JUPITER, Europe's first exascale supercomputer powered by NVIDIA hardware, is delivering breakthrough results across neu
Europe's first exascale supercomputer, JUPITER, located at Germany's Forschungszentrum Jülich and running on NVIDIA Grace Hopper Superchips and NVIDIA Quantum-X800 InfiniBand networking, is proving the transformative potential of exascale computing across a remarkably wide range of scientific domains.
As the international supercomputing community gathered at ISC in Hamburg, four major projects running on JUPITER demonstrated what exascale systems can now accomplish. According to Thomas Lippert, director of the Jülich Supercomputing Centre and professor at Goethe University Frankfurt, "With JUPITER, Europe doesn't just join the exascale era — it leads it, across the widest range of science and AI of any system worldwide."
The Jülich Brain Atlas project has produced CytoNet, a foundation model for brain microarchitecture analysis. Led by neuroscientist Katrin Amunts and computer scientist Christian Schiffer at Jülich's Institute of Neuroscience and Medicine, the model learns from brain imaging data at cellular scale to map individual cell structures and their relationship to broader patterns of brain organization. With 86 billion neurons and about 100 trillion connections between them, understanding the human brain at single neuron resolution was previously out of reach. Training on JUPITER took under five days using 6.5 petabytes of data from 21 post-mortem brains on 4,096 NVIDIA Grace Hopper Superchips. Amunts emphasized the shift in approach: "For the first time, we're not just using AI to analyze the brain — we're building an agent that can think through the experiment itself. That changes what neuroscience will be, and JUPITER is what makes that sentence possible to say today."
A novel ICON configuration developed by researchers at ETH Zurich, the German Climate Computing Centre, Jülich Supercomputing Centre, Max Planck Institute for Meteorology, NVIDIA, Swiss National Supercomputing Centre and the University of Hamburg won the Gordon Bell Prize for Climate Modelling at SC25 last November. ICON is the first model to simulate a coupled Earth system at one-kilometer resolution, including ocean, atmosphere and land with full biogeochemistry and the complete carbon cycle exchanged between all components. Running on 20,480 NVIDIA Grace Hopper Superchips on JUPITER, the model simulated approximately 146 days of real climate into 24 hours of compute, setting a world record in global climate simulation. Daniel Klocke from the Max Planck Institute for Meteorology explained: "At a global resolution of just one kilometer, many of these interactions emerge directly from the laws of physics rather than being approximated. This gives us an unprecedented view of how the atmosphere, ocean and biosphere work together."
In March, Ericsson and Forschungszentrum Jülich announced a collaboration to develop AI for the continued evolution of 5G and for 6G networks, with JUPITER as the compute engine. The work targets brain-inspired architectures designed to handle complex network operations at far lower energy costs, including AI models for Ericsson's radio and core networks and energy-efficient AI inference at the radio edge.
Researchers at the Jülich Supercomputing Centre, working with the NVIDIA Application Lab, achieved a world first by fully simulating a universal 50-qubit quantum computer, surpassing the previous 48-qubit record. This breakthrough was made possible by JUPITER's coherent, tightly coupled CPU-GPU memory architecture, which allows data exceeding GPU limits to spill seamlessly into CPU memory with minimal performance loss. The simulator, JUQCS-50, will be accessible through JUNIQ, the quantum computer user facility at JSC, enabling researchers to design and stress-test the algorithms that future quantum hardware will run.