Meta CEO Mark Zuckerberg is on a bold mission to build the world’s leading Superintelligence Lab, and he’s sparing no expense. The company is offering unprecedented compensation packages—ranging from ₹800 crore to ₹1,600 crore—to attract elite AI researchers from tech giants like Apple, Google DeepMind, OpenAI, and Anthropic.
In what’s being dubbed one of Silicon Valley’s most aggressive talent poaching efforts, Meta is positioning itself as a major contender in the race to artificial general intelligence (AGI)—AI systems capable of human-like reasoning and learning.
Among the most headline-grabbing reports is a ₹1,600 crore (approx. $200 million) offer made to Ruoming Pang, a former AI researcher at Apple. Pang is believed to have accepted the offer and joined Meta’s Superintelligence Lab. This news comes on the heels of another massive offer—₹800 crore ($100 million)—to Trapit Bansal, a former OpenAI scientist involved in the development of the company's O-series AI models.
While Meta has not officially confirmed these compensation figures, they reportedly include base salaries, massive stock grants, signing bonuses, and multi-year vesting plans—often matching or exceeding CEO pay packages at global financial firms.
Meta’s Superintelligence Lab (MSL) was launched earlier this month as a focused initiative to lead in the AGI space. The lab is co-headed by:
Nat Friedman, former GitHub CEO
Alexandr Wang, former CEO of Scale AI
The lab has rapidly gained attention for its aggressive hiring strategy and its ambition to build a "dream team" of AI experts.
Meta’s poaching spree has already brought in at least 11 high-profile AI researchers, including:
Lucas Beyer, Xiaohua Zhai, Jack Rae, and Johan Schalkwyk from Google DeepMind
Ji Lin, Shengjia Zhao, and Jiahui Yu from OpenAI
Trapit Bansal (OpenAI)
Ruoming Pang (Apple)
These scientists are now part of the team working on Meta’s long-term AGI vision, alongside its flagship Llama series of large language models.
Zuckerberg is not just investing in talent—he’s also scaling up infrastructure. On Monday, he announced plans for a massive AI computing facility called “Prometheus”, described as an “AI data supercluster” that will go live by 2026.
In a detailed Facebook post, he outlined how Meta is constructing multiple “multi-gigawatt” clusters, including:
Prometheus
Hyperion, expected to scale up to 5 gigawatts in the coming years
These infrastructure projects are central to Meta’s mission of training next-generation AI systems at an industrial scale.
In a move reminiscent of Elon Musk’s Tesla “tent factory”, Meta is Normalalso building temporary data centres housed in tents. These prefab setups feature advanced cooling systems and are designed to help deploy models faster—even if they occasionally shut down due to heat stress.
This scrappy strategy underscores Meta’s urgency to accelerate its AI deployments and stay ahead in a highly competitive landscape.
According to Business Insider, Meta was reportedly underwhelmed by the reception to its Llama 4 AI model released earlier this year. In response, Zuckerberg has revised Meta’s AI roadmap, resulting in:
A $14 billion investment in Scale AI for high-quality training data
A relentless hiring spree, featuring multi-crore compensation packages
Infrastructure revamps like Prometheus and Hyperion
Together, these efforts aim to position Meta as a frontrunner in the development of advanced, scalable AGI.
Meta’s billion-dollar push into AGI is not just about models—it’s about assembling the smartest minds, building the biggest compute facilities, and moving faster than competitors. With leaders like Nat Friedman and Alexandr Wang, and a growing list of high-profile hires, Meta is betting big on the next frontier of AI.
Only time will tell whether this massive investment pays off—but one thing is clear: Zuckerberg is all in.