NextFin

Meta Restructures AI Division into Four Teams to Accelerate Superintelligence Pursuit

Summarized by NextFin AI
  • Meta Platforms Inc. has restructured its AI division into Meta Superintelligence Labs (MSL), focusing on developing superintelligence, which aims to surpass human intelligence.
  • The new structure includes four teams: TBD Lab for large language models, FAIR for foundational research, a Products and Applied Research team, and an Infrastructure team for computing resources.
  • CEO Mark Zuckerberg plans to invest up to $72 billion in 2025 for data center construction and talent acquisition to enhance AI capabilities.
  • Despite internal tensions regarding compensation and potential workforce downsizing, Zuckerberg remains committed to the vision of superintelligence transforming Meta's future.

NextFin news, Meta Platforms Inc. has reorganized its artificial intelligence division as of August 19-20, 2025, at its headquarters in Menlo Park, California, to accelerate the development of superintelligence, an AI system that could outperform human intelligence across nearly all domains.

The restructuring divides the AI group, now named Meta Superintelligence Labs (MSL), into four distinct teams: the TBD Lab, which oversees large language models including Meta's Llama series; the Fundamental AI Research (FAIR) lab, which continues foundational AI research; a Products and Applied Research team responsible for integrating AI technologies into consumer products; and an Infrastructure team focused on the computing resources necessary to train AI models.

This reorganization follows months of heavy investment and aggressive recruitment by Meta CEO Mark Zuckerberg, who has committed billions of dollars to attract top AI talent from competitors such as Apple, OpenAI, and startups like Scale AI. The company plans to spend up to $72 billion in 2025, primarily on data center construction and hiring researchers, to support its AI ambitions.

According to an internal memo obtained by Bloomberg and reports from the New York Times and other outlets, Alexandr Wang, former CEO of Scale AI, now leads the TBD Lab and serves as Meta's chief AI officer. The restructuring aims to unify research, product development, and infrastructure efforts under one umbrella to better focus on achieving superintelligence, a concept that goes beyond artificial general intelligence (AGI) by surpassing human intellectual capabilities in virtually all areas.

The move comes amid internal tensions, with some existing AI researchers reportedly dissatisfied due to disparities in compensation between new hires and current staff. Meta is also considering downsizing parts of its AI workforce, which currently numbers in the thousands, as it refines its strategy.

Meta's previous AI model, Behemoth, was abandoned for underperformance, prompting a fresh start with this new organizational structure. The company is also reportedly reconsidering its open-source AI stance, potentially moving toward closed-source models similar to OpenAI's approach.

Despite challenges, Zuckerberg remains committed to the vision that superintelligence will usher in a new era of individual empowerment and secure the future of Meta's digital empire.

Sources for this report include 01net, Bloomberg, The New York Times, Reuters, Gizmodo, TrendForce, India Today, and AInvest, with event dates anchored to August 19-20, 2025.

Explore more exclusive insights at nextfin.ai.

Insights

What is superintelligence and how does it differ from artificial general intelligence (AGI)?

What led to the recent restructuring of Meta's AI division?

How is Meta Superintelligence Labs (MSL) organized after the restructuring?

What are the main goals of the TBD Lab within MSL?

What investments is Meta planning to make in 2025 to support its AI initiatives?

What challenges is Meta facing with its current AI workforce?

How does Meta's approach to open-source AI differ from that of competitors like OpenAI?

What impact could the restructuring have on Meta's competition with other tech companies in AI?

What are the potential implications of superintelligence for society at large?

How might the internal tensions regarding compensation affect Meta's AI research output?

What historical examples exist of tech companies restructuring their AI divisions?

What feedback have users given about Meta's previous AI model, Behemoth?

What role does Alexandr Wang play in Meta's AI strategy following the restructuring?

How does Meta's investment in AI compare to that of its competitors?

What are the long-term goals of Meta's AI division?

What are the main criticisms surrounding Meta's AI development strategy?

Search
NextFinNextFin
NextFin.Al
No Noise, only Signal.
Open App