Meta CEO Mark Zuckerberg announced on Friday that former Openai researcher Shengjia Zhao will lead research efforts at the company’s new AI unit, Meta Superintelligence Labs (MSL). Zhao has contributed to some of Openai’s biggest breakthroughs, including CHATGPT, GPT-4 and the company’s first AI inference model O1.
“We look forward to sharing Shengjia Zhao’s role as a lead scientist at Meta Superintelligence Labs,” Zuckerberg said in a post in a thread on Friday. “Shengjia co-founded a new lab and was our lead scientist from day one. Now that we are on track and our teams are coming together, we have decided to formalize his leadership role.”
Zhao will set up the MSL research agenda under the leadership of Alexandr Wang, former CEO of Scale AI, who was recently hired to lead the new unit.
King, who has no background in research, was seen as a somewhat unconventional choice to lead the AI lab. The addition of Zhao, a reputable research leader known for developing frontier AI models, concludes the leadership team. To fill out the units further, Meta hired several high-level researchers from Openai, Google Deepmind, Safe Superintelligence, Apple, and humanity to draw researchers from Meta’s existing Basic AI Research (FAIR) lab and Generated AI Unit.
Zuckerberg said in his post that Zhao has pioneered several breakthroughs, including a “new scaling paradigm.” The CEO of Meta may be referring to Zhao’s work on Openai’s inference model O1, listed as a basic contributor along with Openai co-founder Ilya Sutskever. AI inference models are an important area of MSL, as META does not currently offer O1 competitors.
In June, Zhao reported that he would join the Meta Super Intelligence Lab, along with three other influential Openai researchers, Jiahui Yu, Shuchao Bi and Hongyu Ren. Meta also recruited three employees from Trapit Bansal, another Openai researcher who worked on AI inference models at Zhao, and Openai’s Zurich office, who worked on multimodality.
Zuckerberg spent quite a bit of time making MSL a success. Meta CEO is recruiting staff from AI Superintelligence Lab. Meta reportedly provided 8- and 9-digit compensation packages to some researchers, some of which are “explosive offers” that expire in a few days.
TechCrunch Events
San Francisco
|
October 27th-29th, 2025
Meta has also stepped up its investment in cloud computing infrastructure. This should help MSL perform the large training runs necessary to create competitive frontier AI models.
By 2026, Zhao and MSL researchers should have access to Prometheus, a 1-gigawatt cloud computing cluster in Meta, located in Ohio. Online, Meta will be one of the first technology companies to have an AI training cluster of the size of Prometheus. One gigawatt is enough energy to enhance over 750,000 homes. This should help Meta carry out the large training runs necessary to create frontier AI models.
With the addition of Zhao, Meta currently has two chief AI scientists, including Yann Lecun, Meta’s Fairrabo leader. Unlike MSL, FAIR is designed to focus on long-term AI research. This may be used in 5-10 years. It is not yet known how exactly the three AI units from Meta will work together.
Nevertheless, Meta currently appears to have a formidable AI leadership team to compete with Openai and Google.
