• Source:JND
Add as a preferred

As companies like Google, OpenAI, Microsoft, and Musk’s xAI seem to be making rapid strides in AI seemingly on a daily basis, Meta has appeared conspicuously quiet. But behind the scenes, Mark Zuckerberg’s company is undergoing a major internal AI shakeup. In a prototype memo that seems to be cut from the same cloth as the rest of the pre-release intelligence within Meta, Zeta AI, a subdivision of Meta, plans to partition the AI division into an immediate product execution team. This team will also include an AGI long-term research team.

Two New AI Divisions at Meta: Products and AGI Foundations

The reorganisation will divide Meta’s AI efforts into two major teams:

  • AI Products Team: This division will be led by Connor Hayes, and it consists of all consumer-centric experiences. Among them are the Meta AI assistant and AI features on Facebook, Instagram, WhatsApp, Messenger, and the AI Studio tools of Meta.

  • AGI Foundations Team: Focus on the multidisciplinary engineering and the reasoning for Llama II models, the Llama series of models developed by Meta AI, is to be co-led by Ahmad Al-Dahle and Amir Frenkel. They will also lead work on voice and video intelligence and multimedia AI.

    • Meta’s Fundamental AI Research (FAIR) group will still function as before, but with no multimedia team–that moves to AGI Foundations.

    Focused Teams to Drive Innovation

    Meta aims to encourage the ownership and reduce reliance pathways as stated by Chief Product Officer Chris Cox. This is one of the reasoning goals for the reorganization of AI initiatives into smaller, more focused driven teams that Meta believes will accelerate innovation and draw in new technical talent.

    Regardless of the major overhaul, Meta has clarified that there are no executive layoffs. Instead, the meta shift is simply internal role shifts and talent reassignments, not cuts.

    Talent Drain: Key AI Researchers Exit Meta

    Meta’s AI division appears to be facing significant headcount losses, alongside the latter reorganization. A number of the former Llama model researchers have reportedly left the company, many of them moving to rapidly rising competitors like Mistral AI, according to Business Insider.

    The most high-profile executive exit Joelle Pineau who led Meta’s FAIR group. Pineau will officially exit the company on May 30, 2025 — a date that stands out as a lose for Meta’s research leadership.

    Meta’s Push to Catch Up in the AI Race

    The reorganization of the latter indicates a desperate need to expedite the gap to competitors such as OpenAI and Google, the latter having recently announced powerful new models like Gemini 1.5 and Veo 3 from Google I/O reinforcing generative AI.

    On the other hand, Meta is working on its own consumer AI projects. The company has rolled out a standalone AI assistant app, which it first advertised during LlamaCon 2025. This follows the integration of Meta AI into its flagship applications. The goal of this app is to provide users with access to Llama's features similar to ChatGPT, situated outside of Meta's social media ecosystems.

    Final Thoughts: Meta’s Next AI Chapter

    Even though the general public may consider that Meta did not make a significant impact during the “AI boom,” this reorganization indicates a clearly defined strategy. We believe that by compartmentalising the practical application of AI into one unit and keeping deep, foundational AI research separate, Meta intends to seriously compete in the AI race.

    Those fundamental shifts would be welcome – but with core talent disengaging and competition gaining pace, time is running out on Zuckerberg’s AI vision.