Elon Musk has backed NVIDIA’s claim that data center water use is far smaller than critics suggest.

Data centers face growing scrutiny over how much water and electricity they consume. NVIDIA says its latest cooling systems can nearly erase the water that older facilities lose to evaporation.

What NVIDIA Claims About Data Center Water Use

NVIDIA’s post cited a March 2026 estimate from the Manhattan Institute that data centers use roughly 0.2% of U.S. freshwater, most of it indirectly through power generation.

The company said its 45-degree Celsius liquid cooling lets AI factories in cool climates run on dry coolers rather than evaporative towers.

That shift can cut facility cooling water from about 2.6 million gallons per megawatt each year to near zero.

NVIDIA made the same case in 2025, claiming its Blackwell systems were 300 times more water efficient than air cooling.

Because cooling can reach 40% of a data center’s electricity, the design also trims power costs. Similar trade-offs now shape the global AI data center race.

Why Elon Musk’s Endorsement Matters

Musk, who runs large NVIDIA-powered clusters through xAI, has repeatedly praised NVIDIA’s latest chips as the backbone of his AI projects.

His endorsement supports NVIDIA’s push against the view that AI growth drains local water supplies. The company describes its system as a closed loop that recirculates coolant instead of consuming fresh water.

“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,” Ali Heydari, director of data center cooling and infrastructure at NVIDIA

Follow us on X to get the latest news as it happens

The Case for Caution

The national figure hides the detail. NVIDIA’s near-zero claim covers direct cooling, the smaller part of the footprint.

U.S. data centers used about 17.4 billion gallons of water directly in 2023, a Berkeley Lab report found, plus another 211 billion gallons indirectly through the electricity they drew.

That indirect share climbs as AI scales, and direct use alone is projected to reach 38 to 73 billion gallons by 2028.

Dry coolers also depend on climate, working best in cool regions and less well in hot, dry states.

The strain is visible at Musk’s own xAI. Its Memphis Colossus site has drawn roughly 1.3 million gallons of drinking water a day from the local aquifer and ran dozens of gas turbines before securing permits, prompting a data center pollution lawsuit and community appeals.

How regulators and water-stressed regions respond may decide whether efficiency gains keep pace with the industry’s expansion and the wider contest for AI capital.

Источник: BeInCrypto