Meta's Mega AI Push: Ex-OpenAI and Apple Researchers Join Superintelligence Lab with Rs 1600 Cr Offers

In a bold move that’s shaking up Silicon Valley, Meta has gone all in on artificial intelligence — not just with tech, but with talent. CEO Mark Zuckerberg has launched an all-out recruitment offensive to build a dream team under the newly formed Meta Superintelligence Lab (MSL), offering compensation packages that rival the highest in corporate history.

Among the eye-catching figures is the reported Rs 1,600 crore (around $200 million) package extended to Ruoming Pang, a top AI researcher from Apple, who is now part of Meta’s ambitious AGI efforts. This follows the Rs 800 crore ($100 million) offer reportedly accepted by Trapit Bansal, an ex-OpenAI scientist and one of the key contributors to the company’s o-series models. These staggering offers are said to include salaries, signing bonuses, long-term stock grants, and vesting options, placing them on par with — or even above — CEO compensation at top global banks.

While Meta hasn’t officially confirmed the figures, the hiring spree is real. The Meta Superintelligence Lab, launched earlier this month, is led by former GitHub CEO Nat Friedman and ex-Scale AI CEO Alexandr Wang. Since its inception, the lab has already brought in at least 11 elite researchers, poached from the likes of Google DeepMind, Apple, Anthropic, and OpenAI. High-profile names include Lucas Beyer, Xiaohua Zhai, Jack Rae, and Johan Schalkwyk from Google, along with Ji Lin, Shengjia Zhao, and Jiahui Yu from OpenAI.

These strategic hires are central to Meta’s larger vision of developing Artificial General Intelligence (AGI) — AI systems capable of human-level cognition, reasoning, and learning. The company’s renewed focus on AGI comes after what Zuckerberg reportedly considered an underwhelming response to Meta’s Llama 4 model earlier this year.

But Meta’s investment in the future of AI doesn’t stop at talent. On Monday, Zuckerberg revealed a sweeping plan to build one of the world’s largest AI infrastructures. At its core is a new “AI data supercluster” called Prometheus, expected to go live by 2026. Prometheus is just one of several “multi-gigawatt” compute clusters in development, alongside another major system named Hyperion, which will eventually scale up to 5 gigawatts.

To speed up deployment, Meta is even building temporary AI data centers in industrial tents, a move reminiscent of Elon Musk’s tent-based Tesla production during the Model 3 era. These makeshift facilities, fitted with prefabricated cooling systems, are designed to get new models online quickly — even if that means temporary shutdowns during peak summer heat.

The company is also pouring $14 billion into Scale AI to ensure access to premium training datasets — a crucial component for advancing AI models toward AGI.

Meta’s unprecedented salary packages, infrastructure expansion, and elite hires send a clear signal: Zuckerberg is serious about leading the next frontier of AI.