Menlo Park, July 25–26, 2025 — Meta Platforms has officially appointed Shengjia Zhao, a former lead scientist at OpenAI and one of the co‑creators of ChatGPT, as Chief Scientist of its newly launched Meta Superintelligence Labs (MSL). Zhao’s role places him at the heart of Meta’s bold initiative to develop AI systems that rival or surpass human-level intelligence.

A Central Figure in an Escalating AI Talent War

Meta’s decision to recruit Shengjia Zhao signals an escalation in the fierce competition for elite AI talent, particularly with OpenAI, Anthropic, and Google DeepMind. Zhao joined Meta after a period at OpenAI where he significantly contributed to the development of GPT‑4, GPT‑4.1, and lightweight mini models such as o3, and led the organization’s synthetic data research efforts.

In announcing Zhao’s appointment on Threads, CEO Mark Zuckerberg described him as a “pioneer” who would shape the lab’s research agenda and scientific direction, working closely alongside Alexandr Wang, Meta’s Chief AI Officer (formerly CEO of Scale AI). According to Zuckerberg, Zhao has been fundamentally involved since the lab’s inception and is now formally recognized as a co‑founder of MSL.

What is Meta Superintelligence Labs?

Established on June 30, 2025, Meta Superintelligence Labs consolidates multiple existing Meta AI research functions—including FAIR (Fundamental AI Research), LLaMA development teams, and product‑focused engineering groups—under a single umbrella aimed at pursuing artificial general intelligence (AGI) and beyond.

Leadership includes Alexandr Wang as Chief AI Officer, Nat Friedman (former GitHub CEO) co‑leading applied research and product development, with Daniel Gross (formerly of Safe Superintelligence) contributing on AI products.

Mission, Strategy, and Staffing

Meta has publicly committed to building “full general intelligence” and releasing its model research under an open-source ethos—a contrast to the more closed frameworks of OpenAI and Google DeepMind. The lab is also expected to focus on integrating multimodal capabilities—combining text, vision, voice, and video—into its foundational models.

To build out this capability, Meta has recruited an elite cohort of researchers from top institutions: OpenAI veterans like Trapit Bansal, Jiahui Yu, Shuchao Bi, Huiwen Chang, Ji Lin, Hongyu Ren, as well as former DeepMind and Anthropic engineers including Jack Rae, Pei Sun, and Joel Pobar.

Zhao’s Vision and Role

In his public post, Zhao said: “I am very excited to take up the role as chief scientist … Looking forward to building ASI [artificial superintelligence] and aligning it to empower people with the amazing team here. Let’s build!”.

As chief scientist, Zhao is expected to define MSL’s research agenda and scientific roadmap, guiding fundamental breakthroughs and ensuring that Meta’s pursuit of AGI aligns with broader societal and ethical goals. He will collaborate directly with Zuckerberg and Wang in a leadership triangle overseeing the lab’s trajectory.

Recruitment and Financial Commitment

Meta’s Superintelligence Labs launch has been backed by a staggering $14.3 billion investment for a 49% stake in Scale AI, ensuring access to cutting‑edge annotation infrastructure key for large‑model training. The company is rumored to have offered compensation packages in the range of tens to hundreds of millions of dollars to recruit top researchers. OpenAI CEO Sam Altman has criticized the strategy as overly pay‑driven, stating it could harm culture, though some recruited individuals have pushed back on the highest bonus claims.

Industry Reactions

The aggressive hiring campaign has drawn mixed opinions. OpenAI leaders have publicly expressed concern over the poaching strategy, with Altman calling Meta’s offers “crazy”, suggesting they prioritize compensation over mission alignment. Google DeepMind CEO Demis Hassabis commented that Meta’s efforts reflect urgency to catch up in the AI race, while analysts note the available talent pool capable of AGI-level breakthroughs may be fewer than 1,000 individuals globally.

Yet for Meta, the move reflects a bet on combining deep financial resources, open-source ethos, and a product-first culture to assert itself as a future leader in AI innovation.

What It Means for Meta and for AI

Meta’s Superintelligence Labs, with Zhao at its helm, aims to tackle the next frontier of AI research. The company envisions foundation models, multimodal systems, and potentially artificial superintelligence (ASI) that can operate at or above human-level cognitive performance across domains.

If successful, under Zhao’s scientific leadership, MSL could redefine AI’s trajectory—shifting from incremental improvements in narrow AI toward general intelligence platforms embedded across Meta’s ecosystem: social media, messaging, virtual reality, and enterprise tools.

However, the initiative carries risks. Critics point to Meta’s history with high-cost R&D arms like Reality Labs, which has burned over $60 billion with limited commercial returns. There’s also skepticism whether recruiting talent via massive financial incentives is sustainable or will produce durable innovation cultures.

Meta’s new unit may also reignite regulatory scrutiny around AI safety, data ethics, and the social impacts of releasing powerful, open-source models—especially given Zuckerberg’s public commitment to transparency and democratization of AI capabilities.