India, November 02, 2026: Vertiv Holdings Co. and NVIDIA have introduced a new AI factory approach that aims to accelerate deployment and increase productivity, with the potential to generate up to 30% more cumulative token output in the first 12 months using the same installed capacity. The announcement was made at NVIDIA GTC Berlin, where the companies detailed how their method combines rapid configuration and virtual validation of pre-engineered reference architectures with factory-built deployment.
The approach is designed to move AI factories from project kickoff to productive capacity faster by leveraging pre-engineered reference designs and shifting more assembly and integration off-site. Polar Data Centers is implementing this model for the initial 41 MW phase of its multiphase AI data center at Herøya, Norway, which is scalable to 100 MW.
Faster configuration and validation
Teams can use pre-engineered AI factory reference designs, such as the NVIDIA Vera Rubin DSX AI Factory reference design and NVIDIA Omniverse DSX Blueprint, along with Vertiv SimReady power and cooling assets, all integrated with Dassault Systèmes’ 3DEXPERIENCE agentic platform. This process allows for rapid configuration and virtual validation of compute, power, cooling, and facility requirements in a shared virtual environment.
According to the companies, this method can reduce the time from requirements to a configured and validated AI factory solution from about 30 days to three days, representing a tenfold improvement. Vertiv is also contributing SimReady OpenUSD assets of its power and cooling equipment to the SimReady Central Library on Hugging Face, enabling engineering teams to discover and reuse validated equipment assets in AI factory digital twins.
Accelerated deployment with factory-built infrastructure
Once configurations are validated, Vertiv’s prefabricated solutions, including Vertiv OneCore and Vertiv SmartRun, allow for more assembly, integration, and testing to occur in the factory. This reduces the infrastructure deployment period from roughly 90 days to 45 days, a 50% reduction, according to modeled scenarios. Combined with faster configuration and validation, the total time from requirements to productive AI capacity can be reduced from approximately 120 days to 48 days, making capacity available about 10 weeks sooner.
At Polar Data Centers’ Herøya site, Vertiv is supplying prefabricated power and thermal modules, Vertiv SmartRun overhead infrastructure, Vertiv SmartMod, and Vertiv Unify software to support the deployment.
Industry perspectives and demonstration plans
Giordano Albertazzi, CEO of Vertiv, said, “Success depends on how quickly operators can move from requirements to productive AI capacity. By starting with pre-engineered reference designs that can be rapidly configured and validated, then extending that approach through factory-built deployment, we help customers reduce complexity, deploy infrastructure faster and bring AI capacity online sooner. Polar Data Centers is demonstrating how that approach can be applied at scale in a real-world AI factory deployment.”
Vladimir Troy, vice president of AI infrastructure at NVIDIA, noted that the NVIDIA Vera Rubin DSX AI Factory reference design helps teams align compute, power, and cooling requirements from the outset, and that simulation with NVIDIA Omniverse DSX Blueprint enables integrated design and earlier conflict resolution. Stéphane Sireau, Vice President, High-Tech Industry at Dassault Systèmes, added that co-design and co-simulation in a virtual twin allows integration issues to be resolved before construction. Andy Hayes, CEO of Polar Data Centers, said Vertiv’s factory-built infrastructure provides a disciplined, repeatable deployment model while maintaining flexibility for customer requirements.
Demonstration at NVIDIA GTC Berlin
Vertiv and NVIDIA plan to showcase elements of the approach at NVIDIA GTC Berlin, including facility requirement-driven infrastructure configuration, virtual validation of power and cooling, and accelerated physical deployment using Vertiv’s factory-built infrastructure. The demonstration will also feature technology from Dassault Systèmes, and Polar Data Centers will present its real-world application of the deployment model.
Vertiv Holdings Co. is headquartered in Westerville, Ohio, USA, and operates in more than 130 countries. Polar Data Centers develops high-performance data centers for AI and high-density computing workloads, while Dassault Systèmes provides the 3DEXPERIENCE platform for virtual twin simulations.
Disclaimer: This article is based on a press release provided by the concerned organisation and has been edited for style and clarity with the assistance of AI tools. BusinessFortnight does not verify, endorse, or take responsibility for the claims, figures, or statements made in the original release.
