Meta is reorganizing its artificial intelligence division and splitting Meta Superintelligence Labs into four distinct teams focused on researching superintelligence products and infrastructure. This internal reorganization was announced in mid-August 2025 and reflects a strategic shift in how the company will run its AI efforts.
The company is now reallocating staff and refining roles to align talent with clear objectives. Executives who lead the AI programs will move into new responsibilities, and some leaders are departing as part of the change. Meta will also explore using third-party models to power some products while it continues in-house work where it adds value.
The company has invested a massive money in its AI capabilities this year. The company has spent billions to acquire data and talent, and a big investment was made into Scale AI earlier in 2025. These shifts increased the AI team to several thousand people and increased the cost base of the business. The new structure will be used to enhance focus and control of costs.
The company has put on hold most of its hiring of AI broadly as it fulfills the reorganization, reports note. It will enable Meta to overall review the areas where the new staff will be created, redeployed, or downsized. The company will continue to develop major functions of its products and launch new AI-related tools while it evaluates the staffing and budgets.
Industry observers interpret the shake-up as a bid to become even more focused on execution and establishing accountability in research and the delivery of products. The competition available to top talent will be high, and the firm will need to demonstrate constant progress in order to attract and maintain top researchers. The output that will be of interest to analysts includes; model releases, product integrations, as well as those of infrastructure performance.