Nvidia introduced DSX POD at GTC Berlin on October 21, a reference design for complete Vera Rubin AI factories that partners build inside modular containers and validate before shipping. Nvidia says the approach cuts deployment at a prepared site from six months to as little as 14 days.
DSX POD moves the integration of computing, networking, cabling, rack-level power and liquid cooling from the data center site to the manufacturing facility. Partners finish rack-to-rack cabling, wiring, piping and functional validation before shipment, separately from site preparation, so units arrive ready to connect to the site's power, cooling and networking.
Seventeen partners are working with Nvidia on the systems: Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn, GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology, Schneider Electric, Vertiv and Wistron. They plan to deliver DSX POD AI factories worldwide in early 2027.
Key facts
- 01Nvidia introduced DSX POD at GTC Berlin on October 21, 2026, as a reference design for Vera Rubin AI factories.
- 02Nvidia says DSX POD cuts deployment at a prepared site from six months to as little as 14 days.
- 0317 partners, including Dell Technologies, HPE, Lenovo, Foxconn and Schneider Electric, are building the systems.
- 04Partners plan to deliver DSX POD AI factories worldwide in early 2027.
- 05Nvidia claims its DSX design saves 54 million liters of water a day at 2GW scale versus cooling towers.
Nvidia pitches the speed as a revenue argument. Factory-tested containers reduce exposure to component delays and error-prone onsite work, which the company says gets customers to model training and production inference, and to generating tokens, sooner.
“An AI factory creates value when it starts producing tokens — not when construction is complete”— Jensen Huang, Founder and CEO of NVIDIA
Nvidia founder and CEO Jensen Huang said DSX POD brings 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.
DSX POD is a modular version of Nvidia's broader DSX AI Factory Platform, the AI data center reference architecture first unveiled at GTC San Jose earlier this year. Nvidia says that platform can cut water used for cooling by up to 100% and use up to 30% less power for the same output.
The cooling design pairs closed-loop liquid cooling with outdoor dry coolers that release heat without evaporating water. At 2GW scale, Nvidia claims that saves 54 million liters of water a day compared with cooling-tower systems, while noting that some climates may need extra mechanical cooling during extreme heat.
On power, Nvidia's DSX MaxLPS allocates power dynamically within a facility's provisioned budget, which the company says yields the same token output with 30% less power. DSX Flex is designed to let data centers adjust workload power use to grid needs, demand spikes and energy prices.
The 14-day figure assumes a prepared site: power, cooling and networking still have to be in place before the containers arrive.
Nvidia also used GTC Berlin to launch its OpenShell Reference Architecture for AI agent safety, expand its open model portfolio with a focus on European sovereign AI, and outline partnerships with Nokia, European automakers and neocloud providers.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
