NVIDIA Unveils DSX POD to Speed AI Factory Deployment in Europe

NVIDIA has introduced DSX POD, a modular design for artificial intelligence factories intended to accelerate infrastructure deployment in Europe and other markets. Announced at GTC Berlin on May 20, 2026, the system uses factory-built containers housing NVIDIA Vera Rubin computing infrastructure, allowing operators to connect pre-integrated units to prepared sites and potentially begin production in as little as 14 days.

According to NVIDIA, the approach could reduce deployment time from approximately six months to two weeks. The company plans to work with a global network of computing and infrastructure partners to deliver DSX POD AI factories worldwide in early 2027.

The design moves much of the installation process from construction sites to manufacturing facilities. Partners integrate computing systems, networking equipment, cabling, rack-level power and liquid cooling inside modular containers before shipment, reducing the amount of specialized labor and technical work required at deployment locations.

NVIDIA said factory-based integration and validation could also reduce exposure to component delays and costly errors during on-site installation. The containers are designed to arrive ready for connection to a prepared site's power, cooling, networking and other supporting infrastructure.

Jensen Huang, founder and CEO of NVIDIA, said the commercial value of AI infrastructure depends on how quickly it begins processing production workloads.

"An AI factory creates value when it starts producing tokens — not when construction is complete," said Jensen Huang, founder and CEO of NVIDIA.

Huang said the new design aims to bring manufacturing-style precision to infrastructure construction while shortening the time needed to deploy computing capacity.

"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," said Jensen Huang, founder and CEO of NVIDIA.

DSX POD combines Vera Rubin systems with the supporting infrastructure required to operate them in standardized, containerized configurations. NVIDIA said partners will complete rack-to-rack cabling, wiring, piping and functional validation before shipping each unit, allowing site preparation to proceed separately from system integration.

The company identifies the time required to generate the first production token as a key measure of deployment progress. Once operational, an AI factory can run models for production workloads, allowing operators to provide computing services to customers.

NVIDIA said its DSX POD initiative involves Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn (Ingrasys), GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology (QCT), Schneider Electric, Vertiv and Wistron. The partners are expected to help manufacture and deliver the modular systems globally beginning in early 2027.

NVIDIA cautioned that its forward-looking statements regarding product availability, deployment speed, customer revenue and profitability are subject to risks and uncertainties. The company cited factors including global economic and political conditions, manufacturing dependencies, technological competition, product defects, market acceptance and regulatory changes.