Wednesday · October 21, 2026

 ·  Daily  ·  Newsletters

Resonance network: Quantum InsiderSpace Insider

Nvidia Claims Its New Cooling System Eliminates Data Centre Water Use, But the Full Picture is More Complicated

Nvidia has unveiled a warm-water cooling system it says can eliminate virtually all water consumption inside AI data centres, with chief sustainability officer Josh Parker telling Axios that the water consumption challenge for data centres is largely solved. The system circulates coolant in a closed loop at 45°C through server racks, emerging at 55°C and dissipating heat via passive radiators without evaporative cooling or fans, meaning no new water is consumed on-site once the system is filled.

The technical achievement at the facility level appears genuine. In favourable climates, Nvidia says the approach can deliver a 100% reduction in on-site water use, while also improving energy efficiency and reducing noise.

However, the company’s accounting draws a boundary around the data centre walls and excludes the water consumed by the power plants supplying electricity. Research suggests that off-site water consumption from electricity generation and chip manufacturing can double or triple a facility’s total water footprint, meaning Nvidia’s solution may address only a quarter to a third of AI data centres’ actual water impact.

The scale of the omission is significant. Natural gas power plants consume 1.17 litres of water per kilowatt-hour generated, while coal plants use 2.2 litres. Fossil fuels currently supply roughly half of all data centre power globally, according to the International Energy Agency, which projects natural gas and coal will account for more than 40% of new electricity needed to meet data centre demand through 2030. Wind and solar, by contrast, use negligible amounts of water. Without a shift in the energy mix powering AI infrastructure, the broader water footprint of data centres will remain substantial regardless of what happens inside the building.

James Dargan
About the author
James Dargan

James Dargan is a writer and researcher at The AI Insider. His focus is on the AI startup ecosystem and he writes articles on the space that have a tone accessible to the average reader.

Trending today

an orange smile on a black background
Business & Markets · Enterprise

Amazon’s Panos Panay Bets on Ambient AI Over Apps as the Future of Personal Computing

Policy & Government · AI Safety

Anthropic Grants METR Researchers Employee-Level Access for Independent AI Risk Assessments

Technology & Infrastructure · Models & LLMs

Skyfall AI Unveils Frontier Engineering World Model SKY-001

A close-up of Benjamin Franklin on a United States one hundred dollar bill
Business & Markets

Fragment Raises $12M Series A to Turn Agents into Financial Operators

AI Agents
Technology & Infrastructure · Applications & agents

Google DeepMind AI Agents Beat Minecraft… After 27 Hours And 605 ‘Deaths’

The AI economy, every weekday morning

The daily briefing on LinkedIn. Free, one tap to follow.

Exclusives

Exclusive

South Korea’s AI G3 Strategy: Decoded

Scale-ups to Watch

10 Switzerland-Based AI Scale-Ups You Need to Know in 2026

network, blockchain, digital, hand, web, community, artificial, intelligence, steering, interfaces, bokeh, future, digitization, transformation, change, blockchain, blockchain, blockchain, blockchain, blockchain, transformation
Exclusive

Why Crypto Could Be AI’s Payment Layer: BlackRock Sees Stablecoins Connecting Commerce and Compute

AI Predictions
Exclusive

Why AI Predictions Often Get The Technology Right But The Timeline Wrong

Scale-ups to Watch

10 CEE & Baltics-Based AI Scale-Ups You Need to Know in 2026

More in Technology & Infrastructure

Latest from the same section
Business & Markets · Enterprise

Anthropic and Nscale Reportedly Near Second $45B AI Compute Deal

9 hours ago
Technology & Infrastructure · Data centres & compute

xMEMS Begins Mass Production of Solid-State Cooling Chips for Edge AI and Data Centers

13 hours ago
Technology & Infrastructure · Data centres & compute

Economist Estimates AI Buildout Would Need $3.55 Trillion in Annual Revenue

9 Oct 2026
Business & Markets · Startups

Veir Raises $110M in Series C Funding to Scale Superconducting Power Systems for AI Data Centers

8 Oct 2026

The AI economy, every weekday morning

The daily briefing plus the weekly Scale-ups to watch edition. Free, no spam, unsubscribe any time.