Vertiv and NVIDIA Debut AI Factory Approach as Polar Deploys in Norway

Vertiv Holdings Co and NVIDIA on October 21, 2026 introduced an AI factory approach at NVIDIA GTC Berlin that combines rapid configuration and virtual validation of pre-engineered reference architectures with factory-built deployment, and Polar Data Centers is applying the factory-built deployment model to the first 41 MW phase of its multiphase AI data center at Herøya, Norway. Under defined assumptions, the approach is designed to generate up to 30% more cumulative token output within the first 12 months with the same installed capacity, Vertiv said in its announcement.

For AI factory operators, securing compute is only part of the challenge, the company said. Traditional infrastructure development can require power, cooling and facility systems to be engineered and integrated through separate, sequential processes, creating complexity and potential delays. The Vertiv and NVIDIA approach starts instead with pre-engineered AI factory reference architectures that can be rapidly configured to customer and site requirements, virtually validated before deployment, and then translated into factory-built infrastructure.

Configure and Validate, Then Deploy

The configuration stage combines the NVIDIA Vera Rubin DSX AI Factory reference design and the NVIDIA Omniverse DSX Blueprint with Vertiv SimReady power and cooling assets, integrated with Dassault Systèmes’ 3DEXPERIENCE agentic platform. Rather than developing an integrated infrastructure design from scratch, teams configure and virtually validate the reference designs against project requirements, interfaces and constraints. In Vertiv’s modeled scenario, this reduces the time from requirements to a configured and validated AI factory solution from approximately 30 days to three days, which the company describes as a 10x improvement. A footnote to the announcement states the figure is modeled, measuring the time to an initial integrated design configured and validated in the digital twin against a 30-day conventional design process, and that results vary by project.

Vertiv is also contributing SimReady OpenUSD assets of its power and cooling equipment to the SimReady Central Library on Hugging Face. According to the announcement, the library helps engineering teams discover and reuse validated equipment assets in AI factory digital twins, reducing asset preparation and helping teams evaluate infrastructure options during design.

Once a configuration is validated, Vertiv’s prefabricated solutions shift more assembly, integration and testing into the factory before equipment reaches the site. These solutions include Vertiv OneCore, a hybrid-built data center solution that integrates critical power, thermal management and supporting infrastructure, and Vertiv SmartRun, a prefabricated overhead infrastructure system for high-density power distribution, liquid cooling, networking and containment. In the modeled scenario, the factory-built approach reduces the infrastructure deployment period from approximately 90 days to 45 days, a reduction of up to 50%. The footnote defines this as on-site power and cooling deployment with Vertiv factory-built infrastructure against a 90-day conventional on-site baseline, again with results varying by project.

Combined with the faster configuration and validation stage, the modeled approach reduces these stages from approximately 120 days to 48 days, the company said, which can make productive AI capacity available about 10 weeks sooner. The 30% token-output figure carries its own footnote: the first 12 months are measured from design kickoff as 317 versus 245 productive days at equal installed capacity, based on 3 + 45 versus 30 + 90 days, and actual results will vary.

Polar Data Centers at Herøya

At its Herøya campus in Norway, Polar is using Vertiv’s factory-built deployment model for the first 41 MW phase of its multiphase AI data center, which is scalable to 100 MW. Vertiv is supplying prefabricated power and thermal modules, Vertiv SmartRun overhead infrastructure and Vertiv SmartMod, supported by Vertiv Unify software. The announcement states the deployment model is designed to reduce on-site complexity and help make high-density, liquid-cooled capacity available to customers sooner.

“An AI factory earns its return in tokens, and every day spent configuring infrastructure or integrating it on site is output that never comes back,” said Giordano Albertazzi, CEO of Vertiv. He said success depends on how quickly operators can move from requirements to productive AI capacity, and that Polar is demonstrating how the approach can be applied at scale in a real-world AI factory deployment.

“Our customers measure us on how quickly capacity is ready for them,” said Andy Hayes, CEO of Polar Data Centers. He said Vertiv’s factory-built infrastructure helps Polar manage the complexity of deploying power, cooling and white-space systems for high-density computing at Herøya, giving the operator a more disciplined, repeatable deployment model while preserving the flexibility to meet customer requirements.

Vladimir Troy, vice president of AI infrastructure at NVIDIA (NVDA ), said AI factories must be designed as one system and that the Vera Rubin DSX AI Factory reference design helps teams align compute, power and cooling requirements from the outset. He said the Omniverse DSX Blueprint allows companies such as Vertiv and Dassault Systèmes to simulate these systems together and carry an integrated design into physical deployment. Stéphane Sireau, vice president for the high-tech industry at Dassault Systèmes, said the companies are engineering the factory as one system, with compute, networking, power and cooling co-designed and co-simulated in a virtual twin, so integration issues are resolved before construction rather than on site.

Elements of the approach were announced earlier in 2026. NVIDIA introduced the Vera Rubin DSX AI Factory reference design and general availability of the Omniverse DSX Blueprint on March 16, 2026, naming Vertiv among the industry companies contributing platforms and SimReady assets to the architecture. That release said Vertiv was using the blueprint to build Vertiv OneCore Rubin DSX, a prefabricated, converged data center infrastructure solution designed to accelerate AI factory deployment and AI output per watt. The reference design spans compute, NVIDIA Spectrum-X Ethernet networking and storage, with documentation covering power, cooling and control systems, and its software libraries include DSX Max-Q, DSX Flex, DSX Exchange and DSX Sim.

On June 1, 2026, Vertiv announced progress on a production-grade digital twin capability for Vertiv SmartRun integrated in the NVIDIA Omniverse DSX Blueprint, describing it as the first phase of a multi-phase AI factory digital twin roadmap. The capability was created using Dassault Systèmes’ model-based systems engineering on the 3DEXPERIENCE platform, and Vertiv said it would demonstrate it at Computex Taipei 2026.

Vertiv and NVIDIA plan to demonstrate elements of the approach at NVIDIA GTC Berlin, including how facility requirements inform infrastructure configuration, how teams can virtually validate power and cooling requirements, and how Vertiv applies factory-built infrastructure to accelerate physical deployment. The demonstration will feature technology from Dassault Systèmes, and Polar Data Centers will show how the deployment approach is being applied in a real-world AI factory. The demonstrations are at Vertiv booth #3055, and Scott Armul, Vertiv’s chief product and technology officer, is scheduled to present Building the AI Factory: Converting Megawatts into Intelligence during a sponsored expo-theater session on October 22, 2026, from 4:00 to 4:15 p.m. CEST.