Meta’s “Superintelligence” Lab: Inside the bold push toward AGI

Author

Eugen Rosenfeld
A CTO & a Solution Architect in Life Sciences at Nagarro. He has more than 20 years experience in different programming languages, technologies and business domains.

Meta Platforms (formerly Facebook) quietly initiated one of its most ambitious AI undertakings to date: a concerted effort to build artificial general intelligence (AGI) AI that can match or surpass human cognitive capabilities across domains. Internally dubbed the “superintelligence” initiative, this project reflects CEO Mark Zuckerberg’s renewed focus on long-term technological leadership. 

Reports emerged that Zuckerberg had begun assembling a top-tier AI team, personally overseeing recruitment and infrastructure. The aim has been to push Meta beyond current models, such as Llama, and leap toward AGI, or what insiders now refer to as "superintelligence." 

What does “Superintelligence” mean at Meta?

While Meta hasn’t formally defined the term, “superintelligence” is increasingly used by executives as shorthand for AGI. In public statements and internal communications, Zuckerberg and others use the two terms somewhat interchangeably. 

According to Reuters, Zuckerberg is targeting "machines that can match or surpass human capabilities.” Bloomberg similarly reported his ambition to “achieve what’s known as AGI.” The phrase "superintelligence" appears most frequently in strategic briefings, internal memos, and media leaks, but it has not been used to brand any external-facing product or platform. 

For now, “superintelligence” remains an internal rallying cry, signifying Meta’s next big leap in AI. When speaking formally, the company refers to this effort as part of its advanced AI or research division, rather than using the label publicly. 

The formation of Meta’s Superintelligence Lab 

The initiative gained structure in June 2025, when multiple sources reported that Zuckerberg was building a new AI research division referred to internally as the “superintelligence group.” This team comprises approximately 50 elite AI researchers, many of whom were directly recruited by Zuckerberg himself.

Key Characteristics: 

This R&D team is Meta’s answer to regaining AI leadership, aiming to create systems far more advanced than current industry benchmarks. 

Why now? 

Meta’s push toward superintelligence stems from a confluence of internal and external pressures: 

The strategic stakes of Superintelligence:
What Meta’s AI ambitions signal

If Meta’s superintelligence initiative is successful, it could reshape its product ecosystem and the entire AI landscape. Here’s a look at what could develop over the next ten years: 

Conscious engineering:
A cultural imperative

Ultimately, sustainability is not just about processes but also about people.

It’s about creating a culture where engineers, designers and product leaders are empowered to ask deeper questions:

This is conscious engineering. A mindset that sees limitations not as boundaries but as creative catalysts.

We encourage it through mentoring, measurement, and shared practice. And we have found it to be one of the most powerful forces for innovation. When sustainability becomes a cultural reflex, it permeates everything from backlog maintenance to sprint demos to production releases.

"Meta's 'superintelligence' initiative is not just about outperforming
human cognition; it is a strategic move to redefine the future of AI. As
nations and industries grapple with the implications of AGI, Meta's bold
move challenges us to rethink our approach to innovation, ethics and
global collaboration".

AGI outlook:
2030–2032 horizon