MyTechDailyNews General: How NVIDIA DSX AI Factory Platform Reframes Europe’s AI Infrastructure for Water, Power, and Grid Flexibility

On MyTechDailyNews General, the interesting part of this AI-infrastructure update is not a single device. It is the shift toward planning whole facilities as connected systems: water, power, cooling, simulation, and operations are discussed together. The update centers on NVIDIA DSX AI Factory Platform, with NVIDIA saying Europe’s ecosystem partners are using it to reduce cooling water use, consume less power for the same AI output, bring capacity online faster, and let workload power consumption support grid reliability.

From parts to one platform

Instead of treating servers, cooling, and power as separate purchases, NVIDIA DSX is positioned as a full-stack AI factory reference architecture that was co-innovated with infrastructure companies. In the company’s description, compute, networking, storage, software, power and cooling are codesigned as one system, optimized from design to operation.

Across Europe’s regions, AI clouds and design, construction, energy, and cooling providers are using DSX to operate infrastructure for enterprises, developers, and governments.

The NVIDIA DSX platform groups several named elements:

• DSX MaxLPS: dynamic power allocation inside the AI factory.
• DSX Flex: workload power consumption coordinated with grid conditions.
• DSX Reference Design: validated infrastructure architectures.
• DSX Sim: simulation for design and operation questions.
• DSX OS: infrastructure management software.
• DSX Exchange: connectivity between computing and facility systems.

Cooling water becomes a design choice

Conventional cooling towers remove heat by evaporating water, which creates a continuing demand for replacement water. DSX Reference Design approaches instead connect liquid-cooled computing with dry-cooler-based heat rejection, so heat leaves the site without evaporative loss. In that closed loop, liquid absorbs heat from computing equipment, and outdoor dry coolers—like large radiators—release the heat into air.

With NVIDIA Vera Rubin infrastructure, the design envelope includes 100% liquid cooling and coolant temperatures of up to 45 degrees Celsius. Where dry coolers replace evaporative towers, facility cooling water consumption can become near zero. The example in the materials is 2-gigawatt scale: NVIDIA DSX can save 54 million liters of water a day relative to conventional cooling-tower-based systems, an amount equal to the daily water use of 180,000 households; local climate still determines whether supplemental mechanical cooling is needed during the hottest periods.

Power, not only space, sets the growth limit

Operators often find that available power—not floor space—decides how much capacity comes online. The platform’s answer is to make the electrical budget smarter. 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. That efficiency can also support more GPU capacity and higher throughput within existing power budgets.

CoreWeave, Firmus, Google Cloud, IREN, Lambda, Nebius, Nscale and Sharon AI are exploring that approach. In a proof of concept across NVIDIA HGX B200 systems, Lambda used DSX MaxLPS and reported 24% more cluster token throughput and 19% more GPU capacity within a fixed power budget.

Grid-responsive AI factories

DSX Flex is designed to make AI factories flexible assets for utilities. Operators can 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 so the facility behaves as a flexible power load, supporting grid reliability and enabling faster grid connection.

During a simulated grid event, Emerald AI, EPRI, Nebius, U.K. National Grid, and NVIDIA tested an NVIDIA Blackwell Ultra GPU cluster and saw reductions of up to 40% in the cluster’s power draw were possible while critical workloads continued running.

Nscale: a concrete efficiency picture

Nscale’s demonstration links power optimization, grid-responsive operations and design inside the NVIDIA DSX platform by using DSX MaxLPS, DSX Flex and DSX Sim.

At an Nscale AI factory in Iceland, an evaluation run used DSX MaxLPS across four NVIDIA GB300 NVL72 racks. Within the same 264-kilowatt provisioned power budget, the run showed 49% higher inference throughput and 37% more managed GPU capacity, suggesting that dynamic power allocation can unlock additional capacity from existing infrastructure.

Nscale is also exploring DSX Flex as a way to align workload power consumption with grid conditions.

For facilities in Portugal and Iceland, Nscale uses DSX Sim to create digital twins that evaluate power, cooling and infrastructure configurations before deployment. Before equipment is installed, those digital twins help Nscale spot potential issues earlier, lower deployment risk, and speed the move from planning to productive capacity.

Simulation assets and design speed

The SimReady Central Library is introduced as a collection of validated power and cooling equipment assets from partners like Vertiv, accelerating creation of DSX AI factory digital twins. Ecosystem partners including PTC are helping create simulation-ready assets with computer-aided design; used with DSX Sim, those SimReady assets let builders model infrastructure configurations and evaluate power and cooling decisions before equipment is installed. Earlier validation can reduce deployment risk and move from planning to construction faster.

NVIDIA DSX OS and production readiness

NVIDIA is expanding its modular, open source NVIDIA DSX OS portfolio with three components in developer preview to help operators turn installed infrastructure into productive capacity:

• NVIDIA Config Manager automates network configuration for AI factories.
• NVIDIA Rack Management Service provides rack-level management of power, inventory, firmware and switches.
• NVIDIA Cluster Readiness Engine runs distributed workloads before production workloads arrive to validate GPU clusters and isolate failures to specific nodes.

These previews sit alongside NVIDIA Infrastructure Controller, NVIDIA AI Cluster Runtime and NVSentinel, supporting automated infrastructure lifecycle management, consistent runtime environments, and stronger health validation, fault attribution, and recovery.

Aleria, Cassava, CoreWeave, Crusoe, Firmus, Firebird, G42, GMI Cloud, IREN, Lambda, Nebius, Nscale, Radiant, Sharon AI and Volta are cloud partners exploring the latest components to improve resiliency and reliability.

Europe’s partner stack: design, power, and cooling

NVIDIA’s industrial software, power and cooling partners are advancing Europe’s AI factory buildout through DSX Reference Design work, lifecycle digital twins, and qualified equipment that connect facility infrastructure with accelerated computing.

• Schneider Electric has developed power and cooling reference designs for NVIDIA Vera Rubin infrastructure and is working with AVEVA on digital twins across infrastructure design, construction and operations.
• Procore has integrated its construction management platform with DSX Sim to help manage AI factory builds.
• Siemens uses DSX Sim to simulate power, cooling and acoustics, while Dassault Systèmes and Vertiv develop digital twin models of Vertiv’s OneCore infrastructure for simulation and optimization through model-based systems engineering.
• ABB is developing SimReady assets with NVIDIA DSX Sim, enabling engineers to evaluate source-to-rack architectures before construction and advancing 800 VDC development through Infinitus Direct Current.
• Trane Technologies and Eaton collaborated with NVIDIA on an integrated design for NVIDIA Vera Rubin DSX AI Factory Reference Design, combining Eaton’s Beam platform and power infrastructure with Trane Continuum cooling and coordinated controls in a grid-to-chip architecture.
• Trane Technologies published seven thermal management system reference designs, including three 250-megawatt DSX-based cooling reference designs.
• Through NVIDIA DSX Ready, builders can select qualified equipment. Vertiv’s 2.3-megawatt CoolChip and Trane Technologies’ LiquidStack GigaModular coolant distribution units have qualified, while Hitachi Energy’s battery energy storage system earned NVIDIA DSX Ready BESS qualification.
• GE Vernova, Hitachi Energy and Siemens Energy are collaborating with NVIDIA on power infrastructure and digital twin initiatives for energy supply and grid integration. Hitachi Energy is developing an integration between its energy management system and DSX Flex to optimize grid, generation and storage resources with workload demand.

FAQ for MyTechDailyNews General readers

Is this only a cooling story?

No. Cooling water savings are one visible effect, but NVIDIA DSX AI Factory Platform is presented as a reference architecture spanning compute, networking, storage, software, power and cooling.

Does near-zero cooling water mean no hot-weather backup?

Not necessarily. The materials say dry-cooler-based designs can reduce facility cooling water consumption to near zero, but local climate still determines whether supplemental mechanical cooling is needed during the hottest periods.

Will grid flexibility interrupt AI work?

The trial evidence says reductions of up to 40% in the cluster’s power draw were possible while critical workloads continued running.

Checklist: what to notice in AI infrastructure news

1. Ask whether the story is a product upgrade or a full-stack architecture; NVIDIA says DSX is codesigned as one system.
2. Look for cooling water assumptions, because NVIDIA DSX at 2-gigawatt scale is described as saving 54 million liters of water a day.
3. Check whether the power results are inside a fixed budget: DSX MaxLPS can enable same output with up to 30% less power.
4. See if grid coordination appears, as DSX Flex adapts workload power consumption to utility needs.
5. Note which partners are already included in the materials, from CoreWeave, Firmus, Google Cloud, IREN, Lambda, Nebius, Nscale and Sharon AI to the broader NVIDIA DSX OS list.

Common misconceptions

• Misconception: liquid cooling always consumes water like a cooling tower. The DSX Reference Design discussion contrasts liquid-cooled computing with dry-cooler-based heat rejection that avoids evaporation.
• Misconception: power efficiency means the data center becomes smaller. The materials describe more GPU capacity or throughput within existing power budgets, not necessarily a smaller facility.
• Misconception: responding to the grid means stopping workloads. The U.K. National Grid trial description says critical workloads continued during reductions.

Action advice for tech readers and teams

1. Read NVIDIA DSX announcements as a resource-configuration story: water, electricity, simulation, and management software are being designed together.
2. For project evaluation, look for DSX Sim digital twins before construction, because simulation-ready assets can help builders evaluate power and cooling decisions earlier.
3. For operations conversations, ask whether DSX MaxLPS or DSX Flex is being explored, since they affect power allocation and grid responsiveness.
4. For infrastructure buying, check NVIDIA DSX Ready equipment categories if relevant, including CoolChip, LiquidStack GigaModular, and BESS qualification examples.

For everyday readers, the useful conclusion is that AI factory news now has several scoreboards: resource use, electrical capacity, grid cooperation, and lifecycle software. NVIDIA DSX ties those scoreboards to Europe’s ecosystem partners, and the listed results suggest why the platform is presented as a way to make AI capacity more workable within real constraints.