Nvidia’s hotter AI data center design cuts water usage

Public concerns over the environmental footprint of data centers have increasingly zeroed in on their massive water and energy consumption. Now, Nvidia is stepping forward with a bold claim: its Rubin generation reference design for a fully liquid-cooled data center has “eliminated massive amounts of power usage and pretty much all water usage.” While this sounds like a major leap forward, it doesn’t fully address the broader worries surrounding AI data centers, including the environmental impact of their construction and the enormous power generation needed to run them. As Gizmodo notes, Nvidia’s own blog post remains silent on the cost difference between building this style of facility versus one that relies on less efficient air cooling. Still, the company asserts that “every cloud provider and data center operator building for [Rubin] is making the transition.”

A key driver of these efficiency gains is running AI servers at higher temperatures, up to 113 degrees Fahrenheit (45 degrees Celsius). Amazon recently made a similar pitch, highlighting higher heat tolerances as a way to boost efficiency in its predominantly air-cooled data centers.

With Nvidia’s approach, “heat is captured directly at the chip and transported through liquid loops operating at much higher temperatures, allowing outdoor dry coolers to reject heat efficiently for much of the year.” This design offers far more flexibility with ambient air temperature, reducing reliance on traditional cooling methods.

Josh Parker, Nvidia’s head of sustainability, claims the reference design slashes water usage “from roughly 2.6 million gallons per megawatt per year for conventional cooling-tower-based systems to near zero , up to a 100 percent reduction.” That’s a dramatic shift, but it’s only one piece of a much larger puzzle.