What happened
NVIDIA announced DSX POD on 21 October 2026 at GTC Berlin. The company describes it as “a reference design for complete NVIDIA Vera Rubin AI factories built by partners inside modular containers, integrated and validated before delivery.”
Key details
- Deployment time: NVIDIA says factory-built units “reduce deployment time at a prepared site from six months to as little as 14 days.” The figure is NVIDIA’s claim and applies to a prepared site.
- How it works: partners complete cabling, wiring, piping and functional validation before the containers ship.
- Partners named by NVIDIA: Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn (Ingrasys), GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology, Schneider Electric, Vertiv and Wistron.
- Timing: NVIDIA says partners are to deliver DSX POD AI factories worldwide in early 2027.
- Efficiency figures: DatacenterDynamics reports water and power figures for NVIDIA’s wider DSX platform, but we did not find them in the DSX POD press release, so we do not repeat them as DSX POD claims.
Why it matters
The pitch is that building AI capacity in containers at a factory, rather than on site, shortens the gap between buying hardware and producing tokens, which is the bottleneck for many operators and countries. All performance and timing claims come from NVIDIA, no customer deployment is described, and delivery is not planned until 2027.
Disclosure: Claude, made by Anthropic, was one of the AI tools used to research and draft this article. Details come from NVIDIA’s press release of 21 October 2026, read the same day.