Back to Data Center
Data Center tools

Data Center Water Usage Calculator (WUE): How Much Water AI Uses

Many data centers rely on evaporative cooling — cooling towers that evaporate water to dissipate heat from servers. It's highly effective, but it consumes enormous quantities of water in the process. A 100 MW AI campus can consume millions of gallons of water every day, making water use a growing concern for facilities built in hot, arid, or drought-stressed regions. Water Usage Effectiveness (WUE) measures how much water a facility consumes per unit of computing energy it delivers, and it has become a key sustainability metric alongside PUE. This calculator estimates a data center's daily and annual water consumption from its IT load and WUE — and puts the number in plain-English context.

IT load(kW)
Data center PUE 1.2
WUE — Water Usage Effectiveness 1.8L/kWh IT energy
Annual operating hours(hrs/yr)
Total facility power
1.20MW

IT load × PUE

Annual IT energy
8,760MWh

8,760,000 kWh/yr

Daily water use
11,412gal/day

43,200 liters/day

Annual water use
4.17M gal/yr

4,165,464 gal / 15,768,000 L

Homes equivalent
38US homes/day

at 300 gal/day avg household use

Results update live as you type. For planning and field-check estimates — always verify against applicable standards and equipment ratings.

How we calculate this →

WUE: measuring water the way PUE measures power

Water Usage Effectiveness (WUE) was introduced by The Green Grid — the same organization behind PUE — as a standard metric for data center water consumption. It is defined as annual site water usage (in liters) divided by IT equipment energy (in kWh). A WUE of 1.8 L/kWh means a facility consumes 1.8 liters of water for every kilowatt-hour of compute it delivers.

The calculation flows from IT load and operating hours to annual IT energy, then multiplies by WUE to get total water volume. Facility-level energy (IT load × PUE) determines the power draw from the grid, but WUE is specifically anchored to IT energy — not total facility energy — so a high PUE facility consumes more grid power but not proportionally more water; they are independent variables.

Evaporative cooling towers work by evaporating water to absorb heat. A 1 MW IT load running 24/7 at WUE 1.8 consumes roughly 1.8 × 8,760,000 Wh = 15.8 million liters (4.2 million gallons) of water per year — about 11,400 gallons per day, equivalent to the daily domestic water use of roughly 38 average U.S. households. Scale to 100 MW of IT load and the figure becomes 1.14 million gallons per day.

Closed-loop liquid cooling systems — where the same fluid is continuously recirculated through a chiller without evaporation — can approach WUE near zero because they recapture and reuse the cooling medium. Air-side economizers (using outside air directly) also reduce water use in cool, dry climates. These are the primary technical paths to reducing a data center's water footprint, which is why liquid cooling adoption matters not only for energy efficiency (PUE) but also for water sustainability (WUE) in water-stressed regions.

Frequently asked questions

A 100 MW IT load AI data center using evaporative cooling at a typical WUE of 1.8 L/kWh consumes roughly 1.14 million gallons of water per day — comparable to the daily domestic water supply of a city of about 15,000 people. Large hyperscale AI campuses planned at 500–1,000 MW of IT capacity can consume several million gallons per day. Microsoft, Google, and Amazon have all faced public scrutiny over data center water consumption in drought-stressed regions including the American Southwest, Chile, and South Africa.

WUE — Water Usage Effectiveness — is a standard metric defined by The Green Grid that measures how much water a data center consumes per unit of IT energy delivered. It is calculated as annual site water consumption (in liters) divided by annual IT equipment energy (in kWh). A lower WUE is better: a closed-loop liquid cooling system can approach 0 L/kWh, while a heavily evaporative-cooled facility in a hot climate might run 3.0–5.0 L/kWh. The industry-wide average is roughly 1.8 L/kWh. WUE was designed to complement PUE — PUE measures energy efficiency, WUE measures water efficiency.

Most large data centers use evaporative cooling — cooling towers or adiabatic coolers — because evaporation is one of the most thermodynamically efficient ways to move heat. Evaporating one gallon of water removes roughly 8,000 BTU of heat, which is far more than is achievable with air alone. The water is consumed in the process: it leaves as water vapor in the exhaust air rather than returning to the facility. Data centers can reduce or eliminate water consumption through several approaches: closed-loop water-cooled chillers (no evaporation — water is recirculated and cooled via a refrigerant cycle), air-side economizers (using outside air to cool servers directly in cool climates, with no water at all), or direct liquid cooling to chips (which moves heat into a recirculating loop rather than into the air). Each involves tradeoffs in cost, climate range, and energy efficiency, but all reduce WUE substantially compared to evaporative towers.