Nvidia says six companies are building on its DSX AI factory platform

Nvidia said that six companies are building on its DSX AI Factory Platform to plan, simulate and run the power grid and the data center as one system. The post names a $300 billion equipment backlog and more than 200 gigawatts of U.S. interconnection queue as the problem, and matters to readers because power access now gates AI capacity.

• Nvidia says GE Vernova, Hitachi Energy, Schneider Electric, Siemens Energy, Phaidra and Emerald AI are building on DSX.
• Phaidra cites about 10% more compute per power envelope; Emerald AI models up to 100 gigawatts.
• Watch for named utility or data center customers and test baselines, none of which the post supplies.

• $300B equipment backlog Nvidia cites; no prior-year comparison given
• 200 GW projects waiting in U.S. interconnection queues, per Nvidia
• ~10% more compute in same power envelope, Phaidra’s claim
• 100 GW modeled unlock on existing U.S. grid, half the 200 GW queue

Nvidia’s blog post is the release describing partner adoption of DSX, which it pitches as a way to plan, simulate and run the grid and the AI factory “from the substation to the rack.” It is a corporate blog, not a filing, and it carries no contract values, customer names or deployment dates.

The partners span equipment makers, engineering software and optimization startups.

What each partner is doing, as Nvidia describes it

Nvidia lists four DSX components. DSX Sim and SimReady assets let engineers simulate the grid, substation and AI factory together, validating generators, switchgear, power systems and cooling “before breaking ground.” DSX Flex syncs AI factory demand with grid conditions and onsite generation, and DSX MaxLPS and DSX Exchange aim to extract more compute per watt and share signals across compute, networking, energy, cooling and operations.

Hitachi Energy’s battery energy storage system has qualified as NVIDIA DSX Ready, which Nvidia says helps customers identify power infrastructure that meets applicable DSX reference architecture design requirements. GE Vernova is “deploying” grid-to-facility and rack power expertise to complement DSX systems. Schneider Electric, with AVEVA and ETAP, is advancing DSX with validated Vera Rubin power and cooling reference architectures and an Omniverse-powered AI factory digital twin.

Siemens Energy is developing a reference design for grid-aware power infrastructure aligned with the DSX Vera Rubin AI factory design. It is also only “exploring” Nvidia’s RAPIDS, Metropolis and Isaac Sim for its Noedra digital twin platform. Phaidra integrated DSX Flex, DSX Max-Q and DSX Exchange into its self-learning agents, and Emerald AI integrated its Conductor platform into DSX Flex to help utilities tap underused grid capacity.

Power is now the biggest bottleneck in infrastructure.

Nvidia blog post, October 22, 2026

What changes for developers and utilities

The practical offer is to move power validation earlier. Simulating generators, switchgear and cooling before construction targets the siloed design process Nvidia names as a root cause of delay and stranded capacity. For developers, vendor equipment tagged DSX Ready could shorten procurement checks against a single reference architecture.

The Flex pitch is aimed at the queue. Nvidia describes it as “turning AI factories from static loads into flexible grid assets,” and Emerald AI’s modeled 100 gigawatts equals half of the 200 gigawatts waiting in U.S. queues. The post does not say the two figures are linked, and a model of existing grid headroom is not a count of connected projects.

There is also a standards effect. Hitachi’s qualification and Siemens’ alignment are measured against Nvidia’s design requirements, so grid equipment makers are conforming to a compute vendor’s specification. The post does not say who audits DSX Ready or what the criteria are.

What the post does not say

It names no utility, grid operator or data center owner that has deployed DSX Flex or agreed to curtail load. Phaidra’s “about 10%” gain has no stated baseline, site or duration, and the 100 gigawatt figure is explicitly modeled. GE Vernova’s role is described as complementing DSX, with no product or commercial terms. Siemens Energy’s Noedra work is exploratory.

The post also gives no pricing, no licensing terms for DSX components, and no dates for the reference designs. It offers no comparison for the $300 billion backlog or the 200 gigawatts, so readers cannot tell whether either is growing.

The BlockWest read. The binding test is not whether simulation shortens design cycles but whether utilities will treat AI load as curtailable when they allocate interconnection. Until a utility accepts flexibility in a connection agreement, DSX Flex is a promise from the demand side. If one does, Nvidia’s reference architecture becomes a template for grid terms, and equipment vendors will be building to it.

Nvidia positions the news around GTC Berlin, where Hitachi Energy, Schneider Electric and Siemens Energy have booths. The open question for Nvidia and Emerald AI is which utility will sign a flexible-load arrangement built on DSX Flex, and for Phaidra, what baseline sits behind the 10%.

BlockWest is a news publication. Nothing here is investment advice. Read our disclaimer and editorial policy.