What is Happening Inside Meta's AI Unit?

Meta is undergoing a profound transformation, rapidly reorganizing its artificial intelligence division to solidify its position as an "AI-first" company. This aggressive shift, as reported on June 14, 2026, involves a comprehensive restructuring that touches every aspect of Meta's AI operations, from leadership and team structures to data supply and infrastructure spending. The company's internal dynamics are being reshaped faster than its organizational capacity can comfortably absorb, indicating the urgency and scale of this initiative. At the core of this new structure is Meta Superintelligence Labs (MSL), which now oversees a layered organization. Beneath MSL, specialized groups are dedicated to various critical functions. These include frontier-model teams, the long-standing Fundamental AI Research (FAIR) group, product teams focused on integrating AI into Meta's offerings, and infrastructure teams responsible for compute and systems. A significant addition is the Applied AI unit, formed in March 2026, comprising approximately 6,500 individuals. This unit, representing about 8% of Meta's total workforce, underscores the immense human capital being deployed to support the company's AI ambitions. The strategic direction reflects a borrowing of the "frontier-lab playbook," similar to approaches seen at other leading AI organizations. However, Meta is applying this model within its unique corporate context, distinguishing itself through its unparalleled distribution network across platforms like WhatsApp, Instagram, Facebook, Messenger, and Threads. This extensive reach allows Meta to integrate AI capabilities into products used by billions of daily users, a significant advantage over many pure AI labs. Financially, Meta's commitment is substantial, with projected capital expenditures for 2026 ranging from $125 billion to $145 billion. This massive investment supports the development of reasoning models, multimodal AI, and the pursuit of frontier capabilities. The company has also deepened its relationship with data providers like Scale, indicating a robust data supply strategy, although it reportedly works with rival vendors as well, suggesting a focus on highly specific frontier-training data. Internally, this rapid transformation has not been without its challenges. There have been reports of employee backlash, particularly concerning new data collection practices, such as gathering computer-behavior data for AI agents. This highlights that while Meta is moving swiftly and spending heavily, the pursuit of superintelligence is also presenting significant organizational and cultural hurdles. Despite these challenges, Meta is actively shipping early models, such as Muse Spark, signaling tangible progress in its journey to become a dominant force in the AI landscape.