By Tracy Turner
Ohio State University
COLUMBUS — Ohio’s expanding data center industry could require major investments in power generation and transmission as artificial intelligence drives growth, raising questions about who pays for the infrastructure needed to serve those facilities.
Researchers at Ohio State University College of Food, Agricultural and Environmental Sciences project that six major technology companies’ data centers could use more than twice as much electricity worldwide by 2030 as they did in 2024. Ohio is among the locations projected to have the highest demand.
The study finds that this growth could place greater pressure on regions where data centers would consume a larger share of available electricity.
Under the study’s middle-range growth scenario, the six companies’ Ohio data centers could consume more than 15 terawatt-hours annually by 2030. To illustrate the scale, 15 terawatt-hours equals roughly one-tenth of the electricity sold to all Ohio customers in 2024, according to the U.S. Energy Information Administration.
Nearly two-thirds of the six companies’ projected data center electricity use would be concentrated in fewer than 10 states and countries. Ohio, Oregon, Virginia, Iowa and Ireland are each projected to exceed 15 terawatt-hours annually by 2030 under the study’s middle-range growth assumptions.
The study focused on Amazon, Microsoft, Google, Meta, Oracle and Apple because of their large data center operations, investments in AI infrastructure and publicly available information about their facilities and energy use.
The researchers project that these companies’ combined annual data center electricity consumption worldwide could increase from about 118 terawatt-hours in 2024 to between 239 and 295 terawatt-hours by 2030, roughly 1% of projected global electricity demand. Those totals include AI and non-AI workloads. A terawatt-hour equals 1 billion kilowatt-hours.
For households, farms and businesses, the implications depend partly on whether data center customers cover the infrastructure costs associated with their demand, said Ani L. Katchova, a professor and holder of the Farm Income Enhancement Endowed Chair in the CFAES Department of Agricultural, Environmental and Development Economics.
“Whether data center growth delivers broad economic benefits for Ohio will depend on how durable the local benefits are, how infrastructure costs are allocated and how large the associated land, water and environmental costs become,” Katchova said.
If other electricity customers bear too much of the cost of serving data centers, households, farms and businesses could potentially face higher electricity costs or greater price volatility, she said. The study does not calculate those potential effects.
Q: How much electricity is 15 terawatt-hours?
A: 15 terawatt-hours is equivalent to 15,000,000,000 kilowatt hours. That number is about 10% of the electricity sold all Ohio customers in 2024.Q: How much electricity does my home use in a year?
A: According to the Office of the Ohio Consumer’s Council, the average Ohio home uses between 750 and 1,000 kilowatt-hours per month, or around 9,000 to 12,000 per year.
Published in Communications Sustainability, the study’s co-authors include Yongyang Cai, a professor, and Danbo Chen, a doctoral student, both in AEDE.
The researchers examined how company expansion plans could affect regions with different electricity supplies, connecting projected data center growth with the electricity available to serve it.
“The main implication of our study is the importance of anticipatory planning,” Cai said. “Utilities and system planners should incorporate large data center loads directly into long-term demand forecasts and should coordinate generation, transmission and interconnection planning with expected data center development.”
The team used large language models to analyze corporate disclosures, policy materials and media reports, then combined that information with projections of computing capacity, equipment efficiency and electricity use.
The CFAES researchers developed an Electricity Demand Pressure Index comparing projected annual data center demand with available electricity. The index shows how large a share of a region’s electricity supply the facilities could use. Oregon, Iowa and Ireland show particularly high relative pressure. Ohio falls in the upper-middle range, while larger systems such as Texas show lower relative pressure.
However, statewide electricity availability does not establish whether a particular location can accommodate new facilities. Even if enough electricity is available statewide over a year, local power lines and connections may not be able to deliver it where and when it is needed. The index does not capture seasonal demand, transmission bottlenecks or connection delays, Cai said.
The estimates are scenarios rather than precise forecasts.
Katchova said Ohio is already using special rate structures intended to make large data center customers pay more of the costs of the electricity infrastructure needed to serve them.
“All in all, the broader question is not simply whether data centers create economic activity, but how the benefits and costs are distributed across households, farms, businesses, local governments and the data centers themselves.”
Additional co-authors were Zijun Zhou of Meta Platforms and Jiahong Qin and Lei Chen of Zhejiang A&F University in China.
(Tracy Turner is a senior communications specialist with Ohio State University College of Food, Agricultural and Environmental Sciences.)
