Meta Platforms plans another overhaul of its artificial-intelligence organisation, splitting the newly formed Meta Superintelligence Labs into four distinct groups covering research, superintelligence models, product integration and infrastructure, according to an internal memo and people familiar with the matter. The change is expected to be announced Tuesday and was outlined by Chief AI Officer Alexandr Wang, who joined in June after Meta bought a 49% stake in his data-labeling firm Scale AI for about $14 billion. The reorganisation aims to accelerate the company’s push toward so-called superintelligence and streamline a division that has ballooned to thousands of employees following an aggressive hiring spree that included nine-figure pay packages for top researchers. While no layoffs are planned immediately, Meta is considering trimming headcount in the AI group and some senior executives are expected to leave, the people said. As part of the shake-up, Meta is also weighing the use of third-party large language models for some products, a shift from its previous strategy of relying solely on in-house technology such as the Llama series. Chief Executive Officer Mark Zuckerberg has signalled that AI spending could lift total capital expenditures to as much as $72 billion this year as the company races to catch OpenAI and Google in building more capable systems.