Meta Platforms has quietly restructured its AI division, carving it into two specialized teams in a bid to accelerate product delivery and deepen its long-term research capabilities.
The move is reportedly also meant to directly compete with rivals like Google and OpenAI at a time when the company is facing stiff competition even from the Chinese social media platform TikTok.
The split is to enhance efficiency for Meta’s AI teams
In a memo circulated internally by Chief Product Officer Chris Cox, the AI arm was reconstituted into an AI Products Team and an AGI Foundations Team, a move designed to sharpen each group’s focus and minimize cross-team dependencies.
Rather than trimming headcount, Meta has simply realigned existing talent – the AI Products Team, now under the leadership of Connor Hayes, will be charged with embedding and refining AI features that users interact with directly.
This includes personalized recommendation engines, smart assistants across Facebook, Instagram, and WhatsApp, as well as the company’s standalone AI application, an interface akin to other leading generative AI platforms. By concentrating on consumer-facing tools, this cohort aims to make everyday social-media experiences richer and more competitive.
Concurrently, the AGI Foundations Team, co-headed by Ahmad Al-Dahle and Amir Frenkel, will pursue Meta’s loftier ambitions of artificial general intelligence (AGI), overseeing open-source large language models like Llama and pushing advances in reasoning, multimedia processing, and voice synthesis.
This split aligns Meta more closely with the organizational models adopted by rivals such as Google, Anthropic, and OpenAI, where product engineering and fundamental research are often housed in separate siloes to prevent either from slowing the other down.
While Meta’s Fundamental AI Research group (FAIR) will continue to operate autonomously, a FAIR sub-team focused on multimedia capabilities has been reassigned to the AGI Foundations unit, symbolizing the renewed emphasis on the fusion of research and future-proof development.
Is Meta lagging behind in the AI race?
According to insiders, the restructuring is not a reaction to imminent layoffs; rather, it reflects CEO Mark Zuckerberg’s strategy to bolster both short-term innovation cycles and long-term technological leadership without sacrificing staff or morale.
Late-April departures, most notably that of Joelle Pineau, the veteran head of FAIR who announced her exit effective May 30, 2025, underscore the pressure on Meta to retain top AI talent even as competitors like Mistral and Anthropic aggressively recruit from its ranks.
Meta’s pivot comes amid growing perceptions that it lags behind in the high-stakes AI race. Tech heavyweights such as Google have unveiled market-first models like Gemini and the visual-analysis VEO3, while OpenAI’s ChatGPT continues to expand its presence across industries.
In response, Meta has already woven its AI assistant into core services, from messaging to content creation, and recently launched an independent app at its LlamaCon conference, granting users standalone access to its generative engine. The freshly minted AI Products Team will steward this application alongside in-platform integrations, with the goal of outpacing rivals on usability and engagement.
At the same time, the AGI Foundations Team will refine the underlying Llama architecture, build next-generation large language and multimodal models, and experiment with advanced reasoning algorithms and voice-interface technologies that could one day underpin broader Meta initiatives.
This two-pronged approach reflects a carefully calibrated growth strategy. Frontline engineers and designers sprint to deliver feature updates and consumer hooks while deep-research scientists methodically tackle the complex challenges of true general intelligence.
“Our new structure aims to give each organization greater ownership and make explicit any remaining dependencies.” Cox.
Cox emphasized that the reallocation of personnel is intended to foster accountability and speed.
By isolating research goals from product roadmaps, Meta hopes to avoid dilution of effort, a chronic risk when the same group juggles both exploratory science and market-driven deliverables.
Furthermore, the company has assured employees that no executive positions will be cut and that the reshuffle merely redistributes existing roles rather than shrinking the workforce, a reassurance aimed at preserving institutional knowledge and reducing anxiety during the transition.
Cryptopolitan Academy: Coming Soon - A New Way to Earn Passive Income with DeFi in 2025. Learn More