Quick Answer
Water usage effectiveness (WUE) is the data center industry’s standard water-efficiency metric, introduced by The Green Grid in 2011 in White Paper #35. The formula: annual site water usage in liters divided by IT equipment energy in kilowatt-hours, expressed in L/kWh. The ideal value is 0.0 — no water used. Real-world numbers span a wide range: published industry averages run about 1.8 to 1.9 L/kWh depending on the survey (Data Center Knowledge and the Environmental and Energy Study Institute, both 2025), while best-in-class operators report far less — Microsoft’s fleet averaged 0.27 L/kWh in FY25 and Meta reported 0.19 L/kWh for 2024. Calculating WUE accurately requires dedicated water metering, and those same meters can document sewer credit claims — so measurement often pays for itself on the utility bill.
What WUE Measures (And What It Doesn’t)
WUE answers one question: how much water does a data center use for every unit of computing work? The Green Grid’s White Paper #35 (2011) defines it as annual site water usage divided by IT equipment energy, in liters per kilowatt-hour. The denominator is the same IT energy figure used for PUE, which keeps the two metrics linked.
The numerator counts water used in the operation of the data center: humidification plus water evaporated on site for cooling or energy production — including cooling tower evaporation, blowdown, and drift, per the same white paper. It deliberately excludes lifecycle water, such as water used to manufacture servers or build the facility.
Unlike PUE, which has an ideal value of 1.0, WUE has an ideal value of 0.0 and no upper bound. A facility rejecting all heat with dry coolers can approach 0; a facility relying on evaporative cooling in a hot climate can exceed 2 L/kWh. For context on the raw volumes behind these ratios, see our pillar guide to how much water a data center uses.
The Formula in Practice: A Worked Example
Take a facility with a 5 MW IT load running at full utilization. Annual IT energy is 5,000 kW × 8,760 hours = 43.8 million kWh. Suppose its cooling towers and humidification consumed 22 million gallons over the year — about 83.3 million liters.
WUE = 83,300,000 L ÷ 43,800,000 kWh = 1.9 L/kWh — the top of the 1.8 to 1.9 L/kWh industry-average range reported by Data Center Knowledge and EESI (2025). Cut consumption to 8.8 million gallons (33.3 million liters) through higher cycles of concentration or partial dry cooling, and WUE drops to 0.76 L/kWh.
Two practical notes on the inputs. IT energy should come from the UPS output or busway metering, not the utility feed, or you will understate WUE by mixing in mechanical loads. And the water figure should cover the same 12-month window, because a summer-only sample will overstate annual WUE for any evaporative system.
Site WUE vs. Source WUE
The Green Grid actually defined 2 metrics. Site WUE counts only water used at the facility. Source WUE adds the water consumed off site to generate the electricity the facility buys: WUEsource = (EWIF × PUE) + site water per IT kWh, where EWIF is the energy water intensity factor of the local grid. White Paper #35 cited a US national average EWIF of 1.8 L/kWh, drawing on NREL’s power-plant water research.
That embedded water is not a rounding error — it usually dwarfs site water. Lawrence Berkeley National Laboratory’s 2024 United States Data Center Energy Usage Report estimates US data centers consumed 66 billion liters directly in 2023, but nearly 800 billion liters indirectly through electricity generation, at a national average of 4.52 L/kWh of indirect water per kWh consumed. A site that cuts its own water use by switching to energy-hungry dry cooling can raise total watershed consumption — a tradeoff The Green Grid flagged explicitly, and one reason Google (2022) notes water-cooled data centers use about 10% less energy than many air-cooled ones.
Industry Averages and Best-in-Class Benchmarks
Public WUE disclosures cluster into 2 groups: the broad industry average and the hyperscale leaders who design for the metric.
| Operator / benchmark | WUE (L/kWh) | Period | Source |
|---|---|---|---|
| Industry average (survey-dependent) | ~1.8–1.9 | 2025 | Data Center Knowledge; Environmental and Energy Study Institute |
| Microsoft global fleet (owned-and-operated) | 0.30 | FY24 | Microsoft datacenter sustainability fact sheets |
| Microsoft global fleet (owned-and-operated) | 0.27 | FY25 | Microsoft datacenter sustainability fact sheets |
| Meta data centers (withdrawal-based) | 0.19 | 2024 | Meta 2025 Environmental Data Index |
| Ideal value | 0.0 | — | The Green Grid, White Paper #35 |
Microsoft’s trajectory shows what design attention does: from a global average of 0.49 L/kWh in 2021 to 0.30 in FY24 — a 39% improvement, per Microsoft’s December 2024 announcement — with next-generation facilities designed to consume zero water for cooling.
Compare across companies carefully. Meta’s 2025 Environmental Data Index calculates WUE from water withdrawal, while others report on a consumption basis, and climate, cooling technology, and utilization all shift the number. A low WUE in a cool, humid region is not the same achievement as a low WUE in Arizona.
How to Measure WUE: Metering Comes First
WUE is only as good as the meters behind it, and most operators do not have them: a 2016 Uptime Institute analysis found that fewer than one-third of data center operators track water usage or use the WUE metric at all. A single utility meter is not enough, especially in mixed-use buildings where cooling water and domestic water share a service line.
Getting a defensible number means submetering cooling tower makeup and blowdown separately, then reading them on the same annual window as IT energy. Evaporation is the difference between the two, which also feeds the calculations behind where cooling water actually goes. Our submetering and sewer credits guide covers meter placement and sizing, and how RPM’s water monitoring works shows what continuous remote reading looks like in practice.
The Limits of WUE
WUE is useful, not sufficient. Know its blind spots before setting targets on it.
- It ignores off-site water by default. Site WUE excludes the roughly 4.5 liters of water embedded in each grid kWh (LBNL, 2024) unless you calculate the source variant.
- It has no scarcity weighting. A liter evaporated in a drought-stressed basin counts the same as a liter in a water-rich one. Google’s 2025 Environmental Report addresses this outside the metric, using a water risk framework that led it to pick air cooling at high-risk sites like Mesa, Arizona.
- Annual averages hide seasonal peaks. Evaporative systems consume the most in the hottest weeks, when watersheds are most stressed.
- Definitions vary. Withdrawal-based and consumption-based WUE figures are not directly comparable.
- It can fight your energy goals. Driving WUE to zero with dry cooling raises PUE and indirect water; the 2 metrics must be managed together, as The Green Grid warned when it introduced them.
What WUE Means for Your Water and Sewer Bill
For a cooling-tower facility, the water in your WUE numerator is mostly evaporation — water that leaves as vapor and never enters the sewer. Most commercial utilities still bill sewer charges on the full incoming meter, and sewer rates are frequently half or more of a combined water/sewer bill. That gap is money.
The same submeters that produce an auditable WUE also document how many gallons never reached the sewer, which is the evidence utilities require for evaporation credits. RPM Water Equity Solutions files and manages those claims with nearly 200 utilities across 36 states, so a facility that starts measuring WUE is usually one document package away from recovering sewer overpayments.
Ready to Find Out What You Could Save?
If your data center runs evaporative cooling, the water behind your WUE number is water you are likely paying sewer charges on — even though it never reaches the sewer. The meters that prove your WUE can also prove your credit claim.
Request your free assessment today and find out how much you could recover.
The Bottom Line
WUE is becoming what PUE was a decade ago: the number regulators, communities, and customers ask for first. Operators who install the metering now will be ready for disclosure requirements as they spread — and unlike most compliance spending, water metering pays back immediately through leak detection, efficiency gains, and recovered sewer overpayments. Measure it once, benefit 3 ways.
