Nvidia has unveiled a new generation of AI servers capable of running on coolant hotter than a typical hot tub, a design choice the company says marks one of the biggest efficiency leaps in data center history and could eliminate water usage.
The company’s Rubin generation of AI infrastructure is the world’s first to achieve 100% liquid cooling, with every chip and networking component cooled entirely by liquid in a closed loop with no fans anywhere in the system, claimed Nvidia.
The cooling liquid can run at temperatures up to 45 degrees Celsius, or 113 degrees Fahrenheit — hotter than a hot tub, which typically sits between 38 and 40 degrees Celsius. That higher ceiling is the key to the system’s efficiency gains, as it allows facilities to rely on outdoor air rather than energy-intensive mechanical chillers for much of the year.
Water Use Near Zero
According to Nvidia, the design has zero water consumption, eliminating massive amounts of power usage and nearly all water usage, since the closed-loop system depends on dry coolers rather than evaporative cooling towers except for roughly one percent of the year in some climates.
Water usage has been a hot topic in the AI data center world, but the numbers may surprise you.
According to the Manhattan Institute, data centers use 0.2 percent of daily water usage in the U.S. and that number has dramatically decreased in the past few years due to a new… pic.twitter.com/QnlGrLR5ks
— NVIDIA (@nvidia) June 22, 2026
Ali Heydari, the company’s director of data center cooling and infrastructure, said the approach effectively removes water from the cooling equation altogether.
In numerical terms, the architecture can cut facility cooling water use from roughly 2.6 million gallons per megawatt per year under conventional cooling-tower systems to near zero, a reduction of up to 100% in favourable climates.
ALSO READ: Revenue, Subscriptions And Privacy: The Core Challenges Before New WhatsApp Chief Kunal Shah
Cooling has historically devoured a huge share of data center power, and industry estimates suggest raising chiller plant temperatures by just one degree can cut cooling energy costs by about 4%, with a 50-megawatt hyperscale facility saving more than $4 million annually in cooling-related energy and water costs by switching to liquid-cooled infrastructure.
Quieter, Denser, Different
The shift also changes the physical footprint and feel of data centers. Coolant made of 75% water and 25% propylene glycol flows through cold plates sitting directly on processors, removing the need for the noisy fans and rigid hot-aisle, cold-aisle layouts that define traditional facilities.
Richard Whitmore, president and CEO of Motivair, Schneider Electric’s advanced cooling division, said liquid cooling became unavoidable once chip power density crossed a certain threshold.
Nvidia frames the move as essential to keeping AI’s energy footprint in check as compute demand continues to outpace nearly every other category of infrastructure investment.
Heat Reuse Opens New Possibilities
Beyond efficiency gains, the new model creates an opening for waste heat recovery, where residual heat from AI factory operations can be repurposed to warm nearby commercial or residential buildings.
ALSO READ: Cyber Attack On Tata Electronics Threatens To Expose Apple Data, Tesla’s ‘Trade Secret’
That turns what was once a pure energy cost into a potential community asset, with data centers doubling as heat sources for surrounding infrastructure rather than simply consuming power and water.
An Industry-Wide Shift
Because the Rubin platform is built entirely around liquid cooling, every cloud provider and data center operator building on it is being pulled into the same transition.
Nvidia says the geography still matters — a facility in a cold climate has an easier path to chiller-less operation than one in the desert — but even in warmer regions, the move toward 45-degree coolant brings operators closer to eliminating mechanical refrigeration altogether, with chillers needed only on the hottest days of the year.