Land Use Isn't One Number

A continuous 1-GW load consumes 8.76 TWh of electricity per year. One way to compare generation technologies is to ask how much nameplate capacity would be required to produce that amount of annual energy using observed capacity factors.

Using EIA's 2024 U.S. utility-scale capacity factors—90.8% for nuclear, 60.5% for natural-gas combined cycle, 23.2% for solar PV and 34.3% for wind—the corresponding nameplate capacities are approximately 1.10 GW of nuclear, 1.65 GW of combined-cycle gas, 4.31 GW of solar PV and 2.92 GW of wind. These are annual-energy equivalents, not guarantees that each resource can deliver 1 GW in every hour.

The harder issue is land. Different studies use different boundaries. NREL's utility-scale solar work distinguishes total area inside a project boundary from direct area occupied by arrays, roads, substations and other infrastructure. Wind requires another distinction: the perimeter of a wind project can encompass a very large area because turbines are spaced apart, while the land directly disturbed by turbine pads and roads is much smaller and the land between turbines can often remain in agricultural use.

Natural gas presents the opposite methodological problem. The generating station itself can be compact, but a life-cycle land assessment may also include gas production wells, gathering lines, processing facilities and other fuel-supply infrastructure. Nuclear likewise can be measured as a plant site or through a broader life-cycle assessment that includes fuel and supporting infrastructure.

For that reason, the old question 'How many acres does each technology need?' can be misleading unless the land metric is defined first. A fair comparison should use the same system boundary and normalize land use to electricity actually generated.