Quantifying national trends in data center electricity and water use to support load forecasting and infrastructure planning.
Overview
Reliable forecasts are increasingly important as rapid growth in computing demand increases electricity and water consumption. Lawrence Berkeley National Laboratory (LBNL) conducts comprehensive assessments using models, datasets, and analytical tools to track historical trends, quantify the impacts of technological advances and efficiency improvements, and project the future resource needs of U.S. data centers. These assessments help utilities, policymakers, planners, and other stakeholders anticipate load growth and infrastructure needs.
Our Work
Our researchers developed a detailed “bottom-up” model, grounded in shipments of computing equipment and thermodynamic modeling of cooling systems, to estimate total on-site electricity and water demand across data centers in the United States. LBNL has advanced this research for two decades, with estimates extending back to 2000.
Key Findings
- The 2025 Update estimates that data centers could account for 11.8% of total U.S. electricity use by 2030, with scenarios ranging from 9.5 and 15.3%. The previous report projected a range of 6.7 to 12.0% by 2028.
- The 2024 United States Data Center Energy Usage Report provides a comprehensive assessment of data center electricity and water consumption and projects a range of potential demand scenarios through 2028.
- The 2021 article in Environmental Research Letters quantified the spatially detailed water footprints of U.S. data centers, including water used directly for cooling and indirectly to generate electricity.
- The 2020 article in Science revealed that global data center energy use increased by only 6% during the 2010s, while compute instances increased 550%.
LBNL’s bottom-up national data center energy model is a foundational part of DOE’s energy usage reports. This model is regularly updated based on new equipment data and industry insights.
LBNL’s location database categorizes data centers by type and utility power needs to highlight current and emerging data center loads for different electricity utilities and balancing authorities.
The Shape Maker tool generates customizable electricity load profiles for data center, facility, and grid planning.
LBNL’s model estimates location-specific water use from onsite cooling and electricity generation under different cooling-system designs and power-supply scenarios.
Related Publications & Resources
