United States Data Center Energy Usage Report: 2025 Update | Research in context

Lawrence Berkeley National Laboratory · Published 18 June 2026 · Checked 5 October 2026

Authors: Sarah J. Smith, Alex Hubbard, Alex Newkirk, Mohan Ganeshalingam, Billie Holecek, Dale Sartor, Michael Mills and Arman Shehabi

Read the original report

Data period: Modelled historical estimates cover 2016 to 2024; future scenarios cover 2025 to 2030. Primary shipment data covered only the first three quarters of 2025.
Scope: United States data-centre electricity use across servers, storage, networking and facility infrastructure. Cryptocurrency mining is excluded.
Method: A bottom-up model combines equipment shipments and the installed base with per-device power and use assumptions, then applies facility type and power usage effectiveness estimates. It updates the method used in the 2024 report; equipment datasets and power assumptions were compiled through December 2025.
Assumptions: Material inputs include server and accelerator shipments and lifetimes, rated and idle power, utilisation, workload mix, facility types, power usage effectiveness and the electricity forecast used as the national denominator.
Uncertainty: Four sensitivity scenarios and compounded stress-test bounds explore selected inputs. Inference utilisation lacks verifiable data. The scenarios exclude major departures caused by technology, market, supply-chain, regulatory or grid-availability changes; the bounds do not cover every uncertainty.
Funding and interests: Berkeley Lab conducted the work with DOE Industrial Technologies Office support under a DOE contract. Omdia, S&P Global and IDC supplied inputs. IDC datasets were purchased; Omdia paid products were supplied with complimentary access. Industry experts commented on assumptions. These relationships do not establish whether findings are correct.
Reuse: The report states that it is licensed under Creative Commons Attribution 4.0. The underlying proprietary input datasets retain their own terms and are not imported here.
What it cannot establish: It does not establish electricity demand, water use, emissions, grid impact, planning status, local costs, consent or delivered benefits for a particular site. It is not a global estimate or a commitment that a scenario will occur.

Continue from research to a local decision

This report gives wider model context, not evidence about one proposal. Use the community questions guide to ask for local evidence, inspect the recorded projects, or read the publication method. Unknown local conditions stay unknown.

This record explains one source within its stated boundaries. It is not a verdict on a project, a forecast that a scenario will occur or evidence of community consent.