Vertiv has announced progress on a production-grade digital twin capability for Vertiv SmartRun, now integrated in the NVIDIA Omniverse DSX Blueprint. The goal is to make AI factory infrastructure planning more configurable, repeatable, and simulation-ready, as customers push for faster transitions from compute design to build-out physical systems.
From document handoffs to model-based co-design
As AI deployments scale toward higher densities and larger capacities, Vertiv argues the industry’s traditional, document-driven approach—where power, cooling, controls, and deployment teams pass information through siloed workflows—struggles to keep pace.
SmartRun’s digital twin shifts the process toward model-based design, enabling infrastructure to be designed, simulated, and validated as a single system before construction begins.
In practical terms, the digital twin captures system configurations and dependencies within a virtual environment, aiming to reduce late-stage design changes and integration risk, improve engineering confidence via simulation, and accelerate the path from planning to operational readiness.
Interdependent infrastructure planning
Scott Armul, chief product and technology officer at Vertiv, said: “AI infrastructure can no longer be planned one compute generation at a time.”
He added that customers need power, cooling, controls, and deployment workflows to be designed as an interdependent system, with Vertiv’s SmartRun digital twin encoding Vertiv expertise into configurable, simulation-ready building blocks.
Ecosystem momentum with DSX and SimReady assets
Vertiv’s update also links into NVIDIA’s DSX ecosystem approach, which uses OpenUSD workflows and SimReady assets to bring power, thermal, and operational simulations earlier into the design cycle.
Vladimir Troy, vice president of AI Infrastructure at NVIDIA, said the DSX Blueprint helps the ecosystem build and optimise gigawatt-scale AI factory digital twins, and that bringing Vertiv SmartRun into the workflow can help customers evaluate infrastructure choices earlier.
