MENLO PARK, California: Meta CEO Mark Zuckerberg has unveiled plans to spend hundreds of billions of dollars building a network of massive AI data centers—part of a bold effort to lead the global race toward developing artificial general intelligence, or "superintelligence."
The announcement, made via a post on Threads, highlights the scale of Meta's ambitions in artificial intelligence. The company aggressively competes against rivals like OpenAI and Google for top talent and infrastructure dominance.
Meta's upcoming multi-gigawatt data centers, including Prometheus, due online in 2026, and Hyperion, which could scale up to 5 gigawatts, will support what Zuckerberg calls "titan clusters."
"We're building multiple more titan clusters as well. Just one of these covers a significant part of the footprint of Manhattan," he said.
Zuckerberg cited a report from industry site SemiAnalysis, which suggested Meta could be the first AI lab to bring a gigawatt-plus supercluster online.
He defended the massive expenditure by pointing to Meta's strong ad business: "We have the capital from our business to do this."
The company recently reorganized its AI teams under a new division, Superintelligence Labs, following setbacks in its open-source Llama 4 model and the exit of key researchers. Former Scale AI CEO Alexandr Wang and ex-GitHub chief Nat Friedman are now leading the division.
Meta has raised its 2025 capital expenditure guidance to between US$64 and $72 billion, with hopes that AI will drive revenue through apps like Meta AI, ad tools, and smart glasses.
While some analysts praised Meta's ad performance gains via AI, others cautioned that the scale of this investment is aimed at long-term dominance, and returns may take time.
