Key Takeaways:
- Meta has broken up its Meta Superintelligence Labs into four new groups focused on foundational models, research, products, and infrastructure.
- Alexandr Wang, Meta’s chief AI officer, will lead the new TBD Lab dedicated to building foundational AI models.
- The restructuring comes just months after MSL’s creation and follows significant investments in AI talent and infrastructure.
- Meta is weighing licensing third-party models while continuing development of its in-house Llama series.
- The move underscores rising internal tension and intense competition with OpenAI, Google DeepMind, and other rivals.
Meta has restructured its AI operations once again, dissolving the recently established Meta Superintelligence Labs (MSL) into four new groups designed to streamline research and product development. The decision reflects both organizational pressures and the escalating race to build advanced AI systems capable of surpassing human-level performance.
The four new units consist of a lab focused on foundational models, the longstanding Fundamental AI Research (FAIR) division, a products and applied research team, and an infrastructure-focused group. Each will be tasked with a separate mandate while contributing to Meta’s overarching goal of achieving what executives describe as “superintelligence.”
The new TBD Lab will be led by Alexandr Wang, the founder of Scale AI who joined Meta earlier this year as chief AI officer. According to reporting by the Times of India and TechCrunch, Wang’s group will focus on building and refining foundational models, including future iterations of Meta’s Llama series. FAIR, a research-oriented group with a history predating MSL, will continue to pursue long-term research projects. A products team headed by former GitHub CEO Nat Friedman will integrate AI into Meta’s consumer offerings such as Facebook, Instagram, and WhatsApp. Aparna Ramani, who has overseen infrastructure engineering at Meta, will manage the infrastructure division, ensuring data centers and supporting hardware keep pace with the growing computational demands.
While the restructuring has not yet triggered formal layoffs, insiders suggest the company may reduce staff in certain areas. Some employees will be reassigned to the new groups, while others may leave the company altogether. Meta is also exploring the option of licensing external models, both proprietary and open-source, to complement its in-house efforts. As reported by Gizmodo and Decrypt, this reflects a pragmatic recognition that Meta’s ambitions may require a mix of internally developed and third-party technologies.
This shift comes only months after Meta created MSL in June, with a mandate to drive the company’s most ambitious AI research. That effort included aggressive recruiting of talent from competitors like OpenAI and Google DeepMind, with some packages reportedly worth hundreds of millions of dollars. The dismantling of MSL so soon after its launch underscores both the urgency and the volatility in Meta’s AI strategy.
In an internal memo reported by Decrypt, Wang wrote that “superintelligence is coming,” and argued that organizing around research, product, and infrastructure is critical to making progress. The statement reflects Meta’s belief that building superintelligence requires not only advanced models but also tight alignment across product integration and technical infrastructure.
The reorganization highlights a central tension in Meta’s approach to AI. On one hand, the company has committed billions of dollars to compute resources, data center construction, and large-scale hiring. It has also sought to establish itself as a leader in open-source AI through the Llama series, which has been widely adopted by researchers and developers. On the other hand, Meta has faced criticism for lagging behind competitors in delivering polished consumer-facing AI products. Observers have drawn comparisons to Meta’s metaverse pivot, which saw heavy investment but limited user adoption.
Reports from outlets such as TechCrunch and the Times of India suggest that the reorganization was driven partly by internal tensions. Some employees questioned the concentration of power within MSL, while others worried about balancing long-term research against the pressure to ship consumer products. By dividing responsibilities across four groups, Meta hopes to reduce friction and accelerate progress.
The restructuring also highlights how Meta views its position relative to competitors. OpenAI and Google DeepMind remain highly visible leaders in the development of advanced models, while Anthropic and xAI are gaining traction with investors and users. Meta’s emphasis on superintelligence, combined with its ongoing infrastructure build-out, signals that it does not intend to concede leadership in the AI race.
Still, questions remain about execution. Breaking MSL into multiple groups may allow for more focused work, but it could also introduce new coordination challenges. The company must ensure that foundational research informs product development and that infrastructure can support increasingly large and complex models. Analysts note that while Meta’s open-source contributions through Llama have been well received, the company has yet to demonstrate consumer applications that rival the impact of ChatGPT or Gemini.
For Mark Zuckerberg, the restructuring is the latest demonstration of his personal commitment to AI. He has described AI as one of the company’s highest priorities, alongside its social networking platforms and augmented reality efforts. The decision to bring in Alexandr Wang and to split responsibilities across a new leadership structure reflects his willingness to recalibrate quickly in pursuit of results.
The stakes are high. Meta has already invested billions in AI and faces pressure from shareholders to demonstrate that the spending will yield competitive products and long-term value. If successful, the new organizational structure could allow Meta to deliver both cutting-edge models and practical consumer applications. If not, critics may see it as another costly pivot, echoing the company’s experience with the metaverse.
In conclusion, Meta’s decision to break up its superintelligence lab illustrates the turbulence of the current AI landscape. By creating specialized groups for models, research, products, and infrastructure, the company hopes to accelerate progress while addressing internal challenges. Whether this move positions Meta more effectively against its rivals remains uncertain, but it reinforces the central role AI now plays in the company’s identity and strategy.
Learn how AI Agents can supercharge your company’s profits and productivity at TMC’s AI Agent Event in Sept 29-30, 2025 in DC.
Rich Tehrani serves as CEO of TMC and chairman of ITEXPO #TECHSUPERSHOW Feb 10-12, 2026 and is CEO of RT Advisors and is a Registered Representative (investment banker) with and offering securities through Four Points Capital Partners LLC (Four Points) (Member FINRA/SIPC). He handles capital/debt raises as well as M&A. RT Advisors is not owned by Four Points.
The above is not an endorsement or recommendation to buy/sell any security or sector mentioned. No companies mentioned above are current or past clients of RT Advisors.
The views and opinions expressed above are those of the participants. While believed to be reliable, the information has not been independently verified for accuracy. Any broad, general statements made herein are provided for context only and should not be construed as exhaustive or universally applicable.
Portions of this article may have been developed with the assistance of artificial intelligence, which may have contributed to ideation, content generation, factual review, or editing.
