Friday, August 7, 2026
DarkSubscribe
AI Infrastructure · News & Analysis
HomeChips & HardwareReport
Chips & Hardware · Report

NVIDIA's Rubin generation AI servers achieve 100% liquid cooling at 45 degrees Celsius, dramatically reducing data cente

NVIDIA official — first-hand confirmation of roadmap / product.
Official disclosureSlicast · July 1, 2026 · US · Source: NVIDIA Blog

NVIDIA has announced that its newest Rubin generation AI infrastructure represents the world's first AI compute platform with complete 100% liquid cooling — every chip and networking component cooled entirely by liquid in a closed loop with no fans anywhere in the system. This cooling methodology is outlined in the NVIDIA DSX AI factory reference design, which provides best practices for designing, building and operating entire AI factory infrastructure stacks.

The key innovation is raising cooling liquid temperatures to 45 degrees Celsius, or 113 degrees Fahrenheit — hotter than typical hot tubs but still well within the validated operating limits of silicon processors. "The NVIDIA DSX reference design for AI factories has zero water consumption — we have eliminated massive amounts of power usage and pretty much all water usage," said Ali Heydari, director of data center cooling and infrastructure at NVIDIA. "With dry-cooler-based designs, it's a closed-loop system with no evaporative water cooling — outside of maybe 1% of the year when we might need chillers in some climates."

Industry data shows cooling has historically accounted for up to 40 percent of data center electricity consumption. Raising chiller plant temperatures by just one degree can cut cooling energy costs by approximately 4 percent. At hyperscale, a 50-megawatt facility can save over 4 million dollars annually in cooling-related energy and water costs by moving to liquid-cooled infrastructure. In favorable climates, the 45-degree architecture enables chiller-less operation with dry coolers, reducing facility cooling water consumption from roughly 2.6 million gallons per megawatt per year for conventional cooling-tower systems to near zero — up to 100 percent reduction in water use.

The liquid cooling system uses a mixture of 75 percent water and 25 percent propylene glycol flowing through cold plates positioned directly on processors. When coolant enters a fully liquid-cooled chip at 45 degrees Celsius, it exits at roughly 55 degrees Celsius after absorbing heat across the chip surface, yet performance does not degrade. Processors continue operating at full performance because liquid-cooled cold plates maintain device temperatures within validated operating limits.

Additional benefits include reduced noise — air-cooled data centers generate cooling fan noise at or above 85 decibels, loud enough to require ear protection — higher rack density than air-cooled systems, and potential waste heat recovery for nearby commercial or residential buildings. A system that previously occupied six rack units now fits in two, providing more compute in less space. The fully sealed front panels of Rubin servers contrast with the perforated bezels of air-cooled servers, reflecting the fundamental architectural redesign required for complete liquid cooling across all components.

Read the original
NVIDIA's Rubin generation AI servers achieve… · Slicast