Mark Zuckerberg has already gone through an audacious pivot: rebranding Facebook as Meta and betting big on the Metaverse. It is now regarded as one of the most prominent technological blunders in recent history. Horizon Worlds, which was promoted as the future of human interaction, failed due to awkward VR gear, uninspiring content, and user disinterest. Zuckerberg was almost alone in supporting that vision, and as the Metaverse narrative collapsed, he bore the brunt of the derision. Now, with Meta Superintelligence Labs (MSL), Zuckerberg is trying an even bigger shift, one that may either cement his place among the great business visionaries or signal the decade’s greatest AI overreach.

This time, Zuckerberg isn’t alone. In 2025, he selected Alexandr Wang, the 28-year-old Scale AI creator, as Chief AI Officer. Wang now manages Meta’s AI section, which was reorganized into four independent labs:

1. TBD Lab — focuses on creating the next generation of AI superintelligence, to push AI beyond present technical limitations. The term “frontier-scale model training” most likely refers to their efforts to design and train extremely large-scale AI models on advanced computing platforms, which would be required to achieve superintelligence. This research entails overcoming significant engineering hurdles to train models of unprecedented scale and capacity, as well as pursuing “new directions” in AI research.

1. FAIR — its objective is to promote the state of the art in AI through open research that benefits everyone, with an emphasis on fundamental scientific advancements in machine learning. This involves investigating fundamental issues in reasoning, prediction, and unsupervised learning. The FAIR lab continues to prioritize long-term, fundamental research over product-specific AI applications.

1. Products & Applied Research — the team’s primary goal is to turn breakthrough AI models and research discoveries into practical, consumer-ready solutions. They prioritize improving user experiences by creating and deploying new AI-powered features and functions across Meta’s platforms.

1. MSL Infra — directly addresses the tremendous engineering issues of constructing and sustaining large-scale AI infrastructure. Responsible for constructing vast, multibillion-dollar data centers and training clusters needed to create its next-generation AI. It is vital to ensure dependability and minimize training interruptions. MSL Infra must provide solid methods for managing faults and ensuring consistent training performance over millions of GPU hours.

On paper, the combination appears formidable. Zuckerberg, who once connected billions of people through social networks, now wants to understand everyday life. Wang, the AI infrastructure wunderkind, is in charge of designing and operating superintelligence engines. However, youth and talent do not ensure long-term success. Vision needs more than ambition; it necessitates governance, discipline, and the capacity to translate innovations into sustainable markets. For the worldwide tech community, this distinction is essential. A second visionary is not an insurance plan. It is a risk that the synergy between Zuckerberg’s cultural reach and Wang’s technological expertise could result in commercial supremacy, not merely flashy headlines.

Meta reported earnings of $7.14 per share on $47.5 billion in revenue for the quarter ending June 30. Earnings per share increased 38% from the previous year and were much higher than Wall Street analysts’ expectations of $5.88. It also anticipated sales for the current quarter to be between $47.5 billion and $50.5 billion, above analysts’ projections. Meta’s stock rose more than 9% in after-hours trading on the impressive results. The company’s stock has climbed 16% since the beginning of the year.

The claim comes after Meta CEO Mark Zuckerberg outlined his strategy for AI “superintelligence” in a video and blog post. According to his blog post, he wants everyone to have access to a personal AI superintelligence, which would increase productivity and allow them to spend “more time creating and connecting.”

Meta has been spending heavily to lure top AI talent away from competitors like OpenAI, Google, and Apple for its new Meta Superintelligence Labs team. The corporation is also spending tens of billions of dollars to construct enormous AI data centers. Meta Chief Financial Officer Susan Li stated that recruiting in “high priority” sectors such as AI is projected to increase the company’s overall workforce this year and next. She said that rising pay from Meta’s investments in elite AI talent will be the second-largest driver of spending growth next year.

Meta is collaborating with tech titans such as OpenAI, Google, and Anthropic to achieve superintelligence, the theoretical point at which AI outperforms humans in all areas of knowledge work. It is expected that achieving that milestone would drastically transform the economy and the way people work, potentially opening up major new commercial opportunities for organizations that can deliver the technology.

The stakes may be especially high for Zuckerberg, who wants Meta to be more than simply a social media company and has refocused it on AI following an aborted foray into the Metaverse. The corporation is under pressure to deliver on the billions of dollars it has spent on data centers and semiconductors, as well as a developing smart glasses business that depends on the success of its AI initiatives. And the corporation is lagging behind competitors, having reported delays in deploying the biggest version of its new Llama 4 AI model.

Despite its aggressive spending, Meta said on Wednesday that its capital expenditures in the second quarter totaled $17 billion, virtually matching Wall Street’s expectation of $16.48 billion. It also reduced but did not raise its full-year capital expenditure projection, offering investors a clearer picture of its spending strategy.

The trajectory of MSL cannot be judged solely on a distant horizon. Evidence of success or failure will emerge over the following 12–18 months. Meta’s ability to incorporate agentic intelligence into WhatsApp, Instagram, or Horizon, as well as monetize experiences beyond advertising, will provide early signs. In other words, the next year will reveal Meta’s ability to execute, but the following decade will determine if it can dominate.