A record 35 Nvidia AI HPC supercomputers are in development across Europe — equipping more than 3-million researchers with next-generation infrastructure for continental AI, accelerated science and industrial innovation.
The systems represent Europe’s largest one-year expansion of supercomputers, spanning national supercomputing centres, AI factories and academic research institutions. Built on full-stack Nvidia AI infrastructure, the systems will support research across climate science, healthcare, clean-energy decarbonization, quantum computing and fundamental science.
The Nvidia Blackwell and Nvidia Hopper platforms are powering the majority of Europe’s AI factory buildout, with 800 AI exaflops deployed or announced since last year. With Nvidia Quantum InfiniBand networking, Nvidia CUDA-X libraries, Nvidia NIM™ microservices and Nvidia AI Enterprise software, Nvidia provides a full-stack platform for science, spanning model training, simulation, inference and agentic AI.
“AI is the new instrument of science, and Europe is building the infrastructure to put it in the hands of millions of researchers,” says Jensen Huang, founder and CEO of Nvidia. “With Nvidia accelerated computing, researchers can simulate more complex systems, train scientific AI models and build agentic AI workflows that turn Europe’s data and expertise into breakthroughs for the world.”
Supercomputers including Barcelona Supercomputing Centre’s EuroHPC MareNostrum5 AI upgrade, BavariaAI’s Blue Swan, IT4LIA, HLRS’s HammerHAI and NAISS’s Mimer EuroHPC AI Factory are among those based on advanced Nvidia AI infrastructure.
Europe’s latest AI infrastructure expansion includes advancements at:
- Barcelona Supercomputing Centre’s AI Factory, the first EuroHPC AI-specific installation: Will expand MareNostrum 5 with Nvidia GB300 NVL72 and Nvidia GB200 NVL4 systems, connected by the Nvidia Quantum-X800 InfiniBand platform. Delivering up to approximately 20 exaflops of AI training and 33 exaflops of AI inference performance, the system will accelerate generative AI, climate modeling, health and biotech research, sustainable agriculture, energy systems and government AI services.
- BavariaAI’s Blue Swan: Brings 1,000 GPUs via Nvidia GB200 NVL4 systems and Nvidia Quantum-2 InfiniBand networking to FAU Erlangen and LRZ supercomputing centres. Delivering up to 11 exaflops of AI training and 22 exaflops of AI inference performance, the platform will support Bavaria’s foundation model initiative, advancing open multimodal models for science, public administration, health research, robotics and perception.
- IT4LIA: An AI factory with over 8,000 GPUs via Nvidia GB200 NVL4 systems, Nvidia Quantum-X800 InfiniBand networking and Nvidia AI Enterprise software, delivering 82 exaflops of AI training and 164 exaflops of AI inference performance.
- HLRS’s HammerHAI: Will equip Germany’s first AI factory with over 850 GPUs via Nvidia GB200 NVL4 systems connected with Nvidia Quantum-X800 InfiniBand. Delivering up to approximately 8 exaflops of AI training and 15 exaflops of AI inference performance, HammerHAI will give researchers and industrial users secure AI infrastructure for engineering simulation, large language model inference and scientific discovery.
- NAISS’s Mimer AI Factory, owned by the EuroHPC Joint Undertaking: Hosted at Linköping University, this supercomputer will deploy 100 Nvidia GB200 NVL4 systems, totaling 400 GPUs, with Nvidia ConnectX®-8 networking. Delivering up to 4 exaflops of AI training and about 7 exaflops of AI inference performance, Mimer AI Factory will advance Sweden’s AI science ecosystem across life sciences, materials research, autonomous systems, trustworthy AI and data-driven innovation.
AI for climate and decarbonisation
Nvidia is supporting initiatives that deploy AI infrastructure and software to help researchers apply AI to scientific challenges including climate and Earth systems modeling, biomedical research, and clean-energy technologies such as fusion, hydrogen and carbon capture.
Accelerated simulation and industrial engineering are already enabling breakthroughs in clean-energy research.
Siemens Energy is using the Siemens Xcelerator portfolio, accelerated by Nvidia technologies, including Nvidia Omniverse libraries, CUDA-X and AI infrastructure, to unify design, computational fluid dynamics simulation and manufacturing for gas turbines built to run on up to 100% hydrogen — a complex physics challenge involving extreme heat, fluid dynamics and combustion behavior.
The workflow supports rapid, simulation-driven design iterations for complex gas turbine burner configurations, followed by fast technology validation using additively manufactured combustors. This cuts simulation times by up to 77% to advance hydrogen-capable, low‑carbon gas turbines.
Quantum-GPU supercomputing advances
Nvidia is enabling European supercomputing centres and institutions to develop and run useful hybrid quantum-classical applications using Nvidia CUDA-Q, an open, qubit-agnostic platform for hybrid computing, extending Europe’s leadership in quantum-GPU supercomputing.
CINECA, EuroHPC and Pasqal are integrating a neutral-atom QPU at the CINECA supercomputing centre. The Pasqal hybrid environment is deploying the CUDA-Q platform through integration with Slurm. CUDA-Q provides Pasqal and CINECA with a platform for developing and running quantum applications including for optimization and materials science use cases.
Fraunhofer FOKUS is facilitating the integration of Nvidia CUDA-Q with the quantum programming language Eclipse Qrisp. Qrisp, initiated at Fraunhofer FOKUS and being further developed by the Eclipse Foundation, enables researchers to more easily write complex quantum algorithms which can then be simulated, optimised and run with Nvidia CUDA-Q.
The Barcelona Supercomputing Centre has recently deployed a new analog quantum computer from QPU builder Qilimanjaro Quantum Tech, through the EuroHPC JU initiative. Qilimanjaro integrated Nvidia CUDA-Q into its quantum software development kit, QiliSDK. Qilimanjaro is also working toward making its QPU available in the Nvidia CUDA-Q platform for seamless control of quantum accelerated workflows at BSC.
Researchers at the Jülich Supercomputing Centre, working with a team from Nvidia, set a world record by fully simulating a universal 50-qubit quantum computer. The simulation was run on JUPITER, powered by its Nvidia GH200 Grace Hopper Superchips. Jülich’s JUQCS-50 (for 50 qubits) quantum simulator allows researchers to now test the largest possible quantum problems on supercomputers, to scale quantum computing.