Data Center
Vertiv has announced its collaboration with NVIDIA to advance next-generation physical infrastructure for AI factories, contributing to the NVIDIA Vera Rubin DSX reference design and the Omniverse DSX Blueprint.
The partnership focuses on addressing the growing complexity of AI infrastructure as data centers evolve into high-density “AI factories” with significantly higher power and cooling demands. Vertiv is providing simulation-ready digital assets—referred to as DSX SimReady components—along with validated interfaces and modular infrastructure building blocks aimed at accelerating deployment timelines and reducing integration risks.
As AI workloads scale, operators are increasingly challenged to bring capacity online faster while ensuring system reliability and efficiency. Vertiv’s approach centers on what it describes as “converged physical infrastructure,” integrating power, cooling, controls, and services into a unified system optimized across the full energy and thermal lifecycle.
This model is built on five core elements: standardized building blocks, defined system interfaces, coordinated orchestration, digital continuity, and lifecycle support. Together, these components are designed to simplify infrastructure design, improve coordination across systems, and enhance operational predictability from planning through deployment.
At the heart of the solution is a modular architecture based on 12.5-megawatt infrastructure blocks under Vertiv’s OneCore platform. These blocks can be scaled and configured to support deployments ranging from smaller AI clusters to hyperscale, gigawatt-level facilities, enabling more consistent and repeatable buildouts.
Scott Armul said the shift toward AI factories is fundamentally changing how digital infrastructure is designed and deployed, requiring a move away from fragmented systems toward integrated, simulation-driven models that can be validated before physical implementation.
The collaboration also leverages real-time simulation and system-level modeling, allowing enterprises to test infrastructure performance digitally before construction begins. This is expected to help reduce deployment complexity, shorten time to operational readiness, and optimize performance across power distribution, cooling systems, and chip-level thermal management.
Vladimir Troy emphasized that as AI infrastructure scales in both power and density, enterprises need tightly integrated systems that combine physical infrastructure with digital twin simulation to minimize risk and accelerate deployment.
The joint initiative has resulted in a new design framework, Vertiv OneCore Rubin DSX, which integrates power, cooling, and control systems into a unified, deployment-ready model. Vertiv plans to continue evolving the platform to support future generations of AI compute infrastructure.
The announcement underscores a broader industry shift toward standardized, modular, and simulation-led infrastructure approaches as companies race to build and scale AI capacity more efficiently in an increasingly competitive landscape.
The partnership focuses on addressing the growing complexity of AI infrastructure as data centers evolve into high-density “AI factories” with significantly higher power and cooling demands. Vertiv is providing simulation-ready digital assets—referred to as DSX SimReady components—along with validated interfaces and modular infrastructure building blocks aimed at accelerating deployment timelines and reducing integration risks.
As AI workloads scale, operators are increasingly challenged to bring capacity online faster while ensuring system reliability and efficiency. Vertiv’s approach centers on what it describes as “converged physical infrastructure,” integrating power, cooling, controls, and services into a unified system optimized across the full energy and thermal lifecycle.
This model is built on five core elements: standardized building blocks, defined system interfaces, coordinated orchestration, digital continuity, and lifecycle support. Together, these components are designed to simplify infrastructure design, improve coordination across systems, and enhance operational predictability from planning through deployment.
At the heart of the solution is a modular architecture based on 12.5-megawatt infrastructure blocks under Vertiv’s OneCore platform. These blocks can be scaled and configured to support deployments ranging from smaller AI clusters to hyperscale, gigawatt-level facilities, enabling more consistent and repeatable buildouts.
Scott Armul said the shift toward AI factories is fundamentally changing how digital infrastructure is designed and deployed, requiring a move away from fragmented systems toward integrated, simulation-driven models that can be validated before physical implementation.
The collaboration also leverages real-time simulation and system-level modeling, allowing enterprises to test infrastructure performance digitally before construction begins. This is expected to help reduce deployment complexity, shorten time to operational readiness, and optimize performance across power distribution, cooling systems, and chip-level thermal management.
Vladimir Troy emphasized that as AI infrastructure scales in both power and density, enterprises need tightly integrated systems that combine physical infrastructure with digital twin simulation to minimize risk and accelerate deployment.
The joint initiative has resulted in a new design framework, Vertiv OneCore Rubin DSX, which integrates power, cooling, and control systems into a unified, deployment-ready model. Vertiv plans to continue evolving the platform to support future generations of AI compute infrastructure.
The announcement underscores a broader industry shift toward standardized, modular, and simulation-led infrastructure approaches as companies race to build and scale AI capacity more efficiently in an increasingly competitive landscape.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
