The U.S. data center boom could spike carbon emissions through the end of the decade to the equivalent of adding 5 million to 10 million cars annually to U.S. roadways and stress water supplies significantly, according to a new peer-reviewed study led by Cornell University researchers.
The analysis in Nature Sustainability, three years in the making, aims to quantify the environmental effects of the U.S. artificial intelligence build-out across the country, a task that has been challenging because of the industry’s size and pace of growth. Researchers found that the technology currently is being expanded in a way that will exacerbate emissions and water challenges in many regions, while hindering the ability of large technology companies to reach net-zero targets.
“The future is in our hands. AI is growing fast, but the future doesn’t have to come with a penalty to our climate, to our water,” said Fengqi You, an energy systems engineering professor at Cornell and co-author of the study, which was funded partly by the National Science Foundation.
At the current rate of growth, U.S. data centers would add 24 million to 44 million metric tons of carbon dioxide annually to the atmosphere through the end of the decade, the study said. The build-out also is projected to drain 731 million to 1,125 million cubic meters of water per year, an amount equivalent to the household usage of 6 million to 10 million Americans.
