Data Centers Are Burning Fossil Fuels. It’s Destroying Big Tech’s Climate Goals – CNET

From Ashburn, Virginia, to South Dublin, Ireland, the info facilities that journalist Paris Marx visited all appeared remarkably related: giant industrial services with loud air-conditioning models working to maintain racks upon racks of servers cool. But it was the safety measures across the knowledge facilities that struck Marx greater than something. 

Tall fences. Security cameras in every single place. Locks on each door that required one of many few full-time workers to unlock with a badge. And in a single memorable occasion, standing on an Irish public curb exterior a cluster of Microsoft-owned facilities, there was an eerie feeling of being watched by males on garden mowers, trimming grass that clearly didn’t should be lower.

“It was just really staggering to me that it was at that level that it was so hard to move through these spaces,” Marx instructed me.

The intense safety ought to provide you with an thought of simply how critical Big Tech is about its knowledge facilities. These services are the important thing to the bogus intelligence packages at OpenAI, Meta, Google, Amazon and elsewhere, creating huge quantities of compute and sucking up equally staggering quantities of electrical energy and water. 

There isn’t any stronger metaphor for the manifestation of Big Tech’s energy than the growing sprawl of knowledge facilities. Amazon, Microsoft and Google alone management 59% of hyperscale knowledge heart capability. In his new e book obtainable now, Hyperscale: The Ambition and Excess of Big Tech’s Data Empires, Marx, host of the Tech Won’t Save Us podcast, highlights how world knowledge heart protests are one among many ways in which communities are pushing again towards Big Tech’s energy over us.

Hyperscale calls for

Data facilities existed lengthy earlier than generative AI; we referred to as them server farms, they usually’re what maintain us on-line. But on this AI period, we’ve got hyperscale knowledge facilities.

Before, a big knowledge heart may require round 50 megawatts of energy to maintain every part working. Now, hyperscale knowledge facilities can require triple that quantity of energy. Newer ones, like Meta’s Louisiana knowledge heart, could require 4,500 megawatts (4.5 gigawatts) — staggering demand that officers say is 4 occasions greater than the entire electrical energy utilized in New Orleans on its hottest day in 2025. It’s greater than the 2 gigawatts the Hoover Dam produces.

Hyperscale knowledge facilities are additionally housing extra superior tech. Chips like GPUs are key to creating compute, and newer chips constructed for AI knowledge facilities can run hotter. That makes it even more durable to chill services, so along with air-con, services are utilizing water programs to keep away from servers overheating. It’s a further environmental toll. A hyperscale heart can suck up 1 million to five million gallons of water on daily basis, equal to a city of greater than 10,000 residents, in line with a 2025 estimate.

So how do you retain these services working? The course of to get accepted and linked to a city’s electrical grid can take years. So tech corporations have began constructing their very own energy crops proper subsequent to knowledge facilities for the only goal of protecting them on — good for burning huge quantities of fossil fuels, the principle contributor to the local weather disaster.

Hyperscale impacts

In the 2010s, Amazon, Microsoft, Google and plenty of others pledged some model of changing into carbon impartial (or detrimental) within the subsequent few a long time — that means they might offset the emissions they created with not less than as a lot emission-reducing actions. 

Sustainability reviews from that point present that, general, tech corporations have been measurably reducing their environmental footprint. “And then ChatGPT comes along, and all of a sudden, that’s out the window,” Marx instructed me.

Tech leaders have been candid about permitting generative AI to override their local weather targets. Google reported a 50% enhance in its greenhouse gasoline emissions in 2024 to help its generative AI. Microsoft’s pledge to be carbon detrimental by 2030 “was before the explosion in artificial intelligence,” Brad Smith, the corporate’s president, instructed Bloomberg in 2023, saying the carbon objective was “five times as far away as it was in 2020” due to AI. 

Modern knowledge facilities and their energy crops burn fossil fuels for energy, requiring gasoline generators, big tanks of diesel gasoline stored on-site in case of an influence outage or good old school coal. This has devastating environmental and well being penalties, each instantly for native communities and in the long run for the whole world. 

Communities that neighbor knowledge facilities have reported brown or unsafe consuming water, never-ending noise, and in a single lawsuit towards Elon Musk’s xAI facilities in Memphis, excessive well being dangers; Memphis residents “are feeling it in their lungs because of the fossil fuels that are being burned,” Marx mentioned.

And as a long time of local weather analysis show, burning extra fossil fuels is just going to exacerbate the consequences of local weather change. That’ll result in much more excessive climate, well being dangers and pure disasters. As Marx put it in his e book, “The world will suffer through the worsening natural disasters fueled by the quest for computation by some of the most powerful people on the planet.”

With knowledge facilities and their fossil fuels, we’re basically reshaping the planet for AI. So anti-data heart actions are asking: Is it price it?

“How much computation do we really need here?” Marx requested in our dialog. “How many resources are we willing to dedicate to this vision of the future? And is this technology really delivering in such a way that expending all those resources is justified?”

The reply — for a lot of knowledge heart neighbors in Memphis, Oregon, Chile, Ireland and hundreds of different communities throughout the globe — is an unflinching no.

Katelyn Chedraoui

Reporter 2

Katelyn is a reporter with CNET protecting synthetic intelligence, together with chatbots, picture and video turbines. Her work explores how new AI expertise is infiltrating our lives, shaping the content material we eat on social media and affecting the individuals behind the screens. She graduated from the University of North Carolina at Chapel Hill with a level in media and journalism. You can attain her at kchedraoui@cnet.com.

See full bio