Nvidia Corp ( NASDAQ:NVDA , XETRA:NVD ) has unveiled a modular AI factory design that could reduce on-site deployment from 6 months to 14 days, as chief executive Jensen Huang pushes Europe to build its own artificial intelligence infrastructure.
The DSX POD reference design packages computing hardware, networking, power systems and liquid cooling into factory-integrated modules that can be delivered ready for connection once a site has the required utilities.
NVIDIA introduced the system at its GTC Berlin conference on Wednesday, targeting businesses and infrastructure operators seeking to bring AI computing capacity online sooner and begin generating revenue earlier.
The design will be delivered through a global partner network that includes Dell Technologies, Hewlett Packard Enterprise, Lenovo, Schneider Electric and Vertiv, with deployments planned to begin in early 2027.
The announcement supports Huang's wider argument that Europe needs to produce its own AI systems and services, rather than depend on technology developed elsewhere.
NVIDIA has committed nearly $10 billion to Europe's technology ecosystem and works with more than 12,000 partners across the region, according to the US chipmaker's keynote presentation.
Huang described AI infrastructure as a combination of energy, computing hardware, data centres, models and applications, arguing that investment in the underlying systems must precede widespread adoption.
The keynote also introduced NVIDIA's OpenShell reference architecture, which aims to help businesses deploy AI agents with security controls that restrict access to files, networks and digital tools.
AI agents are software systems that can use models, information and digital tools to perform tasks with limited human intervention, creating additional security risks when they access corporate systems.
NVIDIA also highlighted its open AI models for language, robotics, autonomous driving and scientific computing, which businesses can adapt using their own data.
The company's industrial demonstrations included AI agents supporting manufacturing at Siemens, engineering simulations at Dassault Systèmes and scientific research at Lila Sciences.
NVIDIA's broader strategy is to expand beyond supplying graphics processing units, or GPUs, into the infrastructure, software and development tools needed to build and operate AI systems.
The success of its latest initiatives will depend on whether businesses can deploy the technology economically, secure sufficient power and convert additional computing capacity into sustained demand for AI services.
