Explosive AI growth is emerging as a major climate threat

The AI tools we use every day – from voice assistants to image generators – aren’t floating in some invisible cloud. They run on powerful computers that live in huge data centers.

These centers don’t just sit quietly humming away. They suck up electricity like sponges and burn through water to keep cool. And as AI use explodes, so does the demand for the machines behind it.

The scale of this impact has always been tough to pin down. But now, for the first time, researchers have mapped out what this might actually look like across the United States.

AI’s footprint is massive

By 2030, if AI keeps growing the way it is, it could produce 24 to 44 million metric tons of carbon dioxide every year. That’s like adding 5 to 10 million extra cars on the road.

And it doesn’t stop with emissions. AI infrastructure could also use 731 to 1,125 million cubic meters of water per year. That’s roughly how much water 6 to 10 million Americans use at home in a year.

These numbers are not projections for some distant future. They’re only five years away.

Climate impacts of AI growth

This new analysis comes from a research team led by Fengqi You, a professor in energy systems engineering at Cornell University.

The researchers didn’t just count servers and guess at power use. They spent three years gathering financial, manufacturing, and marketing data across the tech industry.

The team added in location-specific details about power grids, water supply, and climate conditions.

“Artificial intelligence is changing every sector of society, but its rapid growth comes with a real footprint in energy, water and carbon,” said You.

“Our study is built to answer a simple question: Given the magnitude of the AI computing boom, what environmental trajectory will it take? And more importantly, what choices steer it toward sustainability?”

A path forward

The good news is this: if done right, the damage can be scaled back. The study offers a roadmap that could reduce carbon emissions by about 73% and water use by 86% compared to worst-case scenarios.

The plan requires coordination across three key areas: where data centers are built, how power is generated, and how efficiently the centers operate.

“There isn’t a silver bullet,” You said. “Siting, grid decarbonization and efficient operations work together – that’s how you get reductions on the order of roughly 73% for carbon and 86% for water.”

Location makes a big difference

One of the biggest problems? Many of the data centers are being built in places that don’t have a lot of water to begin with.

Nevada and Arizona, for example, are already facing water stress. Building more high-demand tech in these areas only adds to the strain.

Even regions like northern Virginia, which are packed with data centers, are seeing local infrastructure pushed to its limits. That includes power grids and water systems.

The study found that choosing locations with lower water stress and better cooling strategies could cut water use in half. And if those changes are paired with greener electricity and better operations, total water savings could hit 86%.

Some of the best spots are in the Midwest and wind-heavy states like Texas, Montana, Nebraska, and South Dakota. These areas offer a better mix of cleaner power and more water stability.

New York also ranks high thanks to its mix of nuclear, hydro, and renewable energy. But even there, more efficient cooling is essential.

Even clean grids can’t keep up

Clean energy is growing, but not fast enough to keep pace with AI demand. If current trends continue, emissions could actually rise by around 20%.

“Even if each kilowatt-hour gets cleaner, total emissions can rise if AI demand grows faster than the grid decarbonizes,” You said. “The solution is to accelerate the clean-energy transition in the same places where AI computing is expanding.”

The researchers found that even in the most optimistic clean-energy scenario, emissions would only drop about 15%. That still leaves behind 11 million tons of carbon dioxide.

Getting rid of that would take 28 gigawatts of wind energy or 43 gigawatts of solar power – a big lift.

How AI can reduce its own footprint

Improved technology can make a meaningful dent as well. More efficient liquid cooling and smarter server use could trim another 7% of carbon emissions and reduce water consumption by 29%.

Layered onto the other measures, total water savings climb to roughly 32%. But none of this will happen by accident. As tech companies like OpenAI and Google race to build more data centers, time is running out to get things right.

“This is the build-out moment,” You said. “The AI infrastructure choices we make this decade will decide whether AI accelerates climate progress or becomes a new environmental burden.”

The full study was published in the journal Nature Sustainability.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–