Mark Zuckerberg, founder of Meta, said that the social media giant will spend hundreds of billions of dollars for building huge AI calculus centers in the United States.
The first multi-gigawatt data center called Prometheus is expected to come online in 2026, said Zuckerberg.
He said one of the locations would cover an area almost as large as Manhattan (59.1 km/22 miles).
Meta has strongly invested in the efforts to develop what it described as a “superintelligence” – technology, of which she said that they could think about the smartest people.
The company, which has earned most of its money with online advertising, achieved more than 160 billion US dollars in sales in 2024.
In a contribution in his social media platform, threads, Zuckerberg said that Meta had built up several multi-gigawatt clusters, and this one cluster called Hyperion could scale up to five gigawatt over several years.
“We also build several more titanium clusters. Only one of these covers covers a significant part of Manhattan’s footprint,” he added.
Prometheus will be built in New Albany, Ohio, while Hyperion is being built in Louisiana and is expected to be completely online by 2030, said Zuckerberg.
He said Meta would “invest hundreds of billions of dollars … to build superintelligence” and that the centers “names” correspond to their size and impact “.
Karl Freund, main analyst at Cambrian Ai Research, told the BBC: “Zuckerberg intends to spend his way to the top of the AI heap”.
“The talent he sets will have access to some of the best AI hardware in the world,” added Freund.
Meta shares were traded by 1% higher after the announcement, reported Reuters News Agency. So far, the share has increased by more than 20% this year.
There are at least 10,000 data centers all over the world in which the Cloud -Remote -Server -Server Service Save Digital Information -with most in the United States, followed by Great Britain and Germany.
AI-controlled data centers are extremely energy and water-intensive. A study estimates that these centers could consume 1.7 trillion gallons worldwide by 2027. A single AI query – for example a request to chatt – can use about as much water as a small bottle that you would buy in the corner business.
