# Nvidia wants AI factories built in real factories

DevFeed: [Nvidia wants AI factories built in real factories](<https://devfeed.tech/articles/nvidia-wants-ai-factories-built-in-real-factories-92750.md>)

Original publisher: [Read original article](<https://thenewstack.io/nvidia-dsx-pod-containers/>)

Author: Frederic Lardinois

Published: 2026-10-21T11:30:00Z

Content type: news

Language: en

Sources: [The New Stack](<https://devfeed.tech/sources/the-new-stack.md>)

Topics: [NVIDIA DSX](<https://devfeed.tech/topics/nvidia-dsx.md>), [NVIDIA Vera Rubin](<https://devfeed.tech/topics/nvidia-vera-rubin.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2016](<https://devfeed.tech/tags/2016.md>), [about](<https://devfeed.tech/tags/about.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-factories](<https://devfeed.tech/tags/ai-factories.md>), [ai-factory](<https://devfeed.tech/tags/ai-factory.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [aws](<https://devfeed.tech/tags/aws.md>), [berlin](<https://devfeed.tech/tags/berlin.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [conference](<https://devfeed.tech/tags/conference.md>), [data](<https://devfeed.tech/tags/data.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [gtc](<https://devfeed.tech/tags/gtc.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [show](<https://devfeed.tech/tags/show.md>), [up](<https://devfeed.tech/tags/up.md>), [wants](<https://devfeed.tech/tags/wants.md>)

## AI overview

Nvidia's DSX POD is a factory-preassembled data-center building block intended to bring new AI capacity online faster. Each shipping container holds seven Vera Rubin NVL72 racks (504 GPUs), with cabling and networking installed and tested; after delivery, crews connect the unit to its neighbors and complete burn-in tests. Nvidia says it can be ready to produce tokens in as few as 14 days, provided the site already has power, cooling and network connections--a significant caveat because those can be bottlenecks. The design uses warm coolant and dry coolers instead of water-consuming cooling towers, though operators in some climates may need chillers on the hottest days. Nvidia estimates dry cooling could save about 54 million liters of water daily in a 2-gigawatt deployment. Partners are expected to begin worldwide deliveries in early 2027.

## Source excerpt

Nvidia wants new AI capacity to show up at the data center fully built. At its GTC Berlin conference on The post Nvidia wants AI factories built in real factories appeared first on The New Stack.