AMD and Eviden to Power France’s First Exascale Supercomputer

AMD and Eviden, the advanced computing brand of the Atos Group, have been selected to build Alice Recoque, a groundbreaking next-generation exascale supercomputer designed to meet Europe’s growing needs in high-performance computing (HPC) and artificial intelligence. Once deployed, Alice Recoque will become France’s first and Europe’s second exascale system, marking a major milestone in Europe’s digital, scientific and industrial sovereignty.

Led by the Grand équipement national de calcul intensif (GENCI) and operated by the Commissariat à l’énergie atomique et aux énergies alternatives (CEA), this system represents one of the most ambitious and expensive HPC projects ever launched in Europe. The supercomputer leverages the latest generation of AMD’s HPC and AI compute technology, tightly integrated into Eviden’s cutting-edge BullSequana platform to deliver unprecedented computational capability and energy efficiency.

A Pan-European Investment in Sovereign Computing

The overall project represents an investment of 554 million euros, funded through the EuroHPC Joint Undertaking (EuroHPC JU). Funding comes from the Digital Europe Programme (DEP) and the Jules Verne Consortium, a multinational coalition led by France through GENCI and CEA, with key support from SURF in the Netherlands and GRNET in Greece.

The scale and collaborative nature of the project underscore Europe’s long-term ambition: to develop sovereign, energy-efficient computing infrastructure capable of supporting both fundamental science and industrial innovation across the continent.

AMD and Eviden: Powering the Next Generation of European Supercomputing

In their announcement, AMD and Eviden emphasized how Alice Recoque will strengthen Europe’s scientific independence and accelerate progress in emerging fields like generative AI, climate modeling, healthcare, digital twins and advanced manufacturing.

We are committed to enabling the next generation of innovation across AI and HPC,” said Dan McNamara, senior vice president and general manager of Compute & Enterprise AI at AMD.
“The Alice Recoque supercomputer represents a major step forward for European sovereign AI, uniting national ambition, regional collaboration, and AMD’s high-performance and AI compute technologies. Through our collaboration with EuroHPC JU, the Jules Verne Consortium, and Eviden, we are proud to support Europe’s scientific and industrial leadership with a platform purpose-built for scale, efficiency and discovery.

Eviden, which has delivered some of the world’s largest supercomputing systems, highlighted the system’s importance for both sustainability and European autonomy in key technology sectors.

Alice Recoque represents another critical step toward Europe’s digital future, defined by sovereignty, sustainability, and scientific excellence,” said Emmanuel Le Roux, Group SVP and Global Head of Advanced Computing and AI at Eviden.
As a catalyst for breakthroughs in climate modeling, healthcare, advanced materials and AI innovation, it will empower researchers and industries across Europe. Born from a shared European vision, this AI Factory reflects what we can achieve collectively toward a common goal. Eviden is fully dedicated to its success, bringing deep expertise, a collaborative spirit and a long-term commitment to responsible technological leadership.

Alice Recoque: A Full AI-HPC Factory for Europe’s Most Complex Challenges

Unlike traditional supercomputers focused primarily on simulation, Alice Recoque is engineered as a holistic AI-HPC factory. It integrates compute, storage, networking, advanced software stacks, and real-world AI use cases into a unified platform capable of tackling Europe’s most pressing scientific, social and industrial needs.

This includes:

  • Large-scale simulations for earth systems, climate, physics and engineering
  • Massive data analysis, enabling insights from previously unmanageable datasets
  • AI training and inference at scales required for sovereign large AI models
  • Digital twins, enabling real-time modeling of cities, materials, human health and energy systems
  • Industrial R&D acceleration, covering everything from drug discovery to microelectronics

Eviden and AMD describe it as a “live implementation” of their long-standing collaboration. Beyond hardware, the partners are investing heavily in human expertise and AI readiness to ensure research organizations and industries can deploy impactful solutions right away.

Powered by AMD’s Most Advanced AI and HPC Technology

Alice Recoque integrates several next-generation AMD technologies designed specifically for large-scale AI and scientific computing:

Next-gen AMD EPYC™ “Venice” CPUs

Optimized for HPC and AI workloads, offering high memory bandwidth, energy efficiency and next-generation compute density.

AMD Instinct™ MI430X GPUs

The newest accelerators in the MI400 Series, engineered for:

  • sovereign AI compute
  • exascale-class HPC
  • leadership floating-point performance
  • next-generation AI datatypes, including FP4 and FP8

Each MI430X GPU contains:

  • 432 GB of HBM4 high-bandwidth memory
  • 19.6 TB/s of memory bandwidth
  • Optimized leadership AI FLOPs
  • Advanced support for mixed-precision workloads

This configuration allows Alice Recoque to support extremely large AI models, deep scientific simulations, and multi-physics digital twins—all at unmatched memory capacity per accelerator.

AMD FPGAs

Integrated as specialized components for data preprocessing, acceleration of specific algorithms, and advanced data movement operations.

Eviden’s BullSequana XH3500: HPC Designed for Exascale Sustainability

The supercomputer is built on Eviden’s BullSequana XH3500 architecture, interconnected through Eviden’s high-performance BXI networking, and supported by DDN storage.

The BullSequana XH3500 platform is engineered for:

  • Scalability up to and beyond exascale
  • Dense compute floors with fewer racks
  • Reduced power consumption
  • Direct Liquid Cooling (5th generation) using warm water to cool all major components

Alice Recoque will occupy 94 racks, yet thanks to Eviden’s architecture:

  • It uses 25% fewer racks and fewer components than comparable exascale systems.
  • It delivers up to 50% better energy efficiency per GPU.

These advantages are crucial as Europe pushes for greener supercomputing infrastructure to meet climate and emissions goals.

Energy-Aware Computing Through Software Intelligence

Beyond hardware efficiency, Alice Recoque incorporates Eviden’s Argos intelligent software, which provides:

  • advanced real-time monitoring of system thermals and performance
  • intelligent power management
  • energy optimization for AI and HPC workloads
  • tools that help developers maximize efficiency and throughput per watt

This integration ensures the system runs at peak performance while minimizing environmental impact.

Enabling Breakthroughs Across Europe

Alice Recoque is expected to drive transformative advances across fields such as:

Climate & Earth-System Modeling

Exascale resolution enables finer spatial and temporal modeling, improving predictions for extreme weather, ocean dynamics and long-term climate impact.

Material Science & Energy Innovation

Large-scale simulations will accelerate the design of next-generation semiconductors, batteries, clean-energy systems and advanced manufacturing processes.

Healthcare & Personalized Medicine

Digital twins of organs, molecules and biological systems could support precision treatments, clinical decision tools and drug discovery.

Sovereign AI Models

With massive memory and FP4/FP8 support, Alice Recoque can train and deploy large foundation models under European governance and ethical frameworks.

A New Era for European Supercomputing

With Alice Recoque, France and the European Union take a decisive step toward becoming global leaders in sovereign AI and exascale HPC. Blending cutting-edge AMD technologies with Eviden’s proven HPC engineering, the system is designed not just as a supercomputer, but as a continent-wide resource for scientific discovery, industrial innovation and sustainable digital transformation.

Source Link:https://www.amd.com/en/newsroom

Share your love