Meta has outlined a long-term AI strategy focused on significant investments rather than immediate profits. During the company’s Q2 earnings call, CEO Mark Zuckerberg shared Meta’s future vision and highlighted the need for substantial computational resources to support its AI initiatives.
Zuckerberg revealed that Meta is “planning for the compute clusters and data we’ll need for the next few years,” with a particular emphasis on their upcoming AI model, Llama 4.
Meta expects that training Llama 4 will require “nearly 10x more” computing power than its predecessor, Llama 3, which reportedly used 16,000 GPUs. Zuckerberg aims for Llama 4 to be “the most advanced [model] in the industry next year.”
The company’s financial commitment to AI is significant, with projected capital expenditures ranging from $37 to $40 billion for the year, an increase of $2 billion from previous forecasts. Investors were warned to expect “substantial” increases in capital expenditures next year as well.
Despite these heavy investments, Meta CFO Susan Li noted that the company does not expect to generate revenue from generative AI this year. Li emphasized the company’s approach to building flexible AI infrastructure that can adapt to various use cases. She pointed out that the hardware used for training AI models could also be repurposed for inference and, with adjustments, for ranking and recommendations.
Meta’s ongoing AI initiatives, branded as “Core AI,” are already yielding positive results in improving user engagement on Facebook and Instagram. Zuckerberg highlighted the success of a recently launched unified video recommendation system for Facebook, which has already boosted engagement on Facebook Reels more than the initial shift from CPUs to GPUs.
Looking ahead, Zuckerberg envisions AI playing a key role in transforming Meta’s advertising business. He predicted that in the coming years, AI would take over ad copy creation and personalization, enabling advertisers to simply provide a business goal and budget, with Meta’s AI handling the rest.
Despite its considerable AI investments, Meta remains financially strong. Q2 results showed revenue of $39 billion and net income of $13.5 billion, marking year-over-year increases of $7 billion and $5.7 billion, respectively. The company’s user base continues to grow, with over 3.2 billion people using a Meta app daily, and its Threads app, a competitor to X, is approaching 200 million active monthly users.
As Meta navigates the AI landscape, its strategy reflects a long-term vision prioritizing technological development and infrastructure enhancement over short-term financial gains.
