Meta has intensified its pursuit of AI dominance by recruiting Shengjia Zhao, a key contributor to OpenAI’s GPT-4 and former lead scientist behind synthetic data development.
Zhao has joined Meta as chief scientist of its newly formed Superintelligence Labs, a division tasked with foundational AI research and next-generation model development.
Meta CEO Mark Zuckerberg, who announced Zhao’s appointment on Friday, described him as a “pioneer” in artificial intelligence, crediting him for several groundbreaking achievements in the field. Zhao’s recruitment is seen as a major coup in Silicon Valley’s ongoing scramble for elite AI minds, and he will work closely with Zuckerberg and Alexandr Wang, Meta’s new Chief AI Officer and the founder of Scale AI.
The hire follows Meta’s recent $15 billion investment in Scale AI and reflects the company’s massive bet on building artificial general intelligence (AGI). Beyond Zhao, Meta has also quietly absorbed several key figures from OpenAI’s Zurich office — Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai — who had previously worked at Google DeepMind. Together, they form a formidable team under Meta’s Superintelligence Labs, signaling the company’s intent to close the frontier gap between itself and leaders like OpenAI and Anthropic.
Industry insiders view these developments as part of a broader, heated battle over the handful of experts capable of designing and scaling frontier AI models. Databricks VP Naveen Rao likened the competition to “looking for LeBron James,” noting that fewer than 1,000 people globally possess the expertise required to build such advanced systems.
This elite scarcity has driven tech giants to adopt increasingly aggressive tactics. Perplexity CEO Aravind Srinivas recently revealed that a Meta researcher turned down a job offer by saying, “Come back when you have 10,000 H100s” — a nod to the specialized Nvidia chips crucial for training massive models.
Zuckerberg, according to Business Insider, has been personally courting talent, sending direct emails and even inviting AI researchers to private dinners at his home. Meanwhile, OpenAI’s Sam Altman has made his own calls to staff in a bid to curb the exodus. During a July 18 episode of the podcast “Uncapped with Jack Altman,” Altman accused Meta of offering “giant” salaries and stock packages to lure away employees, calling the strategy “crazy.”
“The degree to which they’re focusing on money and not the work and not the mission,” Altman said, “I don’t think that’s going to set up a great culture.”
Demis Hassabis, CEO of Google DeepMind, echoed that skepticism during a Friday appearance on the Lex Fridman Podcast.
“Meta right now are not at the frontier — maybe they’ll manage to get back on there,” he said. “It’s probably rational what they’re doing from their perspective because they’re behind, and they need to do something.”
Despite the skepticism, Meta’s moves appear strategic. With Zhao’s arrival and a reinforced Superintelligence Labs team, the company is attempting to recalibrate its AI trajectory, particularly after lagging behind in recent model releases. Its LLaMA models, though powerful, have struggled to keep pace with OpenAI’s GPT and Anthropic’s Claude series in benchmark evaluations.
The aggressive recruitment spree also comes as Meta navigates increasing pressure to show tangible returns on its AI investments. The company is aiming to make AI the cornerstone of its broader metaverse and product strategy, integrating generative tools across its platforms, including Instagram, WhatsApp, and Facebook.
While Meta and OpenAI both declined to comment on the growing talent war, the high-profile defections and the billion-dollar bids underscore a critical truth in today’s AI race: the future belongs to the labs that can attract — and retain — the very few who know how to build it.
