Data centers are an essential part of the infrastructure that allows us to access the digital services we use daily. These buildings host servers, but also many other equipment that ensure they run smoothly, including cooling systems.
This equipment runs 24/7, which leads to significant electricity consumption. In 2024, data centers alone consumed 415 TWh, which accounted for about 1,5 % of the world’s electricity consumption⁽¹⁾. While this figure may seem low at first glance, it is roughly equivalent to the total electricity consumption of a country such as France (439 TWh in 2023)⁽²⁾.
Furthermore, the production of electricity that powers data centers is responsible for significant greenhouse gas emissions (GHG), especially considering that the countries with the highest electricity consumption for data centers are the United-States of America (45 % of global data center electricity consumption) and China (25 %)⁽¹⁾, two countries with highly carbon-intensive electricity mixes.
Finally, data centers consume large - and sometimes very large - amounts of water to operate their essential cooling systems.
Therefore, data centers have significant environmental impacts, and these impacts can vary greatly from one facility to another and depending on the systems in place. This is why it is so important to have metrics to assess the environmental efficiency of a given data center.

PUE

Of the 3 indicators mentioned in this article, PUE, for Power Usage Effectiveness, is probably the most well-known and the most commonly used.
It is calculated using the following formula :

$$ PUE = \frac{Total\ energy\ consumed\ by\ the\ data\ center}{Energy\ consumed\ by\ IT\ equipment} $$

This energy efficiency indicator was introduced by Malone and Belady in 2006⁽³⁾, before being more widely popularised by The Green Grid, a nonprofit industry consortium, starting from 2007⁽⁴⁾. Since then, PUE has been standardized in 2016 : ISO/IEC 30134-2 and EN 50600-4-2.

As indicated in the formula, this indicator takes into account the energy consumed (in kWh), which includes electricity, but also other types of energy (gas, diesel, or refrigerants for example).
The energy consumed by IT equipment includes the energy consumed by any equipement used for storing, processing or transporting data (computers, monitors, servers, switches, routers, etc…).
The total energy consumed by the data center includes the energy consumed by IT equipment, as well as that consumed by all the other systems (UPS, cooling, heating, lighting, security systems, elevators, etc…).
By definition, PUE is therefore always greater than 1, and the closer its value is to 1, the more energy-efficient the data center is.

In 2023, the average PUE value for data centers in the European Union was 1,6⁽⁵⁾. But this value can vary greatly from one data center to another, in particular depending on the cooling method. Data centers with the best air-cooling technologies achieve a PUE of around 1,3, while data centers using liquid-cooling can achieve a PUE of 1,04. (NB : with the advent of AI, servers are increasingly powerful, thereby increasing the amount of heat per m² that needs to be dissipated, to a point where air-cooling is no longer effective enough. This is why AI requires new data centers, for more power and liquid-cooling.)

Despite its evident value for assessing the energy efficiency of a data center, PUE has its limitations and doesn’t provide information on the pollution associated with the data center’s energy consumption.

CUE

CUE, for Carbon Usage Effectiveness, focuses on the greenhouse gas emissions (GHG). This indicator was introduced in 2010 by The Green Grid⁽⁶⁾ and was later standardized in 2022 (ISO/IEC 30134-8). It is the ratio between the data center’s annual GHG emissions and the energy consumed by IT equipment.

$$ CUE = \frac{GHG\ emissions}{Energy\ consumed\ by\ IT\ equipment} $$

The energy consumed by IT equipment is defined in the same way as for calculating PUE. As for the GHG emissions, the elements to take into account vary from case to case.

Indeed, in the initial definition by The Green Grid, all emissions associated with the total energy consumed by the data center (as defined for the PUE), and all greenhouse gases were considered.

As for the standard ISO/IEC 30134-8, 2 categories of CUE are defined :

• Category 1 (“basic”) only takes into account the CO₂ emissions (and no other greenhouse gas) associated with the electricity consumption of the data center.

• Category 2 (“intermediate”) takes into account the greenhouse gas emissions of all the energy consumed by the data center. This can include GHG emissions caused by refrigerant leakage for example.

CUE is expressed in kg CO₂/kWh (category 1) or in kg CO₂eq./kWh (category 2), and has an ideal theoretical value of 0 kg CO₂eq./kWh, indicating that no GHG emission is associated with the data center’s operations.

And if you hear about CUE category 3 (“advanced”), it is mentioned in the standard for future developments, but is not yet defined.

Interestingly, this indicator is seldom, if ever, used as it only accounts for a fraction of the data (not to mention the discussions regarding the methods and data used for the calculation).

WUE

WUE, for Water Usage Effectiveness, focuses on data center’s water use, and uses the same denominator as for PUE and CUE : the energy consumed by IT equipment.

$$ WUE = \frac{Water\ used\ by\ the\ data\ center}{Energy\ consumed\ by\ IT\ equipment} $$

This indicator was introduced in 2011 by The Green Grid⁽⁷⁾ and was later standardized in 2022 : ISO/IEC30134-9 and EN 50600-4-9.

The energy consumed by IT equipment is defined in the same way as for calculating PUE and CUE. As for the water used by the data center, the elements to take into account vary from case to case.

Actually, the standards define 3 categories of WUE :

• Category 1 (“basic”) takes into account all the water input of the data center.

• Category 2 (“intermediate”) takes into account all the water input of the data center, but excludes discharged water that is being reused for non-industrial purposes.

• Category 3 (“advanced”) takes into account all the water input of the data center as well as the water consumed to produce the energy powering the data center. However, rainwater and discharged water reused for both industrial and non-industrial purposes are excluded.

WUE, as defined in the above-mentioned standards, is expressed in m³/MWh (the version of WUE proposed in 2011 by The Green Grid is expressed in L/kWh). It has an ideal theoretical value of 0 m³/MWh, indicating that no water use is associated with the data center’s operations.

💡 Tip !

Although they are rarely published, PUE, CUE and WUE indicators can help choose a more sustainable hosting provider. However, one indicator alone doesn’t provide enough information to assess the environmental impacts of a data center. For example, PUE can be less important than the carbon intensity of the electricity mix when choosing your hosting provider. Similarly, it is useful to consider the level of water stress in the area of the data center in addition to the WUE for example.

(1) IEA (2025), Energy and AI, IEA, Paris https://www.iea.org/reports/energy-and-ai, Licence: CC BY 4.0
(2) IEA, World Energy Statistics and Balances, IEA, Paris https://www.iea.org/data-and-statistics/data-product/world-energy-statistics-and-balances, Licence: Terms of Use for Non-CC Material
(3) Malone, C., & Belady, C. (2006, September). Metrics to characterize data center & IT equipment energy use. In Proceedings of the Digital Power Forum, Richardson, TX.
(4) The Green Grid (2007, February 20). Green grid metrics: describing datacenter power efficiency. Technical committee white paper.
(5) Joint Research Center (2023)
(6) The Green Grid (2010, December 02). Carbon Usage Effectiveness (CUE): A Green Grid Data Center Sustainability Metric. White paper #32
(7) The Green Grid, (2011). Water usage effectiveness (WUE™): a Green Grid data center sustainability metric. White paper #35.

© Cover picture from vectorjuice, Freepik.