NVIDIA DSX POD Packs an AI Factory Into a Container
By Ganesh Karki
Key facts
- Section
- Tech & AI
- Country
- Germany
- Published
- 18 Nov 2026
NVIDIA DSX POD packs an AI factory into a container: seven Vera Rubin NVL72 racks, 504 GPUs and 252 Vera CPUs, cabled, plumbed and tested in the factory before shipping. Nvidia announced the reference design at GTC Berlin on 21 October 2026, promising to cut deployment at a prepared site from about six months to as little as 14 days, with worldwide deliveries starting in early 2027.
NVIDIA DSX POD at a glance
| Detail | Information |
|---|---|
| Announced | GTC Berlin, 21 October 2026 |
| Per pod | 7 Vera Rubin NVL72 racks, 504 GPUs, 252 Vera CPUs |
| Weight | 28 metric tons |
| Cabling | About 17 kilometres of Spectrum-X cabling |
| Deployment | Six months to as little as 14 days, at a prepared site |
| Shipping | Early 2027, worldwide, via 17 partners |
| Power | 800-volt DC delivery, liquid cooling |
What happened
DSX POD is one delivery form of Nvidia's larger DSX infrastructure platform. Partners build the racks into modular containers, connect the networking, power and liquid cooling, and validate the whole system before it leaves the plant. On arrival the pod connects to the site's power, cooling and network. The 17 named partners include Dell, HPE, Lenovo, Schneider Electric, Vertiv, Foxconn's Ingrasys, QCT, Wistron and Eaton. Nvidia also released three DSX OS infrastructure management components as open-source developer previews.
Frontier Now reports that Vera Rubin is now in full production and that the first worldwide shipments of the modular factories are dated to early 2027. Technology.org notes the pod's closed-loop liquid cooling tolerates coolant up to 45 degrees Celsius, and that Nvidia says the wider DSX platform can cut cooling water use by up to 100 percent and deliver the same AI output with up to 30 percent less power.
Why it matters
AI datacentres are now limited less by chips than by how fast buildings can be fitted out, and moving integration into the factory attacks that bottleneck directly. Nvidia launched the design in Berlin to frame it around Europe's AI build-out, with rack integration in France and system manufacturing in the Czech Republic among more than ten European countries involved. The same logic already reshaped Nvidia's own business: the Nvidia Q3 earnings preview shows datacentre revenue near 102 billion dollars, and demand at that scale needs factories, not just chips. The container also sits inside a tense supply picture, with the tightened United States chip curbs on China redrawing where advanced hardware can be sold.
What is not confirmed
The 14-day figure is Nvidia's own claim and is not independently verified. It is measured only at a prepared site, and getting power, connectivity and the rest of the site infrastructure ready first is not a small matter. Nvidia has not confirmed the container size. The design does not solve access to grid power, the binding constraint on AI factories in Europe, and no customer orders or prices have been published.
Sources
Sources and references
Information checked: 18 November 2026
Related stories
SpaceX Falcon 9 NROL-136 Launch: What to Watch on 21 November
SpaceX Falcon 9 NROL-136 launch is scheduled for 21 November from Vandenberg, carrying reconnaissance satellites for the National Reconnaissance Office.
20 November 2026AO3 Returns After 17 Days Offline After Security Breach
AO3 returns after 17 days offline following a security incident, with some features still disabled, and went down again briefly on 19 November for maintenance.
20 November 2026ChatGPT for Accounting Launches With QuickBooks Link
ChatGPT for Accounting launched this week, connecting the chatbot to QuickBooks, Xero and NetSuite so finance teams can reconcile books and close the month.
20 November 2026Tesla Cybercab Rides Begin in Dallas With V1.5 Cars
Tesla Cybercab rides began in Dallas on 19 November, the second city after Austin to get the steering wheel-free robotaxi, with upgraded V1.5 vehicles.