On June 30, 2025, Meta Platforms CEO Mark Zuckerberg unveiled a sweeping reorganization of the company’s artificial intelligence efforts, launching a new division called Meta Superintelligence Labs (MSL). This ambitious initiative aims to propel Meta to the forefront of AI technology by developing artificial general intelligence (AGI), systems capable of performing tasks as well or better than humans across a broad range of activities.
The MSL will consolidate Meta’s foundational AI model teams, product groups, and the well-regarded Fundamental AI Research (FAIR) division under one roof. At its helm will be Alexandr Wang, the former CEO of Scale AI—a data-labeling startup in which Meta invested a staggering $14.3 billion earlier this month. Wang assumes the role of chief AI officer, charged with steering the company’s aggressive push toward superintelligence.
Joining Wang as co-leader is Nat Friedman, the former CEO of GitHub, who will focus on AI products and applied research. Together, they will guide an elite team of AI researchers and engineers, many of whom were recently poached from rival tech giants like OpenAI, Anthropic, and Google DeepMind. Zuckerberg’s memo to staff, obtained by WIRED and reported by Bloomberg, highlights this talent influx as a key pillar of the new strategy.
The memo states, “We’re going to call our overall organization Meta Superintelligence Labs (MSL). This includes all of our foundations, product, and FAIR teams, as well as a new lab focused on developing the next generation of our models.” The new lab within MSL will particularly concentrate on advancing large language models (LLMs), the technology underpinning conversational AI and other complex applications.
This reorganization comes in the wake of Meta’s recent challenges in the AI arena. The company’s latest open-source model, Llama 4, received a lukewarm reception, while competitors such as Google, OpenAI, and China’s DeepSeek have been gaining momentum. Zuckerberg’s vision for MSL is clear: not just to catch up but to surpass rivals within the next year or so.
To accelerate progress, Zuckerberg has led an aggressive talent acquisition campaign. Reports indicate that Meta has dangled offers with signing bonuses reaching into the eight-figure range, including a disclosed $100 million bonus offered to some OpenAI employees, according to OpenAI CEO Sam Altman. Despite Altman’s confidence that his team would remain loyal, several key AI researchers have defected to Meta’s new superintelligence team.
Among the notable hires are Trapit Bansal, known for pioneering reinforcement learning on chain of thought and co-creating OpenAI’s o-series models; Shuchao Bi, co-creator of GPT-4o voice mode; Huiwen Chang, who contributed to GPT-4o’s image generation and invented novel text-to-image architectures at Google Research; and Jack Rae, who led early large language model efforts at DeepMind. The team also includes Joel Pobar, who previously spent over a decade at Meta, and several other experts with deep experience in AI research and development.
These new recruits are not just an infusion of talent but a strategic move to build a powerhouse capable of delivering breakthroughs in AI products and research. The proximity of their workspaces to Zuckerberg’s and Wang’s offices underscores the hands-on leadership approach being taken.
Meta’s investment in Scale AI and the hiring of its CEO reflect the company’s commitment to data quality and infrastructure, essential components for training advanced AI models. Additionally, Meta has engaged in talks to acquire other AI-focused startups, including Mira Murati’s Thinking Machines Lab and Ilya Sutskever’s Safe Superintelligence, though these discussions have not yet resulted in formal offers.
The company’s renewed AI focus also aims to generate new revenue streams. Zuckerberg envisions leveraging AI advancements to enhance Meta’s AI app, image-to-video advertising tools, and smart glasses, all of which could help offset the massive expenditures of Meta’s Reality Labs division, which has spent over $60 billion since 2020 with limited returns.
This AI arms race is not unique to Meta. Industry-wide, big tech companies are expected to spend an estimated $320 billion on AI in 2025 alone. Microsoft, for instance, invested $650 million in 2024 to acquire much of Inflection AI’s staff, while Amazon has aggressively recruited from competitors like Adept. Despite the vast sums invested, many experts believe the pinnacle of AI technology remains years away.
SoftBank’s Masayoshi Son, a noted tech investor, predicts a breakthrough in the next decade, underscoring the long-term nature of this competition. For Meta, the creation of MSL and the aggressive recruitment drive signal a renewed determination to be a leader in what may be the most transformative technology of the 21st century.
In a company-wide memo, Zuckerberg emphasized the urgency and ambition behind the move: the goal is to “get to the frontier” of AI development and to “overtake” competitors, not merely keep pace. With the Meta Superintelligence Labs now operational and led by some of the brightest minds in AI, the coming months will be critical in defining Meta’s role in the future of artificial intelligence.
