Vertiv and NVIDIA have introduced an AI factory approach designed to accelerate infrastructure deployment and deliver up to 30% more cumulative AI token output within the first 12 months, using the same installed capacity under defined assumptions.

Announced at NVIDIA GTC Berlin on October 22, 2026, the approach combines pre-engineered reference designs, virtual validation and factory-built infrastructure. It aims to help operators move from project requirements to productive AI capacity faster.

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. The facility is designed to scale to 100 MW.

Traditional AI infrastructure projects often require power, cooling and facility systems to be engineered and integrated through separate processes. This can increase complexity and delay deployment. The Vertiv and NVIDIA approach instead uses pre-engineered reference architectures that can be configured for specific customer and site requirements, virtually validated, and then translated into factory-built infrastructure.

The configuration and validation process combines the NVIDIA Vera Rubin DSX AI Factory reference design and NVIDIA Omniverse DSX Blueprint with Vertiv SimReady power and cooling assets. These are integrated with Dassault Systèmes’ 3DEXPERIENCE agentic platform.

According to the companies, the approach can reduce the time required to configure and validate an AI factory solution from approximately 30 days to three days, representing a tenfold improvement in the modeled scenario.

Vertiv is also contributing SimReady OpenUSD assets of its power and cooling equipment to the SimReady Central Library on Hugging Face. The library allows engineering teams to discover and reuse equipment assets in AI factory digital twins. This can reduce asset preparation requirements and help teams evaluate infrastructure options during the design process.

The deployment phase uses Vertiv prefabricated solutions to move more assembly, integration and testing into the factory before equipment reaches the site. These include Vertiv OneCore, which 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, this factory-built approach reduces the infrastructure deployment period from approximately 90 days to 45 days, a reduction of up to 50%.

Combined, the faster configuration, validation and deployment processes reduce the modeled timeline for these stages from approximately 120 days to 48 days. The companies estimate that this could make productive AI capacity available about ten weeks earlier, enabling up to 30% more cumulative token output during the first 12 months after project kickoff, with the same installed capacity.

At its Herøya campus in Norway, Polar Data Centers is using Vertiv’s factory-built deployment model for the first 41 MW phase of its AI data center. Vertiv is supplying prefabricated power and thermal modules, SmartRun overhead infrastructure and SmartMod, supported by Vertiv Unify software.

Giordano Albertazzi, CEO of Vertiv, said AI factory returns depend on token output and the time required to bring infrastructure into operation. He explained that combining pre-engineered reference designs with factory-built deployment can help reduce complexity and bring AI capacity online sooner.

Vladimir Troy, Vice President of AI Infrastructure at NVIDIA, said the NVIDIA Vera Rubin DSX AI Factory reference design helps teams align computing, power and cooling requirements from the outset. He added that the NVIDIA Omniverse DSX Blueprint allows companies such as Vertiv and Dassault Systèmes to simulate systems together and carry integrated designs into physical deployment.

Stéphane Sireau, Vice President, High-Tech Industry, at Dassault Systèmes, said the approach involves co-designing and co-simulating computing, networking, power and cooling within a virtual twin. He noted that resolving integration issues before construction can help bring productive capacity online sooner.

Andy Hayes, CEO of Polar Data Centers, said the factory-built infrastructure model helps manage the complexity of deploying power, cooling and white-space systems for high-density computing at Herøya. He added that the approach supports a more repeatable deployment process while retaining flexibility to meet customer requirements.

Vertiv and NVIDIA plan to demonstrate the approach at NVIDIA GTC Berlin. The demonstration will show how facility requirements inform infrastructure configuration, how teams virtually validate power and cooling requirements, and how factory-built infrastructure can accelerate physical deployment. It will feature technology from Dassault Systèmes, while Polar Data Centers will demonstrate how the model is being applied at its Norwegian site.

Visitors can find the demonstration at the Vertiv booth, number 3055. Scott Armul, Chief Product and Technology Officer at Vertiv, is also 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.

For more information about Vertiv’s power and thermal management products, infrastructure solutions, IT systems and services for critical digital applications, visit Vertiv’s official website. The companies’ AI factory approach focuses on shortening design and deployment timelines to bring productive AI computing capacity online sooner.

Related post

View all