NVIDIA Launches DSX Ready to Qualify Power and Cooling Products for AI Factories
NVIDIA has launched NVIDIA DSX Ready, a new qualification program designed to help AI factory builders identify power and cooling products that meet applicable requirements in NVIDIA’s DSX AI factory reference designs.
The program, announced September 21, 2026, initially covers two infrastructure categories: battery energy storage systems (BESS) and cooling distribution units (CDUs). NVIDIA says the program is intended to give builders greater confidence when evaluating infrastructure products and help reduce integration risk as AI facilities become increasingly complex.
The move comes as AI infrastructure expands toward increasingly large deployments, putting greater pressure on electricity, cooling, water availability, sites and grid connections.
What Is NVIDIA DSX Ready?
NVIDIA DSX Ready connects NVIDIA’s AI factory reference designs with specific partner products that have gone through applicable qualification requirements.
The broader NVIDIA DSX AI factory platform is designed to bring compute, networking, power, cooling, facilities and software together as a single system rather than treating each infrastructure component independently.
That approach is increasingly important for large AI facilities because improving one part of a system can create a constraint somewhere else. A power system, for example, has to work within the requirements of the computing and facility architecture, while cooling infrastructure must be evaluated as part of the overall thermal design.
NVIDIA says DSX Ready gives builders a qualification they can look for when evaluating products, while giving suppliers a defined process for demonstrating that their offerings meet applicable requirements.
Six Products Categories Enter the Program at Launch
NVIDIA’s initial DSX Ready program includes qualified products from several major infrastructure companies.
For battery energy storage systems, the qualified providers announced at launch are:
- Hitachi Energy
- LG Energy Solution
- Tesla
For cooling distribution units, the qualified providers are:
- LG Electronics
- LiquidStack
- Vertiv
NVIDIA said additional infrastructure and software categories will be introduced over time.
Why Battery Storage Matters for AI Factories
Battery energy storage systems can play an important role in large AI facilities because they can interact with the electrical grid and help manage changes in power demand.
NVIDIA’s DSX documentation describes its BESS qualification process as covering systems intended for applications including AI load buffering, demand response and low-voltage ride-through in grid-connected and islanded operating modes.
The qualification process includes technical testing and submission of supporting data to NVIDIA for review.
However, NVIDIA makes an important distinction: passing the BESS qualification does not mean a system has automatically been proven stable for every individual site.
The qualification has a defined technical boundary. Site-specific equipment and engineering considerations still have to be evaluated by the builder.
Cooling Becomes a Critical AI Infrastructure Issue
Cooling is another major consideration as AI computing systems become more power-dense.
The DSX Ready launch includes cooling distribution units, or CDUs, that connect liquid-cooling infrastructure with the broader facility system.
NVIDIA says CDU qualification uses its self-qualification process to determine whether a specific offering meets applicable functional requirements.
That does not eliminate the need for facility-level engineering. A qualified CDU still has to be evaluated against the requirements of the particular AI factory, including its configuration and operating conditions.
NVIDIA’s broader DSX ecosystem already includes simulation-ready infrastructure assets. Its Omniverse DSX documentation, for example, includes Vertiv’s XDU2300 CDU as a SimReady asset with information covering cooling performance, pump characteristics, connections and electrical ratings.
NVIDIA Wants AI Factory Builders to Think Beyond GPUs
The DSX Ready announcement reflects a broader shift in how NVIDIA is approaching AI infrastructure.
As AI workloads expand, building an AI factory is no longer simply a matter of installing more GPUs. Electricity, thermal management, networking, facility design and grid interaction can all affect how much computing capacity can actually be deployed.
NVIDIA has previously described DSX as a platform for designing and operating AI factories across compute, power, cooling, facilities and software. The company has also been working with infrastructure partners on digital twins and validated designs intended to simulate and refine AI factory systems before physical construction.
The DSX Ready program adds another layer by connecting those reference designs with specific products that have gone through NVIDIA’s qualification process.
What Happens Next?
For AI factory builders, the immediate focus is on BESS and CDU products covered by the initial program.
For suppliers, NVIDIA has established qualification paths that require testing or self-qualification depending on the product category. BESS providers run specified tests and submit evidence for NVIDIA review, while CDU providers use the applicable self-qualification process.
NVIDIA says more categories will be added over time.
As AI data centers grow larger and power and cooling constraints become increasingly important, programs such as DSX Ready could give infrastructure buyers a more standardized way to evaluate whether individual components fit within NVIDIA’s broader AI factory architecture.
For now, the program starts with the two infrastructure areas that directly address some of the most important physical requirements of large-scale AI deployment: power storage and liquid cooling.
