How Will AI Change Our Planet?

So he started doing research, requesting information from the city and his local utility company. Today his findings are being used by groups in Colorado Springs to try to block the project.

Data centers have been around for decades. But the demand to build more and bigger facilities has exploded with the growth of AI, because the technology needs massive computing power. Since 2020, the number of data centers in the United States has more than doubled. The U.S. now has more than 3,000, according to Data Center Map, a database that tracks the facilities globally. That is about eight times more than any other country.

Giant technology companies like Amazon, Google, Meta, and Microsoft are investing billions of dollars to build even more data centers. Roughly 1,500 are planned or under construction across the nation (see map, below). Many will be hyperscale data centers. Those are the biggest type, with multiple buildings and rows of computers that stretch as long as a mile.

Communities like Kyle’s are trying to block data center projects—or at least slow them down. Towns and cities across the U.S. have passed temporary bans on building the facilities. And this summer, New York became the first state to impose a one-year moratorium on hyperscale data centers.

“AI is part of our lives, and it’s not something we can reverse,” Kyle says. “My concern is that data centers are happening too fast and without proper public oversight.”

He started doing research. He requested information from the city and his local utility company. Today his findings are being used by groups in Colorado Springs to try to block the project.

Data centers have been around for decades. But the demand to build more and bigger facilities has exploded with the growth of AI. That is because the technology needs massive computing power. Since 2020, the number of data centers in the United States has more than doubled. The U.S. now has more than 3,000. That is according to Data Center Map. It is a database that tracks the facilities globally. The U.S. has about eight times more than any other country.

Giant technology companies like Amazon, Google, Meta, and Microsoft are investing billions of dollars to build even more data centers. Roughly 1,500 are planned or under construction across the nation (see map, below). Many will be hyperscale data centers. Those are the biggest type. They have multiple buildings. And they have rows of computers that stretch as long as a mile.

Communities like Kyle’s are trying to block data center projects. Or at least they hope to slow them down. Towns and cities across the U.S. have passed temporary bans on building the facilities. And this summer, New York became the first state to impose a one-year moratorium on hyperscale data centers. 

“AI is part of our lives, and it’s not something we can reverse,” Kyle says. “My concern is that data centers are happening too fast and without proper public oversight.”

If you’ve ever asked ChatGPT a question, the response might seem effortless. But a lot happens behind the scenes for AI programs to generate answers.

First, your prompt zips to a data center where it is assigned a server—a powerful computer system that stores, processes, and transmits data. A large data center can contain tens of thousands of servers.

The server performs billions of calculations in less than a second to determine the chatbot’s response. Then that answer gets sent back to you. “Just to generate one single word . . . requires huge amounts of energy,” says Shaolei Ren. He is a professor at the University of California, Riverside, who studies AI.

That process happens nonstop at data centers as people turn to chatbots for everything from solving tricky math problems to deciding what to make for dinner. More than 60 percent of Americans use AI at least several times a week, according to Pew Research Center data.

What’s more, companies are adding AI to technology that people already use a lot. Google Search, for example, now includes AI-generated summaries with its search results.

When you ask ChatGPT a question, the response seems effortless. But a lot happens behind the scenes for AI programs to generate answers.

First, your prompt zips to a data center where it is assigned a server. That is a powerful computer system. It stores, processes, and transmits data. A large data center can contain tens of thousands of servers. 

The server performs billions of calculations in less than a second to determine the chatbot’s response. Then that answer gets sent back to you. “Just to generate one single word . . . requires huge amounts of energy,” says Shaolei Ren. He is a professor at the University of California, Riverside. He studies AI.

That process happens nonstop at data centers as people turn to chatbots. They use them for everything from solving tricky math problems to deciding what to make for dinner. More than 60 percent of Americans use AI at least several times a week. That is according to Pew Research Center data. 

What’s more, companies are adding AI to technology that people already use a lot. For example, Google Search now includes AI-generated summaries with its search results.

Powering all that AI content takes energy—and a lot of it. A large data center can use as much electricity as 100,000 households, and the biggest facilities consume significantly more.

Data centers often draw power from the same sources as local communities. The warehouses’ demand for electricity is so great that utility companies often need to build bigger power grids. That has many people worried about rising utility costs and who will foot the bill.

Adding to those concerns, much of the electricity that data centers use comes from burning fossil fuels, such as coal, oil, and natural gas. That process releases carbon dioxide and other greenhouse gases, which trap excess heat in the atmosphere. “[That] can potentially warm up the planet in the long run,” Ren says.

Powering all that AI content takes energy. And it takes a lot of it. A large data center can use as much electricity as 100,000 households. The biggest facilities consume significantly more. 

Data centers often draw power from the same sources as local communities. The warehouses require a huge amount of electricity. So utility companies often need to build bigger power grids. That has many people worried about rising utility costs. And they wonder who will foot the bill. 

In addition, much of the electricity used by data centers comes from burning fossil fuels. They include coal, oil, and natural gas. The process releases carbon dioxide and other greenhouse gases. They trap excess heat in the atmosphere. “[That] can potentially warm up the planet in the long run,” Ren says.

The Financial Times, a global news organization, recently analyzed the 60 largest data centers planned for the U.S. It estimated the facilities will emit approximately 111.8 million tons of carbon dioxide a year. That is roughly equal to the emissions of 24 million gas-powered cars, the study found.

Data centers also need a lot of water. Most of it is used to power them by generating electricity through steam. But the facilities also need water to prevent the servers from overheating. In many data centers, this is done by constantly pumping cool water through pipes near the servers. When it comes to large data centers, this process can use as much water as a small city.

In a typical home or business, water flows down a drain and can be treated and reused. But the water used by data centers usually evaporates, making it more challenging to recapture. As a result, researchers warn that the facilities can threaten water supplies—especially in drought-prone areas.

The Financial Times is a global news organization. It recently analyzed the 60 largest data centers planned for the U.S. It estimated the facilities will emit approximately 111.8 million tons of carbon dioxide a year. That is roughly equal to the emissions of 24 million gas-powered cars, the study found.

Data centers also need a lot of water. Most of it is used to power them by generating electricity through steam. But the facilities also need water to prevent the servers from overheating. In many data centers, this is done by constantly pumping cool water through pipes near the servers. For large data centers, this process can use as much water as a small city.

In a typical home or business, water flows down a drain. It can be treated and reused. But the water used by data centers usually evaporates. That makes it more challenging to recapture. As a result, researchers warn that the facilities can threaten water supplies. That is especially true in drought-prone areas. 

Many people argue data centers are necessary now that AI has become part of daily life. The technology is used just about everywhere, from stores to banks to hospitals. “We need data centers to keep our society functioning,” says Wangda Zuo, an engineering professor at Penn State.

Most experts say the solution is to make data centers more efficient. Some tech companies are looking into renewable energy sources like solar and wind. Scientists are also trying to develop facilities that can run with less electricity and water.

Those changes will happen over time, Ren predicts. “Car emissions were a lot worse 100 years ago, and now they’re more efficient,” the professor says. “The same thing will happen with AI sooner or later.”

Plus, some people point out that AI may one day be smart enough to come up with solutions for climate change. Already, the technology is helping scientists monitor melting sea ice and guiding the gas industry in how to cut emissions.

In the meantime, now that Kyle knows the environmental costs that come with AI, he’s rethinking how often he uses ChatGPT. He is also continuing to voice his concerns about data centers. “I’ve learned that speaking out is a way to be able to get answers and to bring awareness to the cause,” he says. 

—with additional reporting by Lauren Vespoli

Many people argue data centers are necessary now that AI has become part of daily life. The technology is used just about everywhere, from stores to banks to hospitals. “We need data centers to keep our society functioning,” says Wangda Zuo. He is an engineering professor at Penn State.

Most experts say the solution is to make data centers more efficient. Some tech companies are looking into renewable energy sources like solar and wind. Scientists are also trying to develop facilities that can run with less electricity and water.

Those changes will happen over time, Ren predicts. “Car emissions were a lot worse 100 years ago, and now they’re more efficient,” the professor says. “The same thing will happen with AI sooner or later.”

Plus, some people point out that AI may one day be smart enough to come up with solutions for climate change. Already, the technology is helping scientists monitor melting sea ice. And it is guiding the gas industry in how to cut emissions.

In the meantime, Kyle now knows the environmental costs that come with AI. So he is rethinking how often he uses ChatGPT. He is also continuing to voice his concerns about data centers. “I’ve learned that speaking out is a way to be able to get answers and to bring awareness to the cause,” he says.

—with additional reporting by Lauren Vespoli