Nvidia says that some of Europe’s leading ecosystem partners, including CoreWeave, Firmus, Nscale and Google Cloud, are planning to adopt Nvidia DSX AI Factory Platform to build AI factories that minimise water consumption for cooling and use less power for the same AI output.
The platform also helps partners bring new capacity online faster and adapt workload power consumption to support grid reliability.
Nvidia DSX is a full-stack AI factory reference architecture, co-innovated with the world’s leading infrastructure companies. Compute, networking, storage, software, power and cooling are codesigned as one system – optimised from design to operation.
AI clouds and leading design, construction, energy and cooling providers are using DSX to build and operate infrastructure serving Europe’s enterprises, developers, and governments.
“Europe’s technology leadership depends on how efficiently we turn power and infrastructure into intelligence,” says Warren Barkley, vice-president of DSX product management at Nvidia. “Nvidia DSX brings together the technologies and ecosystem expertise operators need to build responsibly, operate efficiently at scale, and bring intelligence to the market faster.”
The Nvidia DSX platform includes DSX MaxLPS for dynamic power allocation, DSX Flex for coordinating workload power consumption with grid conditions, DSX Reference Design for validated infrastructure architectures, DSX Sim for simulation, DSX OS for infrastructure management, and DSX Exchange for connecting computing and facility systems.
These technologies help operators design, deploy and run AI factories that save valuable resources.
Rethinking cooling
Conventional cooling towers remove heat by evaporating water, requiring a continual supply of replacement water. Nvidia DSX reference designs connect liquid-cooled computing with dry-cooler-based heat rejection that avoids evaporation: Cooling liquid absorbs heat from computing equipment and circulates in a closed loop, while outdoor dry coolers – similar to large radiators – release that heat into the air without evaporating water.
Nvidia Vera Rubin infrastructure supports 100% liquid cooling with coolant temperatures of up to 45 degrees Celsius. With dry-cooler-based designs, this architecture can reduce facility cooling water consumption to near zero. At 2-gigawatt scale, Nvidia DSX can save 54-million liters of water a day compared with conventional cooling-tower-based systems – equivalent to the daily water use of 180 000 households. Local climate determines whether supplemental mechanical cooling is needed during the hottest periods.
Delivering the same AI output with less power
Power availability defines how much capacity operators can bring online. DSX MaxLPS and DSX Flex address this constraint.
DSX MaxLPS dynamically allocates power within an AI factory’s provisioned budget, helping operators generate the same number of tokens with up to 30% less power. Operators can also use this efficiency to support more GPU capacity and higher throughput within existing power budgets.
CoreWeave, Firmus, Google Cloud, IREN, Lambda, Nebius, Nscale and Sharon AI are exploring the software.
In a proof of concept across NVIDIA HGX B200 systems with DSX MaxLPS, Lambda demonstrated 24% more cluster token throughput and 19% more GPU capacity within a fixed power budget.
Coordinating AI factory power demand with the grid
DSX Flex helps AI factories become flexible assets for the power grid, enabling operators to adapt workload power consumption to utility needs, demand-response events, and energy prices. It can also coordinate electricity from the grid, onsite renewables and energy storage, helping AI factories operate as flexible power loads that support grid reliability and enable faster grid connection.
In an early trial, Emerald AI, EPRI, Nebius, the UK National Grid and Nvidia demonstrated reductions of up to 40% in the cluster’s power draw on an Nvidia Blackwell Ultra GPU cluster in response to simulated grid events, while critical workloads continued running.