Meta Platforms is reportedly delaying the release of its next-generation artificial intelligence (AI) model, Avocado, to the first quarter of 2026, according to sources cited by CNBC. The move accompanies a strategic shift from open-source to closed-source AI development and a major restructuring of Meta's AI division.
The new Avocado model, originally planned for launch by the end of 2025, will be developed under tighter control by the TBD Lab within Meta Superintelligence Labs (MSL), led by Chief AI Officer Alexandr Wang. This follows a lukewarm market response to Meta's recent open-source Llama 4 series and diminished interest from the developer community. The transition marks a departure from Meta's earlier open-source ambitions, as concerns grew over the architectural exposure of Llama models, including their utilization by Chinese startup DeepSeek.
The report points to dissatisfaction within Meta about the risks associated with open-source projects, prompting a reorientation toward more proprietary AI frameworks. Internal sources told CNBC that this shift influenced leadership changes, most notably the exit of longtime product chief Chris Cox from the AI leadership team. In July 2025, The New York Times highlighted discussions among Wang and team members about moving resources away from open-source models, and CEO Mark Zuckerberg reportedly confirmed that Meta would not release models capable of superintelligence as open-source.
Meta's AI division has faced ongoing challenges, including frequent restructuring, downsizing, and extended overtime. An October 2025 wave of layoffs targeting the Fundamental AI Research (FAIR) lab contributed to the resignation of prominent Meta scientist Yann LeCun. According to The Wall Street Journal, internal tensions have intensified due to pay disparities, bottlenecks in computing resource allocation, and a recruitment process requiring Wang's direct approval, complicating interdepartmental collaboration.
Previously known for fostering cross-department collaboration, Meta's AI division now emphasizes faster development cycles and more centralized communication. The company's recalibrated strategy and organizational changes reflect its effort to better control AI development amidst rising competition and growing complexities in the AI landscape.
Article edited by Jerry Chen
