What does it take to create a superintelligent AI?
For Meta’s Mark Zuckerberg, the answer is simple: hundreds of billions of dollars—and data centres bigger than Manhattan.
In a post on Threads, Zuckerberg revealed Meta’s plans to build massive AI hubs across the U.S.
One facility in Ohio, dubbed Prometheus, is slated to go live in 2026.
Another, called Hyperion, will rise in Louisiana by 2030.
It could eventually scale to a staggering 5 gigawatts—an energy footprint rivaling small countries.
“These aren’t just data centres,” he wrote. “They’re titan clusters. Just one covers a significant part of Manhattan.”
What’s The Goal?
Meta’s mission? To build “superintelligence”—AI powerful enough to surpass the smartest human minds.
And with more than $160 billion in 2024 revenue, they’ve got the cash to back the ambition.
“This is Zuckerberg spending his way to the top of the AI heap,” said analyst Karl Freund.
Meta’s top hires will have access to the world’s best AI hardware.
But not everyone’s cheering. These mega-centres will guzzle energy and water.
One study warns AI could soak up 1.7 trillion gallons globally by 2027. All for what? The next ChatGPT response?
The AI race is heating up—but is the planet ready for the cost?
