Meta’s AI hiring spree has made global waves not only for the staggering salary figures being reported — as high as ₹1,600 crore ($200 million) — but also for what it reveals about the evolving dynamics of talent acquisition in the AI space. While Mark Zuckerberg acknowledged that salaries are indeed high, he clarified in a recent TITV podcast that media figures have been exaggerated and that compute resources — not compensation alone — are the primary magnet for top researchers.
Zuckerberg emphasized that AI scientists today prioritize access to powerful GPUs (like Nvidia’s H100s) over traditional corporate perks or large teams. He noted that researchers often say, “I want the fewest number of people reporting to me and the most GPUs,” underscoring a shift in focus from hierarchical control to computational capability. This preference aligns with the immense infrastructure demands of cutting-edge AI development.
Meta’s strategy is not just about poaching talent; it’s about assembling an elite team with both resources and autonomy. Some of the biggest names in AI — Trapit Bansal (ex-OpenAI), Ruoming Pang (ex-Apple), and several former researchers from Google and OpenAI — have joined Meta’s new Superintelligence Lab. Led by Nat Friedman (ex-GitHub CEO) and Alexandr Wang (ex-Scale AI CEO), the lab aims to develop artificial general intelligence (AGI) — AI that can reason and learn like humans.
To support this, Meta is making massive infrastructure investments. Its AI supercluster Prometheus, due online by 2026, will be a cornerstone, backed by new data centres like Hyperion, projected to scale to 5 gigawatts — among the largest planned AI compute setups in the world. To accelerate deployment, Meta has even adopted a Tesla-style approach: building temporary tent-based data centres to begin model training before full facilities are operational.
The urgency partly stems from internal pressure. Meta’s earlier Llama 4 release received a tepid response, pushing the company to rework its roadmap. A $14 billion investment in Scale AI for premium training data, along with this aggressive hiring and hardware push, signals Meta’s determination to catch up — or surpass — rivals like OpenAI, Google DeepMind, and Anthropic.
In essence, Meta’s AI offensive combines compute supremacy, elite talent, and rapid infrastructure deployment. It's not just a hiring spree — it’s a full-scale arms race to define the future of AGI.
