Prime Highlights:
- Mark Zuckerberg is creating a new “superintelligence” unit dedicated to building artificial general intelligence (AGI).
- Meta is in discussions to invest more than $10 billion in Scale AI, a big step in the AI race.
Key Facts
- The new AGI unit will comprise about 50 of the industry’s best engineers and researchers.
- Meta’s AI model “Behemoth” was delayed, and internally, Llama 4 underperformed.
- The firm will spend as much as $72 billion on next year’s capital expenditures, much of which goes toward AI.
Key Background
Meta CEO Mark Zuckerberg is increasing his focus on artificial intelligence by creating a new top-level research group working on AGI—Artificial General Intelligence. Informally, the “superintelligence” team, the organization will work toward advancing Meta’s AI far beyond existing generative models. Zuckerberg himself is personally leading the creation of this team, actively engaged in recruitment efforts and looking for a new AI research head.
This bold action follows as Meta also experiences internal discontent with its current AI development. The new Llama 4 model was said to have underperformed, and the next-generation “Behemoth” model has been pushed back on account of performance problems. In the meantime, Zuckerberg is centralizing initiatives, simplifying organizations, and focusing sharply on fundamental AI breakthroughs.
Meanwhile, Meta is reportedly in late-stage talks to put over $10 billion into Scale AI, a quick-expanding firm that sells high-quality training data to AI systems. As a part of the prospective agreement, Scale AI founder Alexandr Wang is likely to join Meta’s AGI effort. The transaction would potentially provide Meta with a nearly 49% stake in Scale AI, which is expected to reach $2 billion in revenue next year and is looking for about a $25 billion valuation. Scale AI has also made headlines for winning contracts with the U.S. Department of Defense.
Meta’s renewed emphasis on AGI is backed by significant investment. The firm is projecting 2025 capital spending of between $64 billion and $72 billion, with a significant share dedicated to AI infrastructure. This encompasses investments in bespoke chips, data centers, and AI-driven product development. The recent spate of AI-hiring and restructuring at Meta—followed by multiple rounds of layoffs—is all part of a larger effort to remain competitive in the rapidly accelerating AI race.
Zuckerberg’s initiative is matched by equal-sized investments by competitors. Microsoft has made big bets on OpenAI, Amazon has invested billions in Anthropic, and Google is continually increasing its AI outlays. With this new AGI effort and strategic investments, Meta is looking to take center stage in the next generation of AI innovation.