Meta Forms 'Superintelligence Group' to Pursue Artificial General Intelligence

Meta CEO Mark Zuckerberg is assembling a new team focused on achieving artificial general intelligence (AGI), amid internal dissatisfaction with the performance of its current AI offerings. The team, known internally as the superintelligence group, is part of a broader effort to enhance Meta's AI capabilities.

Zuckerberg has held recruitment meetings with AI experts in recent weeks at his residences in Lake Tahoe and Palo Alto. According to people familiar with the matter, he plans to personally recruit about 50 engineers and researchers, including a new head of AI research. Workspaces at the company's Menlo Park headquarters have been reorganized so that the team can sit close to Zuckerberg.

As part of the initiative, Meta is also preparing to invest billions in Scale AI, a company that provides training data and AI applications to enterprises and governments. The investment, first reported by Bloomberg News, would be Meta's largest external funding to date. Scale AI would be valued at $28 billion, according to The Information.

Scale AI's founder Alexandr Wang is expected to join Meta's team once the deal closes. Based on the new valuation, Wang's stake in Scale AI would exceed $5 billion, while co-founder Lucy Guo would also reach billionaire status. It remains unclear how the transaction would affect their ownership in the company. Meta and Scale AI declined to comment.

Zuckerberg's renewed focus on AI follows internal criticism of Llama 4, Meta's latest large language model. Released in April, the model did not meet expectations in usage or performance. Plans to launch a more advanced version, codenamed "Behemoth," were subsequently postponed over concerns it lacked meaningful improvements over earlier iterations, according to the Wall Street Journal.

The setback prompted Zuckerberg to take a more active role in recruitment and strategy. It has been reported that he has started a WhatsApp group chat called "Recruiting Party" with senior Meta staff to identify candidates. He is personally contacting prospects and maintaining direct communication throughout the hiring process, according to reports.

The superintelligence team is expected to drive future improvements in Meta's AI models and features, including personalization and voice tools. While the exact relationship between this unit and existing AI teams remains unclear, internal transfers are expected.

Meta has invested tens of billions of dollars in AI infrastructure, including capital for future data centers. Zuckerberg has told prospective hires that Meta has the financial resources to support multi-gigawatt data centers without needing external funding, citing the company's advertising business as a stable source of cash flow.

Zuckerberg has reportedly hosted private discussions over meals with researchers and engineers, pitching Meta's ability to fund long-term AI development independently of venture capital.

About the Author

John K. Waters is the editor in chief of a number of Converge360.com sites, with a focus on high-end development, AI and future tech. He's been writing about cutting-edge technologies and culture of Silicon Valley for more than two decades, and he's written more than a dozen books. He also co-scripted the documentary film Silicon Valley: A 100 Year Renaissance, which aired on PBS.  He can be reached at [email protected].

Featured

  • abstract metallic cubes and networking lines

    Call for Speakers Now Open for Tech Tactics in Education: Roadmap to AI Impact

    The virtual conference from the producers of Campus Technology and THE Journal will return on May 13, 2025, with a focus on emerging trends in with a focus on emerging trends in AI, cybersecurity, data, and ed tech.

  • Graduation cap resting on electronic circuit board

    Preparing Workplace-Ready Graduates in the Age of AI

    Artificial intelligence is transforming workplaces and emerging as an essential tool for employees across industries. The dilemma: Universities must ensure graduates are prepared to use AI in their daily lives without diluting the interpersonal, problem-solving, and decision-making skills that businesses rely on.

  • abstract coding

    Anthropic's New AI Model Targets Coding, Enterprise Work

    Anthropic has released Claude Opus 4.6, introducing a million-token context window and automated agent coordination features as the AI company seeks to expand beyond software development into broader enterprise applications.

  • globe surrounded by network connections

    AI Adoption Is Surging, but Infrastructure and Language Gaps Persist

    Artificial intelligence may be spreading faster than previous waves of consumer tech, but a report from Microsoft's AI Economy Institute suggests its benefits are concentrating in a relatively small set of countries, with infrastructure and language emerging as major dividing lines.