As technology companies move forward in their quest to develop advanced forms of AI, the electricity consumption of the data centers that power that technology is expected to increase dramatically—especially in states like Virginia, where so many data centers have been built that one of its states is named. “Data Center Alley.”
Data centers accounted for 4 percent of total electricity consumption in the US last year, and this figure may rise to 9.1 percent by the end of the decade, according to an analysis by the Electric Power Research Institute (EPRI). The “dramatic growth in investment aimed at building and deploying new AI capabilities” raises the alarm about the technology’s energy consumption and environmental impact, the report said.
Big Tech companies like Meta ( META ), Amazon ( AMZN ), Microsoft ( MSFT ) and Google ( GOOGL ), which are rapidly building data centers, saw their electricity consumption more than double between 2017 and -2021, according to EPRI. With AI applications—currently consuming 10 percent to 20 percent of data center electricity—rapidly advancing in capabilities, this number is only expected to rise. A single ChatGPT application, for example, uses 1o times the electricity of a Google query, while AI applications that generate image, sound and video are “unprecedented,” according to EPRI.
Although data center usage is growing rapidly, it will not be paid equally. Fifteen states account for 80 percent of the nation’s data center load, an EPRI analysis found. Virginia, whose data centers consumed about 33.8 million megawatts of electricity last year, about a quarter of the state’s total power, leads the pack.
How did Virginia become a data center?
Virginia’s reliable fiber infrastructure, educated technical workforce and lack of natural disasters have long been a draw to data centers, according to Buddy Rizer, executive director of economic development in Loudoun County, Va. The region is home to “Data Center Alley,” boasting 25 million square feet of operational data centers in what constitutes the world’s largest concentration of resources.
From 2007 to 2016, Loudoun County saw “steady growth” in the data center area, Rizer told the Observer. “Since then, growth has really exploded with the advent of cloud providers, Covid, and now AI,” he added.
Texas ranks behind Virginia in data center energy consumption in 2023 with 21.8 million megawatts used. The state has become a popular data center destination due to its abundance of available land and business-friendly attitude, according to an EPRI study. California, Illinois and Oregon follow Virginia and Texas as the states with the highest data center energy consumption.
By the end of the decade, these areas are expected to become more power-hungry as AI models proliferate, “which typically consume more energy than the data retrieval, streaming and communication systems that have driven data center growth over the past two decades,” according to EPRI research. At the high end of EPRI’s projections, Virginia could see its annual energy use rise to nearly 90 million units by 2030 and account for 46 percent of the state’s total electricity consumption.
The “immediate challenge” with the rapid growth of the data center is the kind that causes the power grid, said Rizer, who added that the uncertainty and its impact on the regional budget is “not good.” The possibility of rising facility prices is also a cause for concern, as is the aesthetics and low noise of the data centers themselves.
To prepare for the AI-related hike in energy consumption, technology companies have begun looking for ways to negate the undesirable aspects of their data centers. Microsoft, Google and Amazon have all struck nuclear power deals in recent months as they bet on clean energy to offset rising carbon emissions from the facilities. Amazon also recently partnered with startup Orbital Materials to explore ways to decarbonize future data centers.
Citing nuclear and local energy solutions, Rizer said he already “sees different types of futures emerging” that could help combat energy problems. “We are doing things better now than we did 10 years ago, and we will be doing things better in five years than we are now,” he predicted.
