As artificial intelligence continues to expand, one concern has grown alongside it: the massive amount of water needed to cool data centres. Now, NVIDIA says it has developed a solution that could dramatically reduce that water consumption – up to 100%. The company has introduced a new “100% liquid-cooled” AI data centre design that aims to eliminate almost all water use inside data centre facilities.
Why Data Centres Use So Much Water
Modern AI data centres house thousands of powerful processors that generate enormous amounts of heat. Traditionally, operators rely on cooling towers, chillers and large amounts of water to keep equipment running safely. As AI workloads become more demanding, cooling systems have become one of the biggest consumers of energy and water in data centres. Industry experts estimate that cooling can account for a significant share of a facility’s total power consumption.
How NVIDIA’s New Cooling System Works
Instead of cooling servers with chilled air, NVIDIA’s new design uses a closed-loop liquid cooling system. Special coolant flows directly to the chips, capturing heat at the source and carrying it away far more efficiently than air.
Read More | India’s Data Centre Industry to Deploy Up to 700,000 GPUs by 2030
The liquid enters server racks at temperatures of around 45°C and exits at about 55°C after absorbing heat. Because the coolant operates at higher temperatures, the heat can be released outdoors using dry coolers rather than traditional water-intensive cooling towers. NVIDIA says the system recirculates the same liquid continuously, meaning it does not require a constant supply of fresh water for cooling.
Water Use Could Drop Close to Zero
According to NVIDIA, a conventional data centre may consume around 2.6 million gallons of water per megawatt each year for cooling. The company claims its new design can reduce on-site cooling water use to nearly zero in many environments. Josh Parker, NVIDIA’s Head of Sustainability, said the company believes the long-standing challenge of water consumption inside data centres is now “largely solved.”
Not a Complete Solution to AI’s Water Footprint
While the technology could significantly reduce water use inside data centres, experts note that it does not eliminate all water consumption associated with AI. Large amounts of water are still used indirectly to generate electricity and manufacture semiconductor chips. Some analysts estimate that these indirect sources can account for a substantial portion of the overall water footprint of AI infrastructure.
Read More | China Plans $295 Billion AI Data Centre Expansion to Challenge US Tech Dominance
As a result, NVIDIA’s breakthrough addresses one major part of the problem, but not the entire environmental impact of AI.
What It Means for the Future of AI
The announcement highlights a growing trend across the technology industry toward liquid cooling as AI systems become more powerful and energy-intensive. By reducing dependence on cooling towers, fans, and chillers, liquid cooling can improve efficiency while lowering water consumption. NVIDIA believes this approach will become a key part of future AI factories and next-generation data centres.
As demand for AI continues to surge worldwide, innovations like these could help data centre operators balance growing computing needs with increasing environmental concerns.
Indrani Priyadarshini is a journalist and editorial professional specialising in technology, artificial intelligence, smart cities, green energy, and digital transformation. With over four years of experience in tech journalism and digital media, she is known for turning complex industry developments into clear, engaging, and insightful stories. Her expertise spans reporting, editorial strategy, digital publishing workflows, and in-depth coverage of emerging technologies shaping the future.
