NVIDIA Says DSX POD Cuts AI Factory Deployment From Months to Weeks

NVIDIA introduced DSX POD on October 21, 2026, at GTC Berlin, a reference design for complete Vera Rubin AI factories that partners build inside modular containers and that are integrated and validated before delivery. The company said the factory-built units reduce deployment time at a prepared site from six months to as little as 14 days.

According to NVIDIA, the containerized units arrive ready to connect out of the box to a prepared site’s power, cooling, networking and other supporting infrastructure. The company said the faster deployment lets customers begin generating production tokens and earning revenue sooner.

Factory Integration Moves Offsite

According to NVIDIA, DSX POD brings Vera Rubin systems together with the infrastructure needed to operate them in repeatable, factory-built, containerized designs that partners can manufacture and deliver at scale. The design moves the integration and installation of computing, networking, cabling, rack-level power and liquid cooling off the site and into the manufacturing facility, and NVIDIA said this reduces the planning, construction, specialized labor and logistics required for a traditional onsite build. The company said factory-tested containers also reduce exposure to component delays and costly, error-prone onsite work, creating a shorter path to model training and production inference.

Before shipment, partners complete rack-to-rack cabling, wiring, piping and functional validation, work that NVIDIA said proceeds separately from site preparation and further shortens the time needed to bring new capacity online. The reference design carries forward the design philosophy of the broader NVIDIA DSX infrastructure platform while using a distinct architecture that integrates and validates systems in purpose-built containers before transport and deployment.

“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 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. NVIDIA defines time to first token as the point when a new AI factory begins running models to serve production workloads, and the company described that earlier start to production as the link between deployment speed and the economics of operating an AI factory.

The Broader DSX AI Factory Platform

In a companion announcement dated October 21, 2026, NVIDIA described NVIDIA DSX as a full-stack AI factory reference architecture, co-innovated with infrastructure companies, in which compute, networking, storage, software, power and cooling are codesigned as a single system. The platform includes DSX MaxLPS for dynamic power allocation, DSX Flex for coordinating workload power consumption with grid conditions, DSX Reference Design for validated infrastructure architectures, DSX Sim for simulation, DSX OS for infrastructure management and DSX Exchange for connecting computing and facility systems.

The companion release said NVIDIA is expanding its modular, open source DSX OS portfolio with three components in developer preview: NVIDIA Config Manager automates AI factory network configuration, NVIDIA Rack Management Service provides rack-level power, inventory, firmware and switch management, and NVIDIA Cluster Readiness Engine runs distributed workloads to validate GPU clusters before production workloads arrive and isolate failures to specific nodes.

NVIDIA said DSX MaxLPS dynamically allocates power within an AI factory’s provisioned budget, helping operators generate the same number of tokens with up to 30% less power, while DSX Flex lets operators adapt workload power consumption to utility needs, demand-response events and energy prices. Lambda used DSX MaxLPS in a proof of concept across NVIDIA HGX B200 systems and demonstrated 24% more cluster token throughput and 19% more GPU capacity within a fixed power budget, according to the release. The company also said its reference designs pair liquid-cooled computing with dry-cooler-based heat rejection, which can reduce facility cooling water consumption to near zero.

According to the companion release, AI clouds and design, construction, energy and cooling providers are using DSX to build and operate infrastructure serving Europe’s enterprises, developers and governments. Warren Barkley, vice president of DSX product management at NVIDIA, said DSX brings together the technologies and ecosystem expertise operators need to build responsibly and operate efficiently at scale.

Partner Ecosystem and Early 2027 Deliveries

NVIDIA named Anduril, Armada, Bull, Crusoe, DayOne, Dell Technologies, Delta, Eaton, Foxconn (Ingrasys), GIGABYTE, HPE, Lenovo, Pegatron, Quanta Cloud Technology (QCT), Schneider Electric, Vertiv and Wistron as the computing and infrastructure partners working to deliver DSX POD AI factories.

NVIDIA said the partners will deliver DSX POD AI factories worldwide in early 2027.