Meta Superintelligence Labs (MSL) is a new AI division of Meta Platforms, launched in June 2025, to develop next-generation “superintelligent” AI models that can learn and evolve with minimal human input. The unit is led by Alexandr Wang (USA), Meta’s Chief A.I. Officer.
He co‑founded Scale AI and brings deep expertise in data annotation and large language model evaluation. Co-leading product strategy is Nat Friedman (USA), former CEO of GitHub, with a focus on AI product development.
Another key figure is Shengjia Zhao (China), Chief Scientist of the lab. He previously led synthetic data research at OpenAI and has deep expertise in theoretical AI and reasoning systems.
The team also includes top hires, such as Pei Sun (China/USA), focusing on robotics and perception, and Johan Schalkwyk (South Africa), specialising in natural language processing and computational linguistics.
Meta has recruited around 50 elite scientists from rivals like OpenAI, Google DeepMind, and Anthropic, offering near‑nine‑figure compensation packages
Who Leads Meta’s Superintelligence Team, And What Are Their Roles And Backgrounds?
Alexandr Wang and Nat Friedman lead Meta’sco-founded Scale Superintelligence Team. Wang, from the USA, is Meta’s Chief AI Officer. He co‑founded Scale AI and sold a 49% stake to Meta for $14.3 billion. He dropped out of MIT and became a self-made billionaire by the age of 24.
Friedman, also from the United States, is a former CEO of GitHub. He co‑leads the Superintelligence Labs, focusing on products and applied research.
These leaders bring deep expertise. Wang excels in data labelling, model evaluation, and AI infrastructure. Friedman adds product strategy and applied AI skills.
Together, they unite Meta’s fundamental AI research (FAIR), product teams, Llama model creators, and a new next-generation model lab under a single mission.
Their arrival signals Meta’s significant commitment to building artificial general intelligence (AGI).
Who Are Some Key Team Members, And What Are Their Expertise, Countries, And Roles?
Meta’s Superintelligence Team is a global powerhouse, featuring top researchers from around the world.
The team is a melting pot of talent, with many hires being immigrants from India, China, South Africa, the UK, and Australia, bringing a rich diversity of perspectives and experiences.
Some notable names:
Trapit Bansal (India/USA): Former OpenAI principal engineer. Expert in meta‑learning, deep learning, and reasoning models. Education: IIT Kanpur (BSc/MSc), PhD at UMass Amherst.
Shengjia Zhao (China/USA): Co‑creator of ChatGPT and GPT‑4. Leads data synthesis and AI safety research.
All the team members, Ji Lin, Huiwen Chang, Hongyu Ren, Jack Rae, Joel Pobar, Pei Sun, Jiahuai Yu, and Johan Schalkwyk, are all immigrants with PhDs and former leads at OpenAI, DeepMind, Anthropic, or Google.
They bring a wealth of expertise in multimodal AI, large‑language models, image generation, speech recognition, perception, reasoning, and scalability, making them crucial for Meta’s AGI development.
How Much Investment And Compensation Is Meta Offering To Attract This Super‑intelligence Team?
Meta is investing massively—both in infrastructure and people. CEO Mark Zuckerberg pledged hundreds of billions of dollars for AI supercluster data centres. Projects like “Prometheus” are slated to launch in 2026, and “Hyperion” may require up to 5 GW of power—massive, even by national supercomputer standards.
On compensation, Meta offered record‑breaking packages:
- Matt Deitke, a 24‑year‑old AI prodigy, accepted a $250 million deal over four years.
- Other top researchers were offered nine-figure bonuses, sometimes $300 million over four years or $50–100 million as a sign-on.
- Meta lured talent from OpenAI, Apple, Google DeepMind, and Anthropic with these massive offers.
Meta’s strategy is bold. It bets on elite talent and heavy infrastructure to race ahead in building personal superintelligence.
Conclusion
Meta’s Superintelligence Team marks a bold leap forward for Meta. It unites leading AI researchers under one mission: to build superintelligent systems that empower individuals. This includes personal AI assistants, creative tools, and smarter platforms across Meta’s services.
The lab was launched in June 2025 and is led by Alexandr Wang and Nat Friedman, drawing talent from elite institutions like OpenAI, DeepMind, Apple, and Anthropic.
Meta is backing the effort with massive investments—ranging from high‑power data centres to nine‑figure compensation packages—to secure top talent and top compute power.
While the vision is grand, the road ahead has challenges. Meta must blend star‑studded talent into a cohesive team and build safe, useful AI, not just powerful AI. Still, the Superintelligence Team signals Meta’s serious intent to shape AI’s future.
