Meta’s founder Mark Zuckerberg has mentioned the social media big will spend tons of of billions of {dollars} on constructing big AI information centres within the US.
The primary multi-gigawatt information centre, known as Prometheus, is anticipated to return on-line in 2026, Zuckerberg mentioned.
He mentioned one of many websites would cowl an space practically the dimensions of Manhattan (59.1 sq km/22.8 sq miles).
Meta has invested closely in efforts to develop what it known as “superintelligence” – expertise that it mentioned might out-think the neatest people.
The corporate, which has made most of its cash from internet advertising, generated greater than $160bn in income in 2024.
In a put up on his social media platform, Threads, Zuckerberg mentioned Meta was constructing a number of multi-gigawatt clusters, and that one cluster, known as Hyperion, might scale as much as 5 gigawatts over a number of years.
“We’re constructing a number of extra titan clusters as effectively. Simply one in all these covers a major a part of the footprint of Manhattan,” he added.
Prometheus shall be inbuilt New Albany, Ohio, whereas Hyperion shall be inbuilt Louisiana and is anticipated to be absolutely on-line by 2030, Zuckerberg mentioned.
He mentioned Meta would “make investments tons of of billions of {dollars}… to construct superintelligence” and that the centres had been given “names befitting their scale and impression”.
Karl Freund, principal analyst at Cambrian AI Analysis, instructed the BBC, “clearly, Zuckerberg intends to spend his method to the highest of the AI heap”.
“The expertise he’s hiring could have entry to a few of the finest AI {Hardware} on the planet,” Freund added.
Meta shares had been buying and selling 1% increased following the announcement, Reuters information company reported. The inventory has risen greater than 20% thus far this yr.
There are a minimum of 10,000 information centres world wide internet hosting the cloud – distant servers that retailer digital data – with most of them positioned within the US, adopted by the UK and Germany.
AI-driven information centres are extraordinarily power and water intensive. One examine estimates that these centres might eat 1.7 trillion gallons of water globally by 2027. A single AI question – for instance, a request to ChatGPT – can use about as much water as a small bottle you’d buy from the corner shop.
