344
Productivity & Workflow355
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative169
Customer Support131
Photography & Imaging156
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Nvidia says a new warm-water cooling system can eliminate most on-site data center water use, but it does not reduce water used by power plants.
In short: Nvidia announced a cooling system that can cut water use inside AI data centers, but it does not address water used to generate the electricity those centers run on.
Nvidia said it has a new “warm-water” cooling system for data centers, which are large buildings filled with computers that run online services and many AI tools. The company says the system can eliminate “pretty much all” water use inside the data center itself.
The idea is to circulate liquid coolant in a closed loop (like the same water moving around in your car’s radiator). Nvidia says the coolant is filled once and then reused for the life of the facility. In the right climates, the company says this can mean a 100% reduction in on-site water use.
The coolant runs into server racks at about 45°C (113°F) and comes out at about 55°C (131°F), carrying heat away from the chips. At those temperatures, outside air can often remove the heat with radiators, without systems that evaporate water, and sometimes without fans.
Cutting on-site water use is helpful, especially in places where water is scarce. But the biggest water use linked to AI can happen outside the data center, mainly at power plants that generate electricity and also in chip manufacturing.
Many fossil fuel power plants use large amounts of water for cooling. TechCrunch notes that these “outside the walls” sources can double or triple a facility’s total water footprint. That means a fix that only counts water used inside the building may address only about a quarter to a third of the overall water use tied to an AI data center.
Source: TechCrunch AI