power usage effectiveness
In plain wordsA score showing how much of a data center's electricity actually reaches its computers: total power drawn divided by the computers' share. A perfect 1.0 would mean no waste; the extra goes mainly to cooling.
A data-center efficiency metric equal to the facility's total energy use divided by the portion consumed by its IT equipment (servers, storage and networking), both measured over the same period. A PUE of 1.0 would mean zero overhead; the excess is consumed by cooling, power conversion and distribution losses, lighting and other infrastructure. The Uptime Institute's annual industry survey has found the average roughly flat at 1.5–1.6 since about 2020 (1.54 in 2025), while large hyperscale operators report about 1.1 across their fleets and some liquid- or immersion-cooled sites report below 1.05. PUE measures infrastructure overhead only: it says nothing about how efficiently the IT equipment turns energy into useful computation, about water use (tracked as water usage effectiveness, WUE, in liters per kilowatt-hour) or about the carbon intensity of the electricity, and it can worsen when the IT equipment itself becomes more efficient.
Origin
Introduced by The Green Grid industry consortium (proposed 2006, promoted from 2007), with Christian Belady often named as its originator; standardized internationally as ISO/IEC 30134-2:2016. The Green Grid added water usage effectiveness in 2011.
Usage note
Figures can be flattered by drawing facility boundaries narrowly or quoting design rather than measured values, so audited annual figures are preferred. Since 2024 the EU has required operators of data centers with at least 500 kW of installed IT power to report PUE every year, under its 2023 Energy Efficiency Directive.
Related
- AI data centerAI data center, n.: A building or campus packed with AI chips, power equipment and cooling, built to train and run AI models. The biggest use as much electricity as a city.
- liquid coolingliquid cooling, n.: Cooling computer chips by running liquid past them instead of blowing air. AI chips now run so hot and are packed so tightly that air alone can no longer keep them cool.
- environmental impact of AIenvironmental impact of AI, n.: The effects of AI on the planet: the electricity, water, materials and land used to make chips, build data centers and train and run models, and the pollution that results.
- rack-scale systemrack-scale system, n.: An AI computer designed as one whole cabinet of chips, cooling and wiring, rather than separate servers, so all the chips inside can work together like one big chip.
Defined in law or standards
ISO/IEC 30134-2:2016 (also EN 50600-4-2:2016); EU: Commission Delegated Regulation (EU) 2024/1364, Annex III (data-centre sustainability indicators)
Mentioned in
AI data centerAI data center, n.: A building or campus packed with AI chips, power equipment and cooling, built to train and run AI models. The biggest use as much electricity as a city., environmental impact of AIenvironmental impact of AI, n.: The effects of AI on the planet: the electricity, water, materials and land used to make chips, build data centers and train and run models, and the pollution that results., liquid coolingliquid cooling, n.: Cooling computer chips by running liquid past them instead of blowing air. AI chips now run so hot and are packed so tightly that air alone can no longer keep them cool.