Nvidia Says Its New AI Cooling Tech Can Bring Water Use Close To Zero. (Image Internet)
New Delhi: Nvidia has announced a cooling design for AI data centres that it says can sharply reduce water consumption as artificial intelligence infrastructure faces scrutiny over energy and environmental impact.
The company’s Rubin-generation infrastructure uses a liquid-cooled system designed to remove heat directly from chips and other components. Instead of depending on air-cooling systems, fans and water-heavy cooling towers, the design uses sealed cold plates through which coolant flows continuously.
The coolant absorbs heat from the hardware and moves through a closed-loop system, meaning it is reused rather than consumed. Nvidia says this approach can help AI data centres operate with far lower dependence on fresh water during normal conditions.
A part of the design is the ability to run the cooling liquid at temperatures as high as 45 degrees Celsius, or 113 degrees Fahrenheit. Nvidia argues that this higher operating temperature improves efficiency because the system does not need as much energy-intensive chilling equipment to keep servers cool.
The company has compared the temperature to a hot tub, saying its newest AI servers can safely operate with coolant even warmer than typical hot-tub water.
Nvidia says its DSX AI factory reference design is built around a fully closed-loop cooling system. In favourable climates, heat can be rejected through dry coolers for much of the year, reducing or nearly eliminating the need for water-based cooling.
Ali Heydari, Nvidia’s director of data centre cooling and infrastructure, said the DSX reference design has “zero water consumption” and cuts power usage linked to cooling.
The announcement comes as global cloud and AI infrastructure faces criticism over resource consumption. Amazon recently disclosed that its global data centres used 2.5 billion gallons of water in 2025, while saying its facilities remain more water-efficient than the industry average.
Nvidia’s latest approach points to a future where powerful AI systems could run with less pressure on local water resources. However, broader concerns around electricity demand, construction impact and AI expansion are likely to remain part of the debate.
