Menu Menu
[gtranslate]

Meta set to invest hundreds of billions on AI data centres

The company’s CEO Mark Zuckerberg has announced that it will be investing hundreds of billions on AI centres that will cover an area almost as large as Manhattan.

Mark Zuckerberg has announced that Meta will be splashing out the cash and investing heavily in AI data centres across the US.

The first will be called Prometheus and is planned to be booted up in 2026. The CEO said that one of its data centre sites will cover an area ‘nearly the size of Manhattan,’ which is 59.1 squared kilometres.

Zuckerberg posted on Threads – an X competitor that was founded by Meta – stating that the company was building several ‘multi-gigawatt clusters.’ One of these is called Hyperion and it is planned to eventually scale up to five gigawatts in the coming years.

‘Prometheus will be built in New Albany, Ohio,’ Zuckerberg posted. ‘While Hyperion will be built in Louisiana and is expected to be fully online by 2030.’

 

View on Threads

 

These big steps are part of a wider strategy to stay ahead in the race for superintelligent AI language models. Meta are up against major rivals like Google, ChatGPT, and others.

Zuckerberg has largely sidelined his original Metaverse vision in favour of going ‘all in’ on artificial intelligence products, throwing eye-watering amounts of money at research, development and production.

In his statement, the tech mogul said that Meta would ‘invest hundreds of billions of dollars.’

Meta reportedly generated over $160 billion USD in 2024, mostly from online advertising. It’s recently incorporated new AI features into WhatsApp and Instagram, with more features planned later down the line.

We also recently wrote about how Meta was looking to invest $15 billion USD in Scale AI, a startup that specialises in data labelling and model evaluation. It’s clear that Zuckerberg is hungry for top-tier talent and resources in order to make his AI dreams a reality.

As the BBC reports, AI-driven data centres are extremely costly for the environment, using excessive amounts of energy and water in order to function.

Any one query to an AI language model uses roughly 500ml of water, which quickly adds up. Experts estimate that 1.7 trillion gallons of water could be consumed globally to power data centres by 2027.

Still, the climate cost won’t stop most of us from diving into ChatGPT on a regular basis.

Gen Z in particular are seeing their job market prospects change as a result of AI, and are developing skillsets with text prompts and AI-based workflows in mind. Not that they’re all happy about it, mind.

Accessibility