Meta is undergoing a radical shift in its AI strategy, creating uncertainty within the company. The company is focusing on developing a new frontier AI model codenamed “Avocado,” replacing its existing Llama models.
Meta Changes Strategy
According to the latest information, Avocado was initially planned for release by the end of this year, but is now targeted for launch in the first quarter of 2026. Performance tests are underway to ensure the system is well-received upon release. Meta states that training the model is progressing as planned.
For this new AI move, the company acquired Alexandr Wang, founder of Scale AI, and a team of leading engineers and researchers in June for $14.3 billion. This acquisition raised Meta’s capital expenditure estimate from $66–72 billion to $70–72 billion. Wall Street analysts expect the company to clarify its return on investment and strategic direction.
Meta’s digital advertising business remains strong, with annual revenues exceeding $160 billion and growth of over 20%. AI is enhancing the company’s efficiency by strengthening ad targeting.
However, CEO Mark Zuckerberg doesn’t want Meta to be solely dependent on advertising revenue, and the new AI team lacks experience in online advertising. Zuckerberg emphasizes that without significant AI initiatives, the company could fall behind in the industry.
Internal sources and analysts note that Meta’s strategy is still fragmented, reinforcing the perception that it is lagging behind the rapidly adopted models of top competitors like OpenAI, Google, and Anthropic. This raises a critical question for investors and industry observers regarding Meta’s future position in the AI space.
