Nvidia has launched a new certification program aimed at standardizing the power and cooling infrastructure that surrounds its AI computing hardware, naming Tesla and Vertiv among the first companies whose products have been qualified under the new framework.
The chipmaker introduced the program, called DSX Ready, in a company blog post on September 21. The initiative is designed to help builders of AI data centers, often referred to within the industry as “AI factories,” identify power and cooling products that meet the requirements laid out in Nvidia’s own DSX reference designs for these facilities. According to Nvidia, the program gives infrastructure partners “a defined way to demonstrate that a specific offering meets the requirements for its category,” while giving data center developers “a clear qualification to look for” when selecting suppliers.
DSX Ready launches with two initial product categories: battery energy storage systems, commonly referred to as BESS, and cooling distribution units, known as CDUs. In the battery storage category, Nvidia has qualified products from Hitachi Energy, LG Energy Solution and Tesla, the latter through its grid-scale Megapack battery system. In the cooling category, the company has qualified products from LG Electronics, LiquidStack and Vertiv, with Vertiv’s qualification specifically covering its 2.3-megawatt CoolChip coolant distribution unit.
The move extends Nvidia’s influence well beyond the graphics processing units at the core of its business and into the surrounding electrical and thermal management systems that determine whether a large-scale AI data center can actually operate reliably. A single rack of Nvidia’s most advanced GPUs can now draw more than 100 kilowatts of power, and modern AI computing clusters routinely stack hundreds or thousands of such racks together, with entire facilities scaling to hundreds of megawatts of total power demand, a scale that has made power stability, backup capacity, liquid cooling and system controls central engineering concerns rather than secondary considerations.
For battery storage suppliers seeking qualification, Nvidia requires partners to conduct a defined set of tests and submit supporting data for the company’s review, evaluating products against Nvidia’s reference design specifications rather than approving an entire data center site outright. According to Nvidia’s DSX documentation, the battery qualification process examines grid-adaptive power conversion performance, with testing covering dynamic power response, current limiting, low-voltage ride-through capability and islanded operation, meaning a battery system’s ability to continue supplying power independently if disconnected from the broader electrical grid. Cooling distribution unit providers, by contrast, use a self-qualification testing suite to demonstrate their products meet Nvidia’s requirements for that category.
Nvidia has been explicit that DSX Ready qualification does not replace the site-specific engineering and validation work still required for any individual data center project, and that qualification alone does not guarantee a supplier will win a specific deployment contract or site approval. The company has said additional infrastructure and software categories are expected to be added to the program over time, building on the initial focus on battery storage and cooling.
Vertiv’s inclusion builds on what was already a substantial existing relationship with Nvidia. The company has been collaborating with Nvidia on prefabricated data center systems and digital-twin modeling for AI factory designs, meaning the new qualification formalizes a partnership that had already been producing real products ahead of Tuesday’s announcement.
For Tesla, the qualification places its Megapack battery product directly within the broader AI infrastructure conversation, opening what could become an additional demand pool for Tesla’s energy storage business beyond its traditional focus on utility-scale and renewable energy projects. Battery systems like Megapack can help smooth out power fluctuations, provide backup ride-through capability during outages, and support data center sites that are still waiting on broader electrical grid upgrades to reach full capacity, all functions increasingly relevant as AI data centers strain local power infrastructure in many regions.
The DSX Ready program builds on Nvidia’s broader DSX platform, which the company first introduced earlier this year at its GTC conference in Taipei as a unified framework for designing, deploying and managing AI factories as integrated systems, spanning compute, networking, power, cooling, facilities and software together rather than treating each component as a separate, disconnected concern. Analysts covering the announcement have noted that the significance of DSX Ready lies less in Nvidia directly controlling the physical power infrastructure of a data center and more in the company effectively defining what a properly deployable, Nvidia-compatible AI factory should look like, a dynamic that could steer developers toward Nvidia’s qualified supplier list when specifying equipment for new projects.
Regulatory filings reviewed by financial analysts, covering positions held before the DSX announcement, showed Nvidia held by 285 hedge funds as of the second quarter of 2026, up from 275 funds in the first quarter, with Fisher Asset Management increasing its Nvidia stake by roughly 3% during that period. The same filings showed Tesla’s hedge fund ownership had declined slightly to 116 holders from 123, even as investment firm BAMCO held more than 12.5 million Tesla shares after increasing its position by 5%. Vertiv’s hedge fund ownership, meanwhile, had climbed to roughly 112 holders from 96 in the prior quarter, with AQR Capital Management holding nearly 2.7 million shares after modestly increasing its own stake. Those filings all predate the DSX Ready announcement and reflect investor positioning heading into, rather than in direct response to, the new certification program.
With DSX Ready now formally launched and additional product categories expected to follow, the program is likely to become an increasingly significant factor in how AI data center developers select power and cooling suppliers going forward, giving Nvidia a growing role in shaping the broader physical infrastructure ecosystem surrounding its computing hardware, even as the company has been careful to stress that site-level engineering and independent validation remain necessary steps beyond simple product qualification.
You must be logged in to post a comment Login