Meta has delayed the release of its AI model “Avocado” to at least May after performance concerns compared with Google’s Gemini models.
 |
| Meta pushes back the release of its Avocado AI model to May or later as it works to close the performance gap with rival systems like Google’s Gemini. Image: CH |
Tech Desk — March 15, 2026:
The decision by Meta Platforms to delay the launch of its artificial intelligence model Avocado highlights the increasingly intense race among major technology firms to dominate the next generation of AI systems.
Originally expected to launch in the first quarter of the year, the model’s release has reportedly been pushed back to at least May after internal testing showed its performance trailing behind competing systems developed by Google. According to reports, Avocado currently performs somewhere between Google’s Gemini 2.5 and Gemini 3—a gap significant enough to force Meta to reconsider the timing of its rollout.
The postponement underscores the rapidly shifting competitive landscape in generative AI, where performance improvements are measured in months rather than years. Major technology companies are racing to release models capable of more advanced reasoning, content generation and automation.
For Meta, the stakes are particularly high. Chief executive Mark Zuckerberg has repeatedly positioned artificial intelligence as central to the company’s future strategy, describing a long-term goal of developing systems that approach “superintelligence”—AI that could eventually surpass human capabilities in many cognitive tasks.
To support that ambition, Meta announced capital expenditure plans of between $115 billion and $135 billion this year, much of it aimed at expanding data centers, computing infrastructure and AI research.
Despite those investments, the delay suggests that Meta is still working to close the performance gap with competitors.
Google’s Gemini series has become one of the most visible benchmarks in the generative AI market, and its rapid development cycle has placed pressure on rivals to deliver comparable or superior models.
Reports that Meta executives have discussed the possibility of temporarily licensing Gemini technology to power some of its AI products—although no decision has been made—illustrate how competitive pressures could lead to unusual partnerships even among rival tech giants.
Such a move would highlight the complexity of the AI ecosystem, where companies may simultaneously compete and collaborate depending on technological capabilities and market demands.
Meta has downplayed the significance of the delay, emphasizing that the company is pursuing a broader roadmap of AI releases throughout the year.
Company representatives have said the next model is expected to demonstrate the company’s rapid progress, suggesting that Avocado may serve as one step in a larger sequence of upgrades rather than a definitive breakthrough.
At the same time, Meta is investing heavily in long-term technological independence, including plans to develop its own specialized AI chips. Building in-house hardware could eventually reduce reliance on external suppliers and allow the company to scale its AI infrastructure more efficiently.
The delay also reflects a broader challenge faced by many companies developing large AI models: achieving both technological leadership and commercial readiness.
As models become more powerful, expectations for performance, reliability and safety rise sharply. Releasing a system that appears weaker than competitors could undermine a company’s credibility in the market.
For that reason, Meta’s decision to postpone Avocado may reflect a strategic calculation that it is better to launch later with stronger capabilities than to release a model that fails to match the latest advances from rivals.
In an industry evolving at extraordinary speed, even a delay of a few months can significantly reshape the competitive balance. Whether Avocado ultimately strengthens Meta’s position in the AI race will depend on how effectively the company can translate its massive investments into technological breakthroughs.