Meta Platforms plans to significantly ramp up its investment in artificial intelligence in 2026, nearly doubling capital expenditure as Chief Executive Officer Mark Zuckerberg accelerates his vision for what he calls “personal superintelligence.”

According to Meta’s fourth-quarter earnings report, the company expects AI-related capital expenditure in 2026 to range between $115 billion and $135 billion, almost twice what it spent the previous year following a restructuring of its AI division. Zuckerberg said the spending is focused on building the infrastructure required to train advanced AI models and deliver personalized AI tools to individuals and businesses worldwide.

Much of the investment will be directed toward large-scale data center infrastructure designed to support both current and future AI systems. Zuckerberg described 2026 as a defining year for artificial intelligence, adding that Meta’s expanded infrastructure would be central to achieving long-term breakthroughs.

While no specific products were unveiled during the earnings call, Zuckerberg confirmed that multiple AI-driven products would be released over the course of the year. “We’re not just launching one thing,” he said, noting that Meta is building several AI offerings in parallel. Industry reports indicate that Meta’s AI unit, led by former Scale AI chief executive Alexandr Wang, is developing a next-generation model codenamed Avocado, which is expected to succeed the company’s Llama family of models.

Zuckerberg defended Meta’s decision to develop its own foundational AI models, describing the company as a “deep technology” firm that cannot rely solely on external platforms. Owning core AI infrastructure, he said, allows Meta to shape how future products are built and deployed across its ecosystem, including platforms such as Facebook and Instagram.

The comments build on Zuckerberg’s broader vision for “personal superintelligence,” first outlined publicly last year. He has positioned Meta’s approach as distinct from other AI developers that emphasize large-scale automation and efficiency. Instead, Zuckerberg argues that AI should enhance individual agency and creativity.

In a public letter, he wrote that while superintelligence could dramatically accelerate human progress, its most important impact may be empowering people to improve the world in ways they personally value. He contrasted this with models centered on centralized AI systems automating most work, stressing that Meta’s long-term focus remains at the intersection of technology and everyday human life.