Data-center cooling water use could reach 644 billion litres by 2030 without adaptation
DataNews Editorial Team
A Tech Insider analysis, citing Rystad Energy, estimates that global data centers used 222 billion litres of water for cooling in 2025. The report says this could rise to 644 billion litres by 2030 if operators do not adopt adaptive measures.
- Rystad Energy is cited as estimating global data-center cooling-water use at 222 billion litres in 2025.
- The same estimate is approximately 59 billion US gallons, not a separate consumption metric.
- Cooling-water demand could reach 644 billion litres by 2030 without adaptive measures.
- Closed-loop liquid systems, dry coolers, immersion cooling and software optimisation are among the approaches identified to reduce resource consumption.
- Water availability could become a more material consideration in AI data-center siting and cooling design.
Tech Insider has highlighted rising cooling-water demand in the global data-center sector as AI infrastructure scales. Citing Rystad Energy, the article estimates that data centers used 222 billion litres of water for cooling worldwide in 2025, equivalent to roughly 59 billion US gallons.
The article presents a conditional 2030 outlook: cooling-water use could reach 644 billion litres if adaptive measures are not deployed. This is a forecast rather than an observed consumption figure, and its outcome depends on whether operators change cooling and resource-management practices.
The technologies identified in the article include closed-loop liquid cooling systems, warmer coolant temperatures, dry coolers, immersion cooling fluids and software-based optimisation. These approaches can reduce a facility's dependence on water withdrawal or improve thermal-resource management, depending on the cooling architecture and operating conditions.
For AI-oriented data-center development, water efficiency could become a distinct infrastructure competitiveness parameter alongside power availability. The projected growth in cooling-water demand may increase attention to liquid and immersion cooling, dry cooling, and automated thermal controls. It could also complicate site selection and raise the importance of water-resource infrastructure and engagement with utilities for new facilities.
As AI workloads expand, water availability could become a design and operating constraint alongside power supply. The projected increase may strengthen the case for water-efficient cooling architectures, while making water access and utility engagement more consequential in site selection.
- — / October 23, 2026 · EnglishSource →
This article was prepared by the DataNews editorial team based on the sources listed above.