NVIDIA has unveiled a sweeping initiative to expand AI infrastructure across Europe, deploying over 3,000 exaflops of Blackwell-based compute to support sovereign AI development, industrial transformation, and scientific research. Announced at NVIDIA GTC Paris at VivaTech, the campaign includes national collaborations with France, the U.K., Italy, and Germany, as well as partnerships with European cloud providers and telcos such as Domyn, Mistral AI, Nebius, Nscale, Orange, Telefónica, Swisscom, Fastweb, and Telenor. These efforts aim to establish Europe as a leader in AI innovation while upholding digital sovereignty and regional language diversity.
In Germany, NVIDIA is building the world’s first industrial AI cloud — a massive data infrastructure powered by 10,000 Blackwell GPUs through DGX B200 systems and RTX PRO Servers. Designed to support industrial workloads such as simulation, digital twins, and AI-powered robotics, the factory will accelerate design and production processes for leading European manufacturers. BMW, Maserati, Mercedes-Benz, and Schaeffler are already leveraging this infrastructure to redesign their product development lifecycles using NVIDIA Omniverse and CUDA-X software stacks. Ansys, Siemens, and Cadence are also optimizing their platforms for Blackwell, boosting simulation speeds and enabling AI-first design flows in sectors like automotive, aerospace, and semiconductor manufacturing.
To ensure national AI capacity, France, the U.K., and Italy are working with regional players to build dedicated cloud infrastructure based on NVIDIA’s Grace Blackwell systems. In France, Mistral AI will deploy 18,000 Blackwell nodes across multiple sites by 2026 to power an end-to-end AI cloud platform. In the U.K., cloud partners Nebius and Nscale plan to roll out 14,000 Blackwell GPUs for scalable AI services. Italy’s Domyn is developing its Colosseum supercomputer for regulated industries, such as finance and healthcare, using NVIDIA’s AI stack. These deployments are positioned to support the training and inference of large language models (LLMs), agentic AI, and physical AI applications for enterprises, startups, and the public sector.
NVIDIA is also working closely with European telecom providers to integrate sovereign AI infrastructure into their national networks. Orange is leveraging NVIDIA systems to offer LLM and personal assistant services via its Cloud Avenue platform. Fastweb has introduced MIIA, an Italian LLM trained on its NVIDIA DGX system. Swisscom is deploying its GenAI Studio and AI Workhub on a DGX SuperPOD™, while Telefónica is piloting edge AI services across Spain with hundreds of NVIDIA GPUs. Telenor is building a renewable-powered AI data center in Norway and supporting a partner’s multilingual AI translation system available in over 100 languages. These telco deployments allow enterprises to build and run agentic AI models within secure and compliant regional environments.
To address the critical need for skills and research, NVIDIA is establishing and expanding AI Technology Centers in Germany, Sweden, Italy, Spain, the U.K., and Finland. These centers support AI talent development through NVIDIA’s Deep Learning Institute and foster local research in disciplines like digital medicine, stable diffusion, embodied AI, and Earth systems modeling. In Germany, the Bavarian AI Center will collaborate with the Bayern KI consortium, while Italy and Spain will host new AI factories via the CINECA consortium and Barcelona Supercomputing Center, respectively.
Europe’s AI diversity is further enriched by sovereign LLMs being optimized through NVIDIA’s Nemotron techniques. Builders such as BSC (Spain), Dicta (Israel), Domyn (Italy), H Company and LightOn (France), and academic groups from Sweden, Slovakia, the U.K., and others are refining models that reflect local languages and cultures. These models will be accessible via NVIDIA’s DGX Cloud Lepton and deployed through microservices called NVIDIA NIMs. New integrations with Hugging Face and the Perplexity AI engine will allow enterprises to fine-tune or run inference on sovereign models in real time with cost-efficient infrastructure, preserving data privacy and improving accuracy.
The sovereign models will be distilled for faster inference, improved token generation, and reduced operational costs, benefiting sectors such as public health, education, and national defense. NVIDIA’s collaboration with Perplexity, a leading AI-powered answer engine, ensures that regional models can serve high-value use cases with credible sourcing and in-line citations — empowering European institutions to conduct research, customer service, and decision support in native languages.
“In the era of AI, every manufacturer needs two factories: one for making things, and one for creating the intelligence that powers them,” said Jensen Huang, founder and CEO of NVIDIA. “By building Europe’s first industrial AI infrastructure, we’re enabling the region’s leading industrial companies to advance simulation-first, AI-driven manufacturing.”
Highlights by Region and Sector:
- France: Mistral AI deploying 18,000 Grace Blackwell systems in a phased rollout across multiple sites.
- U.K.: Nebius and Nscale launching new data centers with 14,000 NVIDIA Blackwell GPUs.
- Germany: 10,000-GPU AI factory to support industrial AI workloads in design, robotics, and factory simulation.
- Italy: Domyn’s “Colosseum” supercomputer to support national AI initiatives and regulated sectors.
- Telcos:
- Orange: Cloud Avenue to offer enterprise-grade AI including personal assistants.
- Fastweb: Trained Italian LLM “MIIA” on its NVIDIA DGX AI supercomputer.
- Telenor: Launching renewable-powered AI data center and multilingual AI translation services.
- Swisscom: GenAI Studio and AI Workhub running on NVIDIA DGX SuperPOD.
- Telefónica: Deploying distributed edge AI fabric across Spain.
- AI Centers:
- Bavaria (Germany), CINECA (Italy), BSC (Spain), and national hubs in the U.K., Sweden, and Finland focused on upskilling and scientific AI research.
- Sovereign Models:
- Open-source LLMs optimized with NVIDIA Nemotron techniques for over 24 European languages.
- Integrated into Perplexity AI and Hugging Face with real-time inference via DGX Cloud Lepton.
“Europe’s diversity is its superpower — an engine of creativity and innovation,” said Jensen Huang. “Together with Europe’s model builders and cloud providers, we’re building an AI ecosystem where intelligence is developed and served locally to provide a foundation for Europe to thrive in the age of AI — transforming every industry across the region.”
