JUPITER Supercomputer Powered by NVIDIA Grace Hopper Becomes Europe's Fastest, On Track for Exascale

Hosted at Jülich Supercomputing Centre in Germany, JUPITER delivers over 2x the performance of Europe's next-fastest system using nearly 24,000 NVIDIA GH200 Grace Hopper Superchips, ranking fourth on the global TOP500 list while claiming the continent's top spot for both HPC and AI workloads.

FE
FIRAT Editorial BoardInstitutional Research Desk
Jun 10, 2025
5 min read
Share:
JUPITER Supercomputer Powered by NVIDIA Grace Hopper Becomes Europe's Fastest, On Track for Exascale

Jülich, Germany · 10 June 2025

NVIDIA announced at the ISC 2025 high-performance computing conference that the JUPITER supercomputer, powered by the NVIDIA Grace Hopper platform, has become the fastest supercomputer in Europe. The system, hosted at the Jülich Supercomputing Centre at the Forschungszentrum Jülich facility in Germany, delivers more than a 2x speedup for high-performance computing and AI workloads compared with the next-fastest system on the continent.

JUPITER ranked fourth on the June 2025 TOP500 list of the world's fastest supercomputers and is on track to become Europe's first exascale system, capable of running 1 quintillion FP64 operations per second. The system is owned by the EuroHPC Joint Undertaking, a European Union initiative established to develop a world-class supercomputing infrastructure across Europe.

Architecture and Specifications

JUPITER comprises nearly 24,000 NVIDIA GH200 Grace Hopper Superchips, interconnected with the NVIDIA Quantum-2 InfiniBand networking platform. The system is built on Eviden's BullSequana XH3000 liquid-cooled architecture and is expected to reach over 90 exaflops of AI performance once fully operational.

The GH200 Grace Hopper Superchip combines NVIDIA's Hopper GPU architecture with the Arm-based Grace CPU on a single module, connected by a high-bandwidth coherent interconnect. This design eliminates the traditional PCIe bottleneck between CPU and GPU, enabling the system to handle both large-scale AI model training and traditional HPC simulation workloads within a unified architecture.

Scientific Applications

JUPITER is designed to serve a broad range of scientific and industrial research domains. The system's computing power is directed toward several priority areas:

Research DomainApplicationNVIDIA Platform
Climate and weather modelingHigh-resolution real-time environmental simulations and digital twin of EarthEarth-2
Quantum computing researchQuantum algorithm and hardware developmentCUDA-Q, cuQuantum
Computer-aided engineeringAI-driven simulation for product design and manufacturingPhysicsNeMo, CUDA-X, Omniverse
Drug discoveryAI models for pharmaceutical research and biomolecular scienceBioNeMo

The system's ability to handle both AI training and traditional scientific simulation within a single architecture represents a shift in how supercomputing resources are deployed. Rather than maintaining separate systems for AI workloads and conventional HPC simulations, JUPITER allows researchers to move fluidly between both paradigms.

European Sovereignty and Access

JUPITER was delivered by an Eviden-led consortium, with the system designed, built, and installed in less than nine months — a timeline that Emmanuel Le Roux, senior vice president and global head of advanced computing at Eviden, Atos Group, described as demonstrating the technological leadership of the European consortium.

"With JUPITER's extreme performance, Europe has taken a giant leap into the future of science, technology and sovereignty," said Anders Jensen, executive director of the EuroHPC Joint Undertaking. "JUPITER's computing power will serve as a catalyst for scientific discovery, propelling foundational research across the continent in fields as diverse as climate modeling, energy systems and biomedical innovation."

Thomas Lippert, codirector of the Jülich Supercomputing Centre, emphasized the system's role in advancing the frontier of foundation model training and high-performance simulation, enabling researchers across Europe to tackle challenges of unprecedented complexity.

Kristel Michielsen, codirector of the Jülich Supercomputing Centre, noted that JUPITER will substantially advance quantum algorithm and hardware development, with hybrid quantum-HPC computation benefiting from tools such as the NVIDIA CUDA-Q platform and the NVIDIA cuQuantum software development kit.

Context and Significance

The announcement positions Europe within the top tier of global supercomputing. The world's fastest supercomputer as of June 2025 remains El Capitan, hosted at Lawrence Livermore National Laboratory in the United States and powered by AMD Instinct MI300A APUs. JUPITER's fourth-place global ranking, combined with its energy efficiency leadership, marks a significant milestone for European computing infrastructure.

Jensen Huang, founder and CEO of NVIDIA, framed the announcement within the broader context of AI-driven scientific discovery: "AI will supercharge scientific discovery and industrial innovation. In partnership with Jülich and Eviden, we're building Europe's most advanced AI supercomputer to enable the leading researchers, industries and institutions to expand human knowledge, accelerate breakthroughs and drive national advancement."

German and other European researchers can apply for access to JUPITER through the EuroHPC Joint Undertaking's regular access mode call for proposals.

Sources

  • NVIDIA Newsroom, "NVIDIA Powers Europe's Fastest Supercomputer," 10 June 2025
  • Forschungszentrum Jülich, "JUPITER Supercomputer Propels European Computing Power," June 2025
  • EuroHPC Joint Undertaking, call for proposals for regular access mode
  • TOP500 List, June 2025 edition
Filed Under:#Supercomputing#NVIDIA#HPC#Exascale#EuroHPC#AI Infrastructure

Share this research insight

Help circulate peer-reviewed evidence and institutional briefings.

Share:
FE
Author SpotlightDivision: ReMIT

FIRAT Editorial Board

Institutional Research Desk · Foresight Institute of Research and Translation

The collective editorial and research translation board of FIRAT, synthesising peer-reviewed evidence, policy briefs, and division milestones across our seven foundational research pillars.

Focus:Institutional PolicyResearch StrategyAfrican DevelopmentInnovation
More Research

Related Articles in Science Technology Engineering and Mathematics (STEM)

View all in STEM