TLDR
- Meta’s upcoming flagship AI model, internally called “Avocado,” won’t launch until May 2026 at the earliest.
- Internal benchmarks show the model performing somewhere between Google’s Gemini 2.5 and Gemini 3 versions.
- The social media giant is exploring a temporary agreement to license Google’s Gemini AI technology.
- Meta’s most recent frontier AI model launch was Llama 3, which debuted in 2024.
- This setback arrives as Meta prepares to invest up to $135 billion in capital expenditures throughout 2026.
Meta’s highly anticipated “Avocado” AI system has been postponed to May following disappointing internal evaluations. Sources indicate the company is now exploring a short-term licensing arrangement with Google for its Gemini platform.
What’s Happening With Meta’s Upcoming AI System?
Meta Platforms has postponed the launch of its next-generation AI model, known internally as “Avocado,” shifting the timeline from this month to May at the earliest — with a potential extension to June.
The New York Times broke the story, drawing on information from people with direct knowledge of the situation. Reuters independently verified that internal testing shows the model’s capabilities landing between Google’s Gemini 2.5 and Gemini 3 — essentially missing Meta’s performance targets.
Meta’s last frontier-pushing AI release was Llama 3, which launched in 2024. In the interim, competitors including Google’s Gemini, OpenAI’s GPT series, and Anthropic’s Claude have advanced their positions in the AI race.
The company did introduce two scaled-down Llama 4 variants last year, optimized for resource-constrained environments. These lighter models were intended as precursors to a more robust system dubbed “Behemoth” — though updates on that project have been notably absent.
A Meta representative stated: “As we’ve said publicly, our next model will be good, but more importantly, show the rapid trajectory we’re on, and then we’ll steadily push the frontier over the course of the year as we continue to release new models.” The representative continued: “We’re excited for people to see what we’ve been cooking very soon.”
Meta stock declined approximately 2% during early Friday trading sessions.
Is Meta Considering a Google AI Partnership?
The situation takes an intriguing turn with reports that Meta’s AI leadership team has floated the idea of temporarily licensing Google’s Gemini technology to support its product ecosystem while Avocado undergoes further development. While no final determination has been reached, these discussions are actively underway.
Such an arrangement would represent a significant setback for CEO Mark Zuckerberg’s vision. He has committed substantial resources to his Superintelligence Labs operation, assembling an elite team of AI researchers and engineers. The company’s capital expenditure roadmap for 2026 ranges from $115 billion to $135 billion, with AI infrastructure representing a major component.
A Gemini licensing agreement would effectively validate Alphabet’s AI capabilities — transforming its narrative from AI challenger to preferred provider for industry peers.
Meta has already demonstrated tangible AI benefits across its platforms — enhanced advertising precision, superior content curation algorithms, and proprietary silicon for recommendation systems. The company also commands a leading position in the smart glasses segment.
However, Zuckerberg has made public commitments to developing a “personal superintelligence” — a Meta-built platform that could ultimately rival Apple and competing smartphone ecosystems as the primary personal computing interface. Achieving this ambitious objective demands proprietary leading-edge models, not licensed alternatives.
For the moment, the Avocado postponement remains the focal point. Meta stock traded down approximately 2% in Friday’s early session.
