Is AI Destroying the Environment?
I’ve been hearing a lot about how data centers, especially those associated with AI, are hard on the environment. I didn’t understand why, so I did some research.
First, what are data centers? I wasn’t even sure what the term meant. It was just being used all the time lately. But, in case you’re also not in the-know: data centers are places that hold the physical computers that run things like cloud storage, streaming services, search engines, and AI programming—basically, all of the Internet runs from various data centers. They can range in size, but the ones causing the environmental problems are large, literal campuses.
How do data centers cause environmental problems? Two main reasons: these mega data centers are using a ton of energy and a ton of water. So, I continued down the rabbit hole. Why do they need so much energy?
It turns out that AI runs on something called graphics processing units (GPUs), which were originally created to handle, you guessed it, graphics. Since they were developed in 2006, they’ve been used for many other things, especially running Artificial Intelligence programs. They require a lot of energy because they work at such an accelerated rate. So, you fill a building full of GPUs that need to consume a ton of energy so they can keep running. I’ve read in many different sources that a single data center needs the energy equivalent to power a whole city. This is, of course, putting a strain on the power grid and uses a lot of fossil fuels. This, in turn, releases a lot of carbon dioxide into the atmosphere.
All of this power generates a lot of heat. To cool the equipment down, most data centers use water—a lot of water. According to a report put out by the Columbia River Keeper (an organization protecting the welfare of the Columbia River in Portland, Oregon), a 100 megawatt data center can use as much water in the course of a year as 2,500 people.
Unfortunately, the best water for data centers is potable water. Polluted water can have microbes or heavy minerals in it which can corrode the equipment, leading most data centers to use fresh, treated water. This obviously puts a strain in places especially areas with water scarcity.
So what do we do about these energy beasts that are essentially working the Internet? They aren’t going away, and nor are they inherently bad. There’s a lot we need to figure out with AI in addition to its energy usage, but it’s also here to stay.
I, like most people, use various aspects of the Internet all day long. Obviously, this blog. I utilize Google Drive for drafts and Gmail. I built my website on Squarespace. I like Facebook to keep up with friends and family and to share my blog posts. I listen to music on Pandora. My husband has a Youtube channel.
No, I don’t think the answer is to get rid of data centers. I think their problems point to a much larger problem. We need to utilize renewable energy sources. Yes, data centers are eating up a huge chunk of energy, but they aren’t the only ones. As a nation, most of the world, we are reliant on electricity. We’re all lit up, plugged in, burning, and chugging fossil fuels.
We don’t have to be. We know there’s other ways to power the world.
What’s your thoughts? And for some lighter reading, check back in on Monday, March 9th, for a glance at St. Patrick’s Day here in New Orleans.
Thanks for reading!
💜
Laura
P.S. If you’re interested to read more about data centers and the environment, here are the websites I used as sources:
Engineers often need a lot of water to keep data centers cool | ASCE
What Are the Environmental Impacts of Artificial Intelligence? - Union of Concerned Scientists
FINAL - DATA CENTER 2025.09.22 Data Center background brief
Why AI Consumes So Much Energy and What Might Be Done About It - Kleinman Center for Energy Policy
