.png){kind=small}
The expansion of data centers in the U.S., caused by massive increases in demand for AI software capabilities, is fueling unprecedented levels of growth in U.S. energy demand, placing significant stress on the country’s aging grid system.
A recent report, “Strategic Industries Surging: Driving U.S. Power Demand,” by consulting firm GridStrategies, found that these data centers, which are essential for cloud computing, AI, and other digital and web services, require vast amounts of energy. This issue is particularly prevalent in "hot spot” states like Texas and Virginia, which are home to some of the largest concentrations of data centers in the world. The five-year electric load growth forecast has increased five-fold.
As the digital and data economies grow, energy producers face increasing difficulty meeting this demand while managing a grid already strained by aging infrastructure and extreme weather events. These regions could see current energy demands quadruple as companies attempt to build these centers as quickly as possible, raising peak energy demand by 128 GWs by 2029. This represents a 16% increase in energy demand by 2029, while previous estimates predicted demand to increase by only around 3% yearly for the next five years.
While 3% annual growth may sound reasonable, the report notes that would represent six times the planning and construction of new generation and transmission capacity than is now on the table. The rapid growth comes after an extended period where the utility industry has seen little year-over-year growth in demand. However, the report also notes that the estimates could be underestimated or overestimated, as the technology is rapidly developing and advancing, and that uncertainty also presents a challenge.
Data centers require large amounts of energy on a constant basis, with their load expected to rise further in the coming years. This rise in demand comes at a time when the national, as well as regional, grid is already struggling with reliability issues, more frequent outages, and the pressure to transition toward renewable energy.
While solar and wind power are expanding, their intermittent availability and insufficient energy storage capacities pose challenges for supporting not only the current levels of energy demand, but also the projected future demand as data center demand grows. AI data centers currently account for 4.4% of national energy demand, but McKinsey projects that AI may represent 70% of energy demand growth by 2030.
Addressing this growing energy need requires large-scale, immediate investment in grid modernization. Upgrading transmission lines, expanding energy storage, and creating a diverse mix of reliable energy sources is crucial. Without these improvements, the growing presence of data centers could exacerbate grid instability and further increase the risk of blackouts in areas with large energy demands.
The report indicates that if the grid can’t keep up with demand it could jeopardize the nation’s economic, technological, and geopolitical leadership. Meeting this rising demand is of strategic importance to the U.S. and major regulatory and policy shifts will be needed to do so.