NVIDIA unveils AI push for Europe at Berlin keynote

NVIDIA used Jensen Huang's keynote in Berlin to unveil new AI infrastructure, software and robotics products for Europe, framing the announcements as part of a broader push to expand its role in the region's AI buildout.

At GTC Berlin, Huang, the company's founder and chief executive officer, said European groups including L'Oréal, Revolut, Spotify, Merck, Vodafone, Zara, Porsche, Nestlé and HSBC are working with NVIDIA on AI projects. He also said the company has been travelling across Paris, London, Stockholm and Dublin, is working with more than 12,000 partners in Europe, and is committing nearly USD $10 billion to the region's technology ecosystem.

Huang laid out NVIDIA's argument that AI depends on a full stack, from energy supply and data centre infrastructure to models and applications. He said European companies and governments should build AI as well as use it.

"After decades of advancement in the computer industry, we are now at the beginnings of a new industrial revolution," said Huang.

Data centre design

One of the main announcements was NVIDIA DSX POD, a reference design for AI factories based on the company's Vera Rubin systems and built by partners inside modular containers. NVIDIA said the units are integrated and validated before delivery, reducing onsite deployment time from six months to 14 days once utilities are in place.

"This is a complete transformation of the world supply chain," said Huang. "This will change the way that people install and build data centers, and hopefully, as a result, velocity goes up."

NVIDIA also outlined the broader DSX AI Factory Platform, which it said can cut water use for cooling by up to 100% and deliver the same AI output with up to 30% less power. The pitch reflects growing focus among European operators and policymakers on energy consumption, cooling methods and the speed of data centre construction.

Agent safety

Huang devoted much of the keynote to AI agents, which he described as systems that do more than answer prompts because they can use memory, knowledge and external tools. He argued that security controls need to change with each deployment because agents may be given access to files, software, networks and other resources.

To address that, NVIDIA introduced the OpenShell Reference Architecture for Safe AI Agents. It is built on NVIDIA OpenShell, an open-source runtime that uses policy controls and sandboxing between an agent and the systems it can affect.

"The most important thing is to make sure the agent is secure," said Huang.

He added that NVIDIA is also launching the Agent Knowledge Foundation, a set of models and libraries aimed at linking agents to business data held in documents and databases. NVIDIA highlighted Nemotron tools for document parsing, query embedding and reranking, alongside new Kumo models for predictions from structured data.

"Ultimately, what turns these super agents, these agents into super agents, is that you give them access to your company's knowledge," said Huang.

Open models

NVIDIA also used the event to promote its portfolio of open models as starting points for industry-specific AI work. Huang listed Nemotron for language, Cosmos for world foundation models, GR00T for humanoid robotics, Alpamayo for autonomous vehicles, Ising for quantum-classical computing and Earth-2 for physics AI.

He said customers can post-train those models with their own datasets to create what he called "sovereign intelligence". The idea is likely to resonate in Europe, where debate over AI sovereignty, local control of data and industrial policy has intensified.

"You could start with an open model, something that's incredibly good like Nemotron, and with your own data, post-train it for your use case and achieve, of course, super intelligence," said Huang. "That is a very sensible thing to do, and all the technology exists today."

Industry use

The keynote included examples from life sciences, manufacturing, engineering, telecoms and transport. NVIDIA said Lila Sciences is using agents to help design a potential B-cell lymphoma therapy, Siemens is using agents to prepare products for production in a digital twin of its Erlangen factory, and Dassault Systèmes is using engineering agents to refine vehicle designs through simulation.

In telecoms, Huang said AI will change network operations and spectrum efficiency. He also said NVIDIA had made its first commercial phone call on T-Mobile's network with Nokia after almost eight years of work.

"Now CUDA makes phone calls," said Huang. "AI will revolutionize the way we do telecommunications. It will enhance our spectrum efficiency tremendously."

In healthcare, NVIDIA said it is working with Siemens Healthineers on AI-enabled radiation therapy that combines treatment-planning agents, digital twins and live imaging data. Huang described the system as a form of physical AI because it must track changes in the patient's body and tumour during treatment.

"The tumor is changing in shape. The physical body is changing while you're in therapy," said Huang. "The robot, this radio adaptive radiotherapy system, has to perceive the changing environment, understand where the tumor is, segment it just perfectly and direct the radiotherapy accordingly."

Robotics and vehicles

In transport, Huang said Mercedes-Benz, Volkswagen Commercial Vehicles, Lucid, Einride and Volvo Cars are among the groups adopting or working with NVIDIA's Hyperion platform and related software for automated driving. He also said Volkswagen's van business, JLR, Volvo and Mercedes-Benz next-generation vehicles will be Hyperion-compatible.

NVIDIA also announced Agile ONE SN, a reference humanoid built on Isaac GR00T with 31 degrees of freedom. Huang said it is intended to shorten setup times for robotics researchers from weeks to days.

He closed by showing RTX Spark laptops and a demonstration involving Hugging Face's Microduck robots, which he said were created on RTX Spark, trained in the cloud and tested through simulation before being deployed physically.

"Everybody can build agents. Every company in Europe will become an AI company ... and we're going to have a thriving AI economy here in Europe," said Huang.