NVIDIA unveils Vera Rubin DSX POD: containerized AI factory deploys in 14 days

At GTC Berlin, Nvidia introduced the Vera Rubin DSX POD reference design, packing a complete AI factory into a modular shipping container, targeting one of the biggest bottlenecks for AI data centers: deployment and scaling speed. According to multiple accounts, on a prepared site the deployment cycle can shrink from 6 months to 14 days—going from months down to days—offering a plug-and-play solution for rapidly growing AI compute demand.

Confirmed

• The DSX POD is a full-stack, modular AI factory reference architecture built on the Vera Rubin platform; each container includes 17 kilometers of cabling and ships factory pre-integrated and pre-tested
• Partners including Dell, HPE, Lenovo, Schneider Electric, and Vertiv are involved, with the units manufactured in their factories
• Also announced at the event was the OpenShell Reference Architecture, likewise aimed at rapid deployment of AI data centers/factories
• The post gives two official framings of the speedup—"months down to days" and "6 months down to 14 days"—with the latter applying to prepared sites

2026-10-21 ~ 2026-10-21 · 7 related posts

• [source] NVIDIA unveils DSX POD: an AI supercomputer in a shipping container with 17km of wiring — LinusEkenstam · 2026-10-21
• NVIDIA Unveils DSX POD, Cutting Datacenter Scaling From Months to Days — kimmonismus · 2026-10-21
• NVIDIA DSX cuts AI factory cooling water to near zero and power use by 30% — nvidia · 2026-10-21
• NVIDIA DSX POD slashes AI factory deployment from six months to 14 days — nvidia · 2026-10-21
• [source] NVIDIA's new OpenShell architecture cuts AI factory deployment from 6 months to 14 days — testingcatalog · 2026-10-21
• [source] NVIDIA unveils Vera Rubin DSX POD, shipping-container AI factories that deploy in days — NandoDF · 2026-10-21
• NVIDIA's DSX POD: a 1GW supercomputer costs $50B but yields $40B a year — LinusEkenstam · 2026-10-21