Data centre water usage evaluates direct water withdrawal for facility operations, separating on-site measurement through water usage effectiveness from indirect power generation consumption.
How much water do data centres use?
Data centre water usage covers water withdrawn for cooling and, indirectly, the water required for power generation. Intake can include potable, recycled, or alternative supplies. Consumed volume differs from gross withdrawal: the U.S. Geological Survey defines consumptive use as water withdrawn that is evaporated, incorporated, or otherwise removed from the local water environment. Operational demand depends on cooling design, climate, and seasonal load cycles.
That figure splits into two parts. Direct water relates to on-site cooling, where water withdrawn can differ from net water consumed depending on the operator reporting boundary and source type, including potable or recycled supplies. Indirect water is consumed upstream during power generation. Both require precise reporting boundaries to evaluate.
Peak and daily water requirements vary between facilities according to campus scale, heat rejection methods, and seasonal weather patterns. Evaluating industry figures requires defined measurement boundaries across jurisdictions and reporting standards. Technical assessments evaluate operator reporting boundaries, separating total intake from net water consumed while identifying water sources, such as potable or recycled supplies, alongside specific operating periods.
Why do data centres need so much water?
Servers generate heat. Operating hardware beyond safe thermal thresholds causes throttling or failure. Cooling is essential to keep the facility running. Water is an effective medium to reject heat, though the chosen method depends on facility design, thermal requirements and local utility conditions.
One common method is evaporative cooling. Warm air or water passes through a cooling tower, where evaporation transfers heat away. Evaporated water represents net consumption rather than simple withdrawal, meaning it leaves the local system. Determining the impact requires verifying whether the intake uses potable, recycled or other water sources.
This is why a hot, dry region can be both attractive and risky for an operator. Dry air makes evaporative cooling more effective, yet dry regions are exactly where water is scarce. The engineering logic and the community's water security pull in opposite directions.
What is water usage effectiveness (WUE)?
Water usage effectiveness measures annual cooling and humidification litres divided by annual IT-equipment kilowatt-hours, as defined in Microsoft reporting. A lower ratio reflects less water per unit of IT electricity, not delivered computational output or total facility water performance.
Water usage effectiveness, or WUE, tracks site water use against IT energy consumption, though calculation boundaries differ across operators. For example, Microsoft calculates WUE by dividing annual litres of water used for humidification and cooling by the annual kilowatt-hours used to power IT equipment. WUE sits alongside power usage effectiveness, where balancing water draw against electrical consumption depends on evaporative, dry, or hybrid cooling selections under local climate conditions.
Site WUE, an on-site metric, measures water use relative to IT energy over a defined reporting period. It requires a specified reporting boundary for water withdrawn or net water consumed, whereas assessing indirect water consumed during electricity generation requires reviewing the off-site power supply.
How do AI data centres change water demand?
High-density computing for artificial intelligence training and inference produces elevated thermal loads compared with conventional architectures. Concentrating compute power into compact physical footprints increases heat intensity per square metre. Facility teams must review whether specific thermal management strategies, heat transfer designs, or secondary cooling systems are needed across the data hall to support these higher rack densities without assuming identical water demands.
While internal direct-to-chip or immersion loops recirculate liquid, primary heat rejection may still evaporate water. Buyers must verify the complete heat-rejection design, external heat exchangers and the measured reporting boundary rather than assuming total facility water savings.
The relationship between high-density workloads and water use depends on mechanical architecture and local conditions. Although dense computing increases heat output, cooling designs differ in water requirements. Internal closed-loop configurations isolate internal fluids, but the primary heat rejection plant may or may not use evaporative cooling at the facility boundary. Total water impact must be evaluated through the chosen cooling system, ambient climate, and indirect water use associated with grid electricity generation.
Can data centres reduce their water use?
Yes, and several proven approaches are in use. The choice usually comes down to cost, climate, and local water availability.
Procurement teams should examine external heat rejection, not just internal loops. Key checks include whether facilities draw potable or recycled water, local utility supply availability, seasonal reliance on evaporative chilling, and how the operator measures on-site water withdrawal and consumption.
Each cooling design involves technical trade-offs. Air-based systems can increase electricity demand, affecting indirect water use depending on the power generation mix. Recycled water requires accessible local infrastructure. Determining the right approach requires evaluating utility capacity, water source types and operational requirements for the specific site.
How does water usage affect where data centres get built?
Water access is a key site-selection factor alongside power and connectivity. When assessing an asset, critical questions include whether utility infrastructure supports expected cooling volumes, whether recycled water is accessible, and how withdrawal limits affect facility design. Verifying infrastructure capacity early provides clarity on technical feasibility.
Site feasibility due diligence examines applicable utility capacity, planning controls, water allocation rules and catchment drought conditions. Verification questions assess whether projected facility water demand creates supply constraints, planning exposure or operational risk for the site.
When acquiring, disposing of or developing a data centre asset, buyers and their technical advisers evaluate utility supply terms, withdrawal boundaries and recycled water viability. Data Centre Axis reviews commercial requirements and sourced information to support acquisition mandates, disposals and site sourcing, while technical assessments are handled by separately scoped specialists.
Frequently asked questions
How much water does a data centre use per day?
Daily cooling water requirements vary according to IT capacity, cooling architecture, facility size, and local climate. Water intake fluctuates across seasons, with peak demand typically linked to higher ambient temperatures and the operational profile of evaporative heat rejection systems on site.
Do all data centres use water for cooling?
Not necessarily. Water use varies by cooling design, climate and reporting boundaries. Evaporative systems consume water through phase change, whereas air or closed-loop designs have different consumption profiles. Total impact depends on both on-site cooling and indirect power generation.
What is a good WUE for a data centre?
A lower WUE figure reflects reduced water use relative to IT energy consumption within a given reporting boundary. Specific targets vary with regional climate, cooling methodology, and utility resources. Because reducing water use can increase electrical draw, operators assess WUE alongside PUE to balance site infrastructure requirements and local utility conditions.
Does AI use more water than normal computing?
Key procurement questions address actual workload IT electricity, complete facility heat rejection, the nominated water source and the applicable reporting period.
Is data centre water use regulated in Australia?
Water availability and allocation considerations vary across utility jurisdictions and local catchments. Planning evaluations often review projected water intake and discharge requirements alongside broader infrastructure capacity. Project teams must determine whether planning frameworks require evidence on alternative sourcing, recycled water access, or specific heat rejection designs when assessing utility connections and site feasibility.
