Meta's AI Overhaul: Superintelligence Ambitions Spark Internal Divisions

Meta Platforms Inc. has temporarily halted hiring for its artificial intelligence division following a period of aggressive recruitment that included high-profile acquisitions and substantial financial outlays. The company confirmed the hiring freeze to CNBC, describing it as part of “organizational planning” for its new superintelligence initiatives. The move, first reported by the Wall Street Journal, comes as the firm restructures its AI operations into four distinct teams: a superintelligence research group, an AI products team, an infrastructure division, and a long-term exploration division. These teams fall under the umbrella of Meta Superintelligence Labs, a name reflecting CEO Mark Zuckerberg’s vision for AI systems capable of outperforming the smartest humans in cognitive tasks [1].

The restructuring is part of a broader effort to streamline operations after rapid expansion in AI talent acquisition. In one of its most significant moves, Meta spent $14.3 billion to acquire a 49% stake in Scale AI, bringing on its founder, Alexandr Wang, who now leads the company’s Llama series of open-source large language models. In recent months, Meta has offered nine-figure compensation packages to attract AI researchers from competitors like OpenAI and Google, sparking a high-stakes recruitment war in Silicon Valley [1].

The reorganization has also led to internal tension, particularly between long-time Meta AI researchers and newly hired executives. For instance, Shengjia Zhao, co-creator of OpenAI’s ChatGPT, was recently appointed as Meta’s chief AI scientist and has reportedly questioned the work of existing team members. Meanwhile, several veteran researchers, including Joelle Pineau, Angela Fan, and Loredana Crisan, have left the company to join AI startups or other tech firms [3]. The shift in leadership and strategy has also seen some of Meta’s open-source approach to AI models being reconsidered, with new teams opting to pursue closed-source development for its next generation of AI systems.

Zuckerberg emphasized the company’s commitment to AI during an investor call, describing it as a driver of “individual empowerment” and a key to enhancing Meta’s core advertising business. The company has budgeted up to $72 billion in capital expenditures for 2025, with the majority allocated to data center construction and AI hiring [3]. Analysts suggest that the reorganization reflects a strategic pivot to consolidate research and product development under a more focused structure, aiming to accelerate the development of superintelligence and compete with industry leaders.

Despite the hiring freeze, Meta’s long-term investment in AI remains substantial. The company is also exploring the use of third-party AI models, potentially including open-source and closed-source options from other firms. This shift marks a departure from its prior strategy of relying solely on in-house AI development. Whether these changes will yield competitive AI products remains to be seen, but the reorganization underscores the high-stakes nature of the global AI race [3].

Source:

[1] Meta Brakes Massive AI Talent Recruitment Spending Spree, Mark Zuckerberg TBD Superintelligence Lab (https://www.cnbc.com/2025/08/21/meta-brakes-massive-ai-talent-recruitment-spree-mark-zuckerberg-tbd-superintelligence-lab.html)

[2] Meta Freezes AI Hiring, WSJ Reports (https://www.reuters.com/business/meta-freezes-ai-hiring-wsj-reports-2025-08-21/)

[3] Zuckerberg Again Overhauls Meta's A.I. Efforts (https://www.nytimes.com/2025/08/19/technology/mark-zuckerberg-meta-ai.html)