Nvidia has unveiled DSX, a comprehensive platform integrating hardware, software, and simulation capabilities to streamline AI factory deployment, optimise energy use, and promote industrial AI scalability.
Nvidia has moved to formalise the way AI infrastructure is planned and run with the launch of DSX, a full-stack platform it says is intended to give builders a common framework for AI factories from the first design sketches through to day-to-day operations.
Announced at GTC Taipei, DSX brings together open-source software libraries, APIs, reference designs, Nvidia’s accelerated computing platforms and partner technologies in an effort to align chips, systems, software, facilities and operational tools around a single infrastructure model. Nvidia says the goal is to cut token costs, shorten the route to first production and make it easier to turn available power into useful AI output.
Jensen Huang, Nvidia’s founder and chief executive, framed the platform as a broader shift in how AI infrastructure is delivered. “We’re not just shipping chips , we’re giving every infrastructure builder a complete playbook to build AI factories,” he said. “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.”
The company is positioning DSX as more than a reference architecture. It is intended to cover compute, networking, storage, facility planning, power, cooling, controls, simulation and operations, reflecting the scale and complexity now associated with industrial AI build-outs.
Among the newer pieces is DSX MaxLPS, a set of technologies Nvidia says is designed to maximise token performance per megawatt within a fixed power envelope. It combines 45-degree-Celsius liquid cooling with in-rack technologies aimed at improving performance per watt, and Nvidia claims this can allow operators to run up to 40% more GPUs at their most efficient operating point with little effect on workload performance.
DSX OS, meanwhile, is an open-source, modular software layer for AI factory operations. Nvidia says it supports lifecycle management, intelligent scheduling, runtime consistency, health automation, resilience, multi-tenant operations and platform services. Together with DSX Reference Design and DSX Sim, it forms part of a wider push to make AI infrastructure more predictable to deploy and easier to optimise at scale.
DSX Sim is intended to let operators model and validate infrastructure decisions before a facility is built, then continue using the same simulation layer during deployment and operations. Nvidia’s documentation says the approach is built around digital twins for AI factories, with the Omniverse DSX Blueprint providing an end-to-end framework for gigawatt-scale environments.
The platform also reaches beyond the data hall. DSX Flex is designed to connect AI factories to power-grid services, allowing workloads to respond to grid signals such as load shedding, demand response and pricing events. Nvidia says that could help preserve grid reliability and affordability while creating room for further AI growth. DSX Exchange is designed to support secure exchange of compute, network, energy, power and cooling signals between IT systems, operational technology and operational agents.
The industrial implications are clear. As AI infrastructure becomes more power-intensive, the challenge is no longer only how to add more compute, but how to integrate it with constrained electricity supply, cooling systems and facility-level controls. Nvidia’s pitch is that DSX can make those trade-offs measurable and manageable before capital is committed, rather than after a site is already under construction.
The company is also building an ecosystem around the platform. Cloud providers including CoreWeave, Crusoe, Firmus, IREN, Lambda, Nebius, Nscale and Yotta Data Services are deploying DSX Sim, DSX MaxLPS and DSX OS components, according to Nvidia. The company says these roll-outs are meant to reduce risk, improve GPU utilisation and bring capacity online faster.
On the hardware side, Dell Technologies, HPE, Lenovo and Supermicro, alongside ASUS, Foxconn, GIGABYTE, Pegatron, Quanta Cloud Technology, Wistron and Wiwynn, are developing DSX-ready systems and contributing simulation-ready assets. Quanta Cloud Technology and Pegatron are working with Dassault Systèmes on a live digital twin configurator for AI factories, aimed at automating rack-to-facility design with less effort and better quality.
Nvidia is also extending DSX into software ecosystems through partners such as Cadence, PTC and Siemens. Other software firms, including Aible, BeyondAI, Bhashini, DCAI, Mirantis, OpenNebula Systems, Rafay, Red Hat, Sarvam, Simplismart, Spectro Cloud, Supermicro, vCluster and Vultr, are adopting DSX OS components for areas such as lifecycle management, security, multi-tenancy and resilience.
One of the most notable tests of the platform is a commercial, multi-megawatt pilot with Emerald AI and Silicon Valley Power. Nvidia says the project is intended to show how an AI factory can dynamically shift its power demand in response to utility signals without disrupting AI workloads, a model that could prove increasingly relevant as utilities look for ways to manage large, flexible loads.
For industrial decarbonisation professionals, the significance of DSX lies in its attempt to treat AI infrastructure less as a collection of servers and more as an energy system: measurable, simulation-driven, grid-aware and increasingly tied to cooling, operations and power constraints. Nvidia is betting that the next phase of AI expansion will be won not just on silicon, but on how efficiently the whole factory is designed and controlled.
- https://www.techedt.com/nvidia-dsx-gives-infrastructure-builders-a-full-stack-platform-for-ai-factories – Please view link – unable to able to access data
- https://nvidianews.nvidia.com/news/dsx-infrastructure-ai-factory – NVIDIA has introduced the DSX platform, a comprehensive framework for building AI factories. This platform integrates open-source software libraries, APIs, reference designs, and NVIDIA’s accelerated computing platforms, aiming to streamline the design, deployment, and operation of AI infrastructure. By aligning chips, systems, software, facilities, and partner technologies, DSX seeks to reduce token costs and accelerate time to first production. The platform is designed to enhance operational reliability and scalability, enabling operators to convert available power into higher AI output efficiently.
- https://www.nvidia.com/en-us/data-center/products/dsx/ – The NVIDIA DSX platform unifies design, simulation, operations, and ecosystem technologies to help build AI factories optimized for the lowest token cost. It defines how AI factories are designed, built, and optimized across the full stack—from chips and systems to infrastructure software, facilities, and partner technologies. DSX combines open software libraries, workflow guides, and reference designs with NVIDIA compute platforms and co-designed OEM infrastructure to enable a broad ecosystem of software and hardware solutions.
- https://docs.nvidia.com/dsx – NVIDIA DSX Documentation provides comprehensive resources for building and operating AI factories. It includes DSX Sim, a suite of simulation technologies that enable partners to design, validate, and operate AI factories before and after physical deployment. The documentation also covers NVIDIA Omniverse DSX Blueprint for AI Factories, an end-to-end framework to design, simulate, build, and operate gigawatt-scale AI factory digital twins.
- https://www.tomshardware.com/tech-industry/artificial-intelligence/nvidia-announces-reference-design-for-gargantuan-gigawatt-scale-omniverse-dsx-data-centers-single-data-center-requires-a-nuclear-reactors-worth-of-power-generation – At the GTC 2025 summit, Nvidia unveiled the Omniverse DSX Blueprint, a reference design for gigawatt-scale AI data centers dubbed ‘AI factories.’ These facilities are intended for training and running large AI models and require massive energy comparable to a nuclear reactor. The DSX design integrates Nvidia’s Omniverse digital-twin platform with real-world engineering data, allowing partners to plan, simulate, and optimize all components of AI data centers, including power, cooling, compute, and networking.
- https://www.itpro.com/infrastructure/dell-technologies-unveils-massive-expansion-to-dell-ai-factory-with-nvidia-at-gtc-2026 – At Nvidia GTC 2026, Dell Technologies announced a major expansion of its Dell AI Factory with Nvidia, introducing new enterprise-scale AI services and hardware solutions. These include enhanced data analytics and processing capabilities, new server racks with Nvidia’s latest chips, and the Dell Lightning File System—a high-performance storage system targeting heavy AI workloads. Since its 2024 launch, Dell’s AI Factory has delivered 240+ updates and is used by over 4,000 customers, including the Wellcome Sanger Institute.
- https://www.youtube.com/watch?v=cfNPrs-7xwQ – In this video, Jensen Huang, founder and CEO of NVIDIA, explains the DSX platform, an Omniverse digital twin platform for AI factories. He discusses how monthly delays can cost billions, as AI factory revenue equals tokens per watt, and how AI agents dynamically optimize infrastructure when the digital twin becomes the operator. The video highlights the importance of time and efficiency in scaling AI factories and the role of DSX in achieving these goals.
Noah Fact Check Pro
The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.
Freshness check
Score:
8
Notes:
The article was published on June 1, 2026, and reports on NVIDIA’s announcement of the DSX platform at GTC Taipei. A press release from NVIDIA on May 31, 2026, also announced the DSX platform, indicating that the article is based on this press release. The content appears to be original and not recycled from other sources. However, the reliance on a press release may limit the freshness score. The earliest known publication date of substantially similar content is May 31, 2026. No discrepancies in figures, dates, or quotes were found. The article includes updated data and does not recycle older material. Overall, the content is fresh and original.
Quotes check
Score:
7
Notes:
The article includes a direct quote from Jensen Huang, NVIDIA’s CEO: “We’re not just shipping chips — we’re giving every infrastructure builder a complete playbook to build AI factories.” This quote matches the one in NVIDIA’s press release from May 31, 2026. No earlier usage of this exact quote was found, suggesting it is original. However, the reliance on a press release for the quote raises concerns about its independent verification. The wording of the quote is consistent across sources, with no variations noted. Overall, the quote appears to be original but lacks independent verification.
Source reliability
Score:
6
Notes:
The article originates from Tech Edition, a niche publication. While it provides a detailed account of NVIDIA’s DSX platform announcement, the source’s limited reach and potential biases may affect the reliability of the information. The article does not appear to be summarising or aggregating content from other publications. However, the lack of independent verification and the niche nature of the source warrant a cautious approach.
Plausibility check
Score:
8
Notes:
The article’s claims about NVIDIA’s DSX platform align with information from NVIDIA’s official press release and other reputable sources. The details about the platform’s components, such as DSX MaxLPS and DSX OS, are consistent with NVIDIA’s announcements. The language and tone are appropriate for the topic and region. No excessive or off-topic details are present. The report lacks specific factual anchors, such as names, institutions, and dates, which could enhance its credibility. Overall, the claims are plausible but would benefit from more specific details.
Overall assessment
Verdict (FAIL, OPEN, PASS): FAIL
Confidence (LOW, MEDIUM, HIGH): MEDIUM
Summary:
The article provides a detailed account of NVIDIA’s DSX platform announcement, with content that is fresh and original. However, it relies heavily on NVIDIA’s press release for information, and the source’s limited reach and potential biases affect the reliability of the information. The lack of independent verification from other reputable sources raises concerns about the credibility of the content. Therefore, the overall assessment is a FAIL with MEDIUM confidence.
