake Tahoe residence for private discussions. Despite currently trailing rivals with its LLaMA 4 model, Meta is placing its bets on a next-generation system, internally codenamed “Behemoth,” expected to launch later in 2025.
“Together we are building an elite, talent-dense team that has the resources and long-term focus to push the frontiers of superintelligence research,” Zuckerberg emphasized, underscoring Meta’s bold push into the future of AI.
