Meta is reportedly working on a new AI model called 'Avocado' and it might not be open source

Meta is developing a new AI model called "Avocado" that multiple outlets report may be proprietary rather than open-source, with a tentative 2026 release and being built inside the "TBD" group of its AI Superintelligence Labs led by Chief AI Officer Alexandr Wang. The move signals a strategic pivot from Meta's earlier open-source posture after delays and developer dissatisfaction with Llama 4, internal shake-ups including FAIR layoffs and the departure of Yann LeCun, and reflects CEO Mark Zuckerberg’s push to catch OpenAI and Google as Meta plans to invest heavily in AI (roughly $600 billion), a shift that could materially alter Meta's developer ecosystem and competitive positioning though the implications for Llama remain unclear.

Analysis

Meta is developing a new AI model called "Avocado," reported by CNBC and Bloomberg as potentially proprietary rather than open-source, with a tentative release timeline in 2026. The model is being built inside the "TBD" unit of Meta's AI Superintelligence Labs under Chief AI Officer Alexandr Wang, who reportedly favors closed models.

The reported pivot follows operational friction around Llama 4 — the "Behemoth" release has been delayed, developers have been reportedly unimpressed, and internal changes include layoffs at FAIR and the departure of chief AI scientist Yann LeCun. Mark Zuckerberg has publicly signaled a more selective open-source stance, citing safety and competitive urgency while committing large-scale investment to AI.

A move to a closed Avocado model would materially change Meta's developer ecosystem and distribution model for Llama derivatives and is intended to better position Meta against OpenAI and Google; Meta has stated an intent to spend roughly $600 billion on AI initiatives. Near-term investor risks are concentrated in execution (Avocado and Llama 4 quality/timing), developer adoption/backlash, and uncertain market reaction, which is reflected in mixed sentiment and a negative per-ticker sentiment score for META.

Key Decisions for Investors

• Reassess exposure to META and consider reducing position or hedging near-term given heightened execution risk on Avocado and ongoing Llama 4 delays
• Monitor concrete catalysts — official Avocado milestones, Llama 4 release quality, senior-exec moves, and capital allocation updates tied to the $600 billion plan — before increasing exposure
• If maintaining a long-term position, use staged sizing with downside protection because a proprietary pivot could improve monetization but risks developer adoption and reputational costs
• Compare relative exposure to OpenAI/Google-exposed strategies and be prepared to rotate if Meta's closed approach triggers developer flight or regulatory scrutiny