Europe has launched its largest-ever expansion of AI supercomputing infrastructure, with 35 new NVIDIA-powered AI and high-performance computing (HPC) systems now under development across 23 countries. The initiative will provide next-generation AI infrastructure to more than 3 million researchers and marks a significant step in Europe’s effort to strengthen technological sovereignty, scientific research, and industrial innovation. The announcement was made at ISC High Performance 2026, where NVIDIA revealed that the new systems will support a broad range of applications, including climate science, healthcare, clean energy, quantum computing, advanced manufacturing, and fundamental scientific research. According to NVIDIA, more than 90% of Europe’s AI Factory deployments now rely on NVIDIA infrastructure, with approximately 800 AI exaflops either deployed or announced since 2025.
Europe Accelerates Its AI Infrastructure Buildout
Europe’s latest wave of AI infrastructure reflects a growing effort to build sovereign computing capacity capable of supporting both public-sector and industrial AI initiatives. The 35 systems span national supercomputing centers, EuroHPC AI Factories, research institutions, and university-led programs. Together, they represent the largest one-year deployment of supercomputing infrastructure in Europe’s history. NVIDIA’s Blackwell and Hopper platforms form the backbone of most of these deployments, while NVIDIA Quantum InfiniBand networking, CUDA-X libraries, NIM microservices, and AI Enterprise software provide the software stack needed for AI training, simulation, inference, and agentic AI workloads.
NVIDIA founder and CEO Jensen Huang described AI as a transformative scientific instrument capable of accelerating discoveries across multiple disciplines. “AI is the new instrument of science, and Europe is building the infrastructure to put it in the hands of millions of researchers,” Huang said during the announcement.
AI Factories Drive Continental Expansion
Several flagship projects are leading Europe’s AI Factory buildout.
MareNostrum5 AI Upgrade
The Barcelona Supercomputing Center will significantly expand its MareNostrum5 system through the EuroHPC AI Factory initiative. The upgraded infrastructure will deploy NVIDIA GB300 NVL72 and NVIDIA GB200 NVL4 systems connected through NVIDIA Quantum-X800 InfiniBand networking. The upgraded platform is expected to deliver approximately 20 exaflops of AI training performance and 33 exaflops of AI inference capability. Researchers will use the system for generative AI development, climate modeling, biotechnology, healthcare research, agriculture, energy systems, and government AI services. Barcelona Supercomputing Center Director Mateo Valero Cortés said the project will provide European researchers with advanced tools to address some of the world’s most complex scientific challenges.
BavariaAI’s Blue Swan
Germany’s BavariaAI initiative is developing the Blue Swan platform across Friedrich-Alexander University Erlangen and the Leibniz Supercomputing Centre. The deployment includes 1,000 GPUs based on NVIDIA GB200 NVL4 systems and Quantum-2 InfiniBand networking. Once operational, Blue Swan will deliver up to 11 exaflops of AI training performance and 22 exaflops of AI inference performance. The platform will support Bavaria’s efforts to build open multimodal foundation models for healthcare, robotics, public administration, perception systems, and scientific research.
Italy’s IT4LIA Project
Italy’s AI Factory initiative, IT4LIA, represents one of the largest announced deployments. The project will utilize more than 8,000 GPUs through NVIDIA GB200 NVL4 systems and Quantum-X800 networking. According to NVIDIA, the infrastructure will deliver approximately 82 exaflops of AI training capability and 164 exaflops of AI inference performance. The system will support AI model development across cybersecurity, meteorology, manufacturing, climate research, and agriculture while strengthening Europe’s broader AI ecosystem
Germany’s HammerHAI
Germany’s first AI Factory, HammerHAI, will deploy more than 850 NVIDIA GPUs connected through Quantum-X800 InfiniBand. The facility is expected to provide approximately 8 exaflops of AI training performance and 15 exaflops of AI inference capability. Researchers and industrial users will use the platform for engineering simulations, large language model inference, and advanced scientific workloads.
Sweden’s Mimer AI Factory
The Mimer AI Factory, hosted at Linköping University and owned by the EuroHPC Joint Undertaking, will deploy 100 NVIDIA GB200 NVL4 systems comprising 400 GPUs. The system will provide approximately 4 exaflops of AI training performance and around 7 exaflops of AI inference capability. Applications will span life sciences, materials research, autonomous systems, trustworthy AI, and data-driven innovation.
AI Expands Into Climate and Energy Research
Alongside supercomputing deployments, NVIDIA is supporting projects focused on climate science and energy decarbonization. Researchers are increasingly using AI infrastructure to improve Earth system modeling, biomedical research, carbon capture technologies, fusion energy research, and hydrogen-based energy systems. These applications require massive computational resources and advanced simulation capabilities. One notable example involves Siemens Energy, which is leveraging NVIDIA technologies within the Siemens Xcelerator portfolio. The company uses NVIDIA Omniverse libraries, CUDA-X software, and accelerated computing infrastructure to design gas turbine burners capable of operating on up to 100% hydrogen. These simulations address complex fluid dynamics, combustion behavior, and extreme heat conditions. According to NVIDIA, the workflow reduces simulation times by up to 77%, enabling faster validation of hydrogen-capable low-carbon energy technologies.
Europe Deepens Leadership in Quantum-GPU Computing
The expansion also strengthens Europe’s position in hybrid quantum-classical computing. Several leading institutions are integrating quantum processing units (QPUs) with NVIDIA’s CUDA-Q platform, which provides a unified environment for developing and running hybrid quantum applications. At Italy’s CINECA supercomputing center, EuroHPC and Pasqal are integrating a neutral-atom quantum processor that will support optimization and materials science applications. Meanwhile, Fraunhofer FOKUS is integrating CUDA-Q with the Eclipse Qrisp quantum programming language, helping researchers develop and optimize complex quantum algorithms.
The Barcelona Supercomputing Center has also deployed a new analog quantum computer from Qilimanjaro Quantum Tech. The company has integrated CUDA-Q into its QiliSDK software stack and plans to enable seamless quantum-accelerated workflows within the center’s research environment. In Germany, researchers at the Jülich Supercomputing Centre and NVIDIA recently achieved a new milestone by fully simulating a universal 50-qubit quantum computer on the JUPITER supercomputer using NVIDIA GH200 Grace Hopper Superchips. The achievement establishes a new benchmark for large-scale quantum simulation and provides researchers with the ability to test increasingly complex quantum workloads before deployment on physical quantum hardware.
Europe Builds the Infrastructure for the AI Era
The deployment of 35 new AI supercomputers marks a defining moment for Europe’s AI strategy. By combining AI factories, sovereign compute infrastructure, advanced networking, climate research, and quantum computing initiatives, Europe is building a foundation designed to support scientific discovery and industrial competitiveness for years to come. With 800 AI exaflops already deployed or announced and millions of researchers gaining access to next-generation computing resources, the continent is positioning itself as one of the world’s largest centers for AI-driven innovation.
