Key Takeaways
- Meta’s upcoming AI model, internally called “Avocado,” won’t arrive until May 2026 at the earliest.
- The system’s capabilities currently rank somewhere between Google’s Gemini 2.5 and Gemini 3 benchmarks.
- The social media giant is exploring a short-term licensing agreement for Google’s Gemini AI technology.
- Meta’s most recent flagship AI release was Llama 3, which launched in 2024.
- This setback comes despite Meta’s projected $135 billion capital spending plan for 2026.
Meta’s highly anticipated “Avocado” artificial intelligence model has been pushed back to May following disappointing internal performance tests. Sources indicate the company is now exploring a licensing arrangement with Google for its Gemini platform.
Inside Meta’s AI Model Postponement
Meta Platforms has shifted the timeline for its next-generation AI system, dubbed “Avocado” internally, from an April launch to May at the earliest — with some estimates suggesting a June release.
The New York Times broke the story, drawing on insider information about the delay. Reuters independently verified that Avocado’s current capabilities place it somewhere between Google’s Gemini 2.5 and Gemini 3 models — falling short of Meta’s competitive targets.
Meta’s last major AI breakthrough came with the Llama 3 release in 2024. Meanwhile, competitors like Google’s Gemini, OpenAI’s GPT series, and Anthropic’s Claude have continued advancing their technologies.
The company did launch two streamlined Llama 4 variants last year, optimized for devices with constrained processing power. These releases were intended to precede a more robust model nicknamed “Behemoth” — though that project has gone quiet.
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.” They continued: “We’re excited for people to see what we’ve been cooking very soon.”
Meta stock experienced approximately 2% losses during Friday’s opening trading session.
Is a Google AI Partnership in Meta’s Future?
The situation takes an intriguing turn here. Senior figures within Meta’s artificial intelligence teams have reportedly floated the idea of securing a temporary license for Google’s Gemini platform to support the company’s offerings while Avocado development continues. While no final determination has been reached, these discussions are actively underway.
Such a move would represent a significant strategic reversal for CEO Mark Zuckerberg. He has committed substantial resources to his Superintelligence Labs initiative, assembling an elite team of AI researchers and engineers. The company’s 2026 capital expenditure forecast ranges between $115 billion and $135 billion, with substantial portions dedicated to AI infrastructure development.
A licensing deal with Gemini would essentially grant Alphabet a prominent victory — transforming its image from AI follower to sought-after provider whose technology competitors are willing to pay for.
Meta has already generated tangible value from AI applications — enhanced advertising precision, refined content curation algorithms, and proprietary silicon for recommendation engines. The company also maintains market leadership in smart eyewear.
However, Zuckerberg has openly pledged to create a “personal superintelligence” — a Meta-built platform capable of rivaling Apple and other mobile operating systems as the primary everyday computing interface. Achieving that vision demands a proprietary leading-edge model, not a borrowed solution.
For the present, the Avocado postponement remains the dominant narrative. Meta stock traded down approximately 2% in Friday’s early session.
