Nvidia packs Vera Rubin AI factories into prebuilt containers

Nvidia introduced DSX POD at GTC Berlin on Oct. 21. It is a reference design for complete Vera Rubin AI factories that partners build inside modular containers, then integrate and validate before shipping. Nvidia says the units arrive ready to connect to a prepared site's power, cooling and networking, cutting deployment from six months to as little as 14 days.

That clock starts once the site is ready. Nvidia's keynote summary says the timeline applies after site utilities are in place, so preparing the site and its utilities, including the power connection, sits outside it. What moves to the factory is the work inside the container: computing, networking, cabling, rack-level power and liquid cooling.

Jensen Huang framed the product around revenue. In his keynote, as reported by Data Center Knowledge, he said a gigawatt-scale data center costs $50 billion and that a three-month delay can be worth $10 billion in revenue, figures that are his own estimates. In Nvidia's release he said an AI factory creates value when it starts producing tokens.

Nvidia named 17 partners, including Dell Technologies, HPE, Lenovo, Schneider Electric and Vertiv, to deliver DSX POD factories in early 2027. Yahoo Finance reported Nvidia's statement that the pods will not replace conventional data centers and are a way to start AI work before a permanent facility is ready. Container-based data centers already exist, Data Center Knowledge noted, so Nvidia's bet is on standardizing them.

Nvidia also says its DSX designs pair closed-loop liquid cooling with dry coolers and could save 54 million liters of water a day at a 2-gigawatt deployment compared with cooling towers, though some climates still need supplemental cooling. Greenpeace protesters interrupted the keynote shortly before the announcement, citing water supplies and electricity bills. Deliveries in 2027 will show whether factory-built pods shorten real projects or move the wait to power.