Nvidia (NASDAQ: NVDA) announced at GTC Berlin that European cloud providers and infrastructure companies are adopting its DSX AI Factory Platform, a full-stack reference architecture that integrates compute, networking, storage, power and cooling into a single codesigned system.

The platform uses closed-loop liquid cooling paired with dry coolers to eliminate water evaporation during heat rejection. Nvidia said that at a 2-gigawatt deployment scale, this approach could save 54 million liters of cooling water per day compared with conventional cooling-tower systems, an amount it equates to the daily water consumption of 180,000 households. The company's Vera Rubin infrastructure supports 100% liquid cooling at coolant temperatures up to 45 degrees Celsius.

Nvidia's DSX MaxLPS software dynamically allocates power within a provisioned budget, which the company said can deliver the same token output using up to 30% less power. In a proof of concept using Nvidia HGX B200 systems, Lambda demonstrated 24% more cluster token throughput and 19% more GPU capacity within a fixed power budget. Nscale reported 49% higher inference throughput and 37% more managed GPU capacity across four Nvidia GB300 NVL72 racks within a 264-kilowatt power budget. CoreWeave, Firmus, Google Cloud, IREN, Lambda, Nebius, Nscale and Sharon AI are among the companies exploring DSX MaxLPS.

A separate software component, DSX Flex, enables AI factories to reduce power draw in response to grid conditions. In an early trial involving Emerald AI, EPRI, Nebius, the U.K. National Grid and Nvidia, the system demonstrated reductions of up to 40% in cluster power draw on a Blackwell Ultra GPU cluster during simulated grid events.

Nvidia also launched the SimReady Central Library on Hugging Face, a repository of simulation-ready digital models of physical equipment to accelerate AI factory digital twin creation. Additionally, three new open-source DSX OS components entered developer preview: Nvidia Config Manager, Nvidia Rack Management Service and Nvidia Cluster Readiness Engine.

Ecosystem partners involved in the buildout include Schneider Electric, Siemens, Dassault Systèmes, AVEVA, ABB, Eaton, GE Vernova, Hitachi Energy, Siemens Energy, Trane Technologies and Vertiv.