Artificial intelligence is increasing pressure on data centre cooling, making water monitoring a growing sustainability issue for the sector.
AMS Instrumentation & Calibration has highlighted ISOIL Industria analysis, warning that AI demand is creating an urgent and often hidden pressure on water.
The concern is not only the electricity required to run data centres. It is also the water used to cool the servers that process AI workloads.
ISOIL says the response starts with awareness, measurement and accountability across the integrated water cycle.
Why does data centre cooling use water?
Data centres generate large amounts of heat. Servers need constant cooling to prevent overheating and maintain performance.
According to ISOIL, many data centres use water cooling systems because they can be far more efficient than air cooling.
The article outlines several cooling methods.
Some systems spray water onto a heat-exchange surface and allow it to evaporate, lowering the temperature of the air that is then ducted through the facility. Cooling towers can also remove excess heat through evaporation. Other systems use chillers and closed-pipe circuits, in which water circulates and absorbs heat from equipment.
Each approach creates a need to understand how much water is being used, where losses occur and whether cooling systems are operating efficiently.
- Want more information on what Inside Water is up to? Sign up for our weekly email, landing in inboxes with the latest news.
- Do you want the magazine delivered to your letterbox? Sign up here to subscribe.
What is AI’s hidden water footprint?
The ISOIL article highlights estimates from research on the water footprint of AI queries.
It reports that asking ChatGPT 20 to 50 questions can consume about half a litre of water.
It also reports that if 100 million weekly users submitted one prompt each, weekly water consumption could reach up to five million litres.
AMS said projected water demand linked to AI could reach 6.6 billion cubic metres by 2027, including indirect water use associated with power generation. These figures underline why data centre cooling is becoming a water-sector issue, not only a technology-sector issue.
As AI workloads expand, water utilities, regulators, facility owners and technology companies will need clearer visibility over cooling-related demand.
Why does measurement matter?
ISOIL says full awareness of a facility’s water footprint is the first step in adopting efficiency solutions.
That requires measurement tools that can operate reliably in demanding industrial environments.
Real-time flowmeters can help data centres monitor consumption as it occurs. Leak detection systems can identify inefficiencies before losses increase. Water quality probes can help operators monitor cooling water and avoid contamination risks that could affect equipment or operator safety.
This shifts water management from a periodic reporting exercise to an operational control issue. It also provides facilities with better information for compliance, sustainability reporting, and water stewardship decisions.
What role can instrumentation play?
AMS said ISOIL is calling for smarter monitoring across the integrated water cycle. That includes tools that can track water use at the facility level and support better control of cooling systems.
In practical terms, data centres need to know how much water enters cooling processes, how much is lost through evaporation or leakage, and how water quality changes during operation.
Without those measurements, water efficiency claims are difficult to verify.
With real-time instrumentation, operators can identify abnormal consumption, detect leaks, optimise cooling performance and reduce avoidable waste. The same data can also support conversations with utilities and communities where data centre demand competes with other water needs.
Why is this a sustainability issue?
AI may improve productivity, but the infrastructure that supports it consumes physical resources. Data centre cooling links digital growth to local water availability, especially in regions already affected by drought, climate variability or competing demand.
ISOIL’s position is that innovation cannot be separated from accountability. If AI demand continues to increase, water use must be measured at the facility, cooling system, and process levels. That makes flow measurement, leak detection and water quality monitoring part of the sustainability response.
The core message is simple: data centres cannot manage the water they cannot see.
As AMS said, the future of water will depend on how industry measures, manages and protects every litre.
