Mark Zuckerberg to build Manhattan-sized AI data center in Meta’s superintelligence drive

In a move to solidify Meta’s position in the race for artificial general intelligence (AGI), CEO Mark Zuckerberg has announced plans to build a data center nearly the size of Manhattan. The facility, described as a multi-gigawatt titan cluster, will support the company’s ambitious AI initiatives, including its newly formed Superintelligence Labs division. As part of its broader AI strategy, Meta plans to spend hundreds of billions of dollars over the coming years, reflecting the scale and urgency of its vision.With superclusters like Prometheus set to come online in 2026 and Hyperion following, Meta aims to outpace rivals like OpenAI and Google in the development of next-generation AI models and infrastructure. Meta’s bold investment marks a new phase in AI competition, aiming to integrate advanced AI across social media, VR, and the metaverse. Analysts say this could fast-track autonomous AI breakthroughs and transform human–technology interaction.Zuckerberg confirmed that Meta is building several massive AI compute clusters, each with the energy footprint of a small city. One such facility, Prometheus, will be the company’s first multi-gigawatt data center, while Hyperion is designed to scale up to 5 gigawatts over time.These efforts are intended to support the creation of superintelligent AI systems—machines capable of outperforming humans in a wide range of tasks. To fund this vision, Meta raised its 2025 capital expenditure estimate to between $64 billion and $72 billion. Zuckerberg cited the company’s strong ad revenue, $165 billion last year, as a key enabler of such spending. “We have the capital from our business to do this,” he told investors.In addition to infrastructure, Meta is aggressively investing in talent. The company recently formed Superintelligence Labs, a new division led by former Scale AI CEO Alexandr Wang and ex-GitHub chief Nat Friedman. Zuckerberg has reportedly led a personal effort to attract top-tier researchers, offering multimillion-dollar packages, some reaching $100 million, to secure talent from rival firms.These hires follow challenges with Meta’s open-source Llama 4 model and the departure of key staff. With new AI applications like Meta AI, image-to-video ad tools, and smart glasses in development, the company is working to diversify beyond its traditional social media roots. Industry analysts suggest Meta could become the first company to bring a gigawatt-plus AI supercluster online, cementing its leadership in the global AI race.