Nvidia launches DSX POD, aims for modular, preintegrated AI factories

Nvidia launched DSX POD AI factories in a move designed to speed up deployments with modular containers that are integrated and validated before delivery. According to Nvidia, the DSX POD reference design can speed up deployment time at a prepared site from six months to as little as 14 days.

The news announced at GTC Berlin aims to address an obvious issue--it takes too long to spin up AI data centers. Modular AI factories could speed up AI factory deployments and feature a smaller footprint. The catch is that you need a prepared site, which also needs power.

Nevertheless, Nvidia's DSX POD will cut integration time. Computing, networking, cabling, rack-level power and liquid cooling are integrating at the manufacturing stage. That upfront integration will cut some of the need for specialized labor, planning and construction at the site.

• Nvidia makes case AI factories, its stack are enduring capital assets

Jensen Huang, CEO of Nvidia, outlined the launch during his GTC Berlin keynote. He emphasized that producing tokens faster leads to more revenue. "With Nvidia DSX POD, we’re bringing factory-built precision and speed to AI infrastructure construction — reducing deployment time from months to weeks so customers can scale capacity and bring revenue-generating production online faster," he said.

Huang added that DSX POD will be shipping globally in 2027. IT is working with its various hardware partners including Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn, GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology, Schneider Electric, Vertiv and Wistron to ship DSX POD systems.

Other items from GTC Berlin include:

• Nvidia launched its OpenShell Reference Architecture for AI agent safety.
• The company expanded its open model portfolio with new additions and pitched Europe, which is building for sovereign AI.
• Nvidia outlined partnerships with Nokia as well as Europe's major automakers and neoclouds.

More:

• Nvidia contributes rack, tray designs to OCP
• Nvidia launches Open Agent Safety Platform to secure AI agents
• Nvidia’s Nemotron hit enterprise AI inflection point