NVIDIA and European partners build AI factories that use less water and power

what happened

NVIDIA said European technology and infrastructure companies are adopting its DSX platform to build AI factories. The announcement was made on October 21 at the GTC conference in Berlin. According to a press release published by GlobeNewswire, the platform is designed to reduce cooling water use and electricity consumption while delivering the same computing output.

The company says a closed-loop liquid cooling system with dry coolers can cut facility cooling water use to near zero. At a 2-gigawatt scale, the savings would reach 54 million liters a day, equal to the daily water use of 180,000 households. NVIDIA says its DSX MaxLPS software can produce the same number of tokens with up to 30% less power.

Pilot projects also reported higher performance. Lambda saw 24% more cluster token throughput within a fixed power budget, and Nscale recorded 49% higher inference throughput. In a trial with grid signals, Emerald AI, EPRI, Nebius, the U.K. National Grid and NVIDIA reported cluster power draw reductions of up to 40%.

why it matters

For markets, this points to a shift in data center priorities, with power and cooling becoming the main constraints on AI infrastructure growth. Demand for grid and climate equipment could rise, which matters for suppliers such as Vertiv, Schneider Electric, GE Vernova and Siemens Energy. The news is a partnership announcement rather than financial results, so its effect on share prices should be judged with caution.

market impact

NVDA and data center equipment makers: stronger demand for cooling and power systems used in AI factories

tags

instruments