Zuckerberg Eyes Superintelligence: Meta’s Bold Bet to Catch Up in AI Race

Meta CEO Mark Zuckerberg is personally recruiting a high-powered team to spearhead the company’s bid for artificial superintelligence, according to multiple sources. This move marks a dramatic shift in strategy amid internal frustration over Meta’s recent AI efforts, including its Llama models and stalled “Behemoth” rollout.

Meta Founder and CEO Mark Zuckerberg speaks at LlamaCon 2025, an AI developer conference in April.

Building a Superintelligence Squad

Over the past week, Zuckerberg has met with potential recruits at his homes in Palo Alto and Lake Tahoe, targeting approximately 50 elite engineers and researchers to join the new “superintelligence” unit. These roles come with nine-figure compensation packages—an eye-catching incentive that underscores Meta’s renewed urgency to lead in AI.

Scale AI CEO Heads to Meta

A centerpiece of this push is the acquisition of a 49% stake in Scale AI for roughly $14.3 billion. As part of the deal, Scale CEO Alexandr Wang, according to Bloomberg and Reuters, will join Meta’s superintelligence team. Wang brings deep expertise in large-scale data labeling—critical for training advanced AI systems.

Responding to AI Missteps

Meta’s recent AI products, like Llama 4, have been met with lukewarm responses, prompting internal concern. Earlier rollout delays of the ambitious “Behemoth” model added to the urgency. Zuckerberg’s hands-on leadership signals a top-down attempt to reset Meta’s AI trajectory.

Recruitment Hurdles Ahead

Industry insiders caution that recruiting talent may prove challenging due to Meta’s internal culture issues, including high bureaucracy and divisive leadership dynamics. Still, the company’s vast GPU infrastructure and massive user base offer compelling long-term opportunities to apply cutting-edge models at scale.

A Risky Moonshot

Meta is reportedly investing between $10–15 billion into this initiative, including hardware, team-building, and data partnerships. While this ambitious investment rivals those of competitors like OpenAI and DeepMind, analysts suggest that real breakthroughs in artificial superintelligence (beyond narrow AI) remain years—or decades—away.