NVIDIA DSX gives infrastructure builders a full-stack platform for AI factories
NVIDIA DSX brings software, reference designs, simulation, and partner systems into a platform for AI factories.
NVIDIA has announced NVIDIA DSX, a platform designed to give infrastructure builders a common framework for creating AI factories, from design and deployment to operations.
Table Of Content
Announced at NVIDIA GTC Taipei, DSX combines open source, modular software libraries, application programming interfaces, reference designs, NVIDIA accelerated computing platforms, and partner technologies. The platform is built to align chips, systems, software, facilities, and partner technologies around AI factory infrastructure, with the aim of reducing token cost and shortening the time needed to reach first production.
NVIDIA said the platform is designed to improve operational reliability and resiliency at scale while helping operators convert available power into higher AI output. The company is positioning DSX as a full-stack approach to AI infrastructure, covering compute, networking, storage, facility design, power, cooling, controls, simulation, and operations.
“We’re not just shipping chips — we’re giving every infrastructure builder a complete playbook to build AI factories,” said Jensen Huang, founder and CEO of NVIDIA. “With the DSX platform, you can simulate the entire factory before you spend a dollar, validate performance before a single rack is installed and operate with the kind of reliability that production AI demands.”
DSX adds software for power and operations
The latest additions to DSX include DSX MaxLPS and DSX OS, two software components aimed at improving how AI factories are designed and run.
DSX MaxLPS is a suite of technologies designed to maximise token performance per megawatt within a fixed power budget. It combines 45-degrees-Celsius liquid cooling with in-rack technologies that optimise performance per watt. NVIDIA said this allows operators to run up to 40% more GPUs at their most energy-efficient operating point with minimal impact on workload performance.
DSX OS is open source, modular software purpose-built for AI factory operations. It supports lifecycle management, intelligence scheduling, runtime consistency, health automation, resiliency, multi-tenant operations, and platform services.
These additions sit alongside existing DSX platform elements. DSX Reference Design provides generation-specific, validated AI factory architectures across compute, networking, storage, hardware cluster design, and facilities infrastructure. DSX Sim provides a high-fidelity simulation layer for modelling, validating, and optimising infrastructure decisions from planning and design through deployment and operations.
DSX Flex connects AI factories to power-grid services, allowing workloads to adapt to grid signals such as load shedding, demand response, and pricing events. DSX Exchange supports secure integration of compute, network, energy, power, and cooling plant signals between IT, operational technology, and operations agents.
Partners build around DSX-ready systems
NVIDIA is expanding DSX through cloud, system manufacturing, software, and infrastructure partnerships.
Cloud partners CoreWeave, Crusoe, Firmus, IREN, Lambda, Nebius, Nscale, and Yotta Data Services are deploying DSX Sim, DSX MaxLPS, and DSX OS components. NVIDIA said these deployments are intended to reduce risk, improve GPU utilisation, and bring AI cloud capacity online faster.
Dell Technologies, HPE, Lenovo, and Supermicro, together with ASUS, Foxconn, GIGABYTE, Pegatron, Quanta Cloud Technology, Wistron, and Wiwynn, are building NVIDIA DSX-ready systems. These companies are also contributing simulation-ready assets to support full-stack AI factory deployments at global scale.
Model-based systems engineering is a key part of the DSX ecosystem. Quanta Cloud Technology and Pegatron are working with Dassault Systèmes to create a live AI factory digital twin configurator, designed to automate rack-to-facility design with improved quality and reduced workload. System manufacturer adoption of DSX Sim also expands the NVIDIA Omniverse DSX Blueprint ecosystem, with software partners including Cadence, PTC, and Siemens.
DSX Flex moves into a grid-responsive pilot
DSX Flex is being used in a commercial, multi-megawatt pilot with Emerald AI and Silicon Valley Power. The pilot is intended to demonstrate grid-responsive AI factories that can dynamically adjust power consumption in response to utility signals while protecting AI workload performance.
NVIDIA said the approach is designed to help safeguard grid reliability and affordability for customers while unlocking additional power capacity to support AI growth.
Other ecosystem partners are adopting DSX OS components for lifecycle management, multi-tenancy, security, health automation, resilience, and platform services. They include Aible, BeyondAI, Bhashini, DCAI, Mirantis, OpenNebula Systems, Rafay, Red Hat, Sarvam, Simplismart, Spectro Cloud, Supermicro, vCluster, and Vultr.
