Bull & NVIDIA

The complexity of AI and Exascale computing requires more than hardware. It requires deep integration, joint engineering, performance optimisation and a shared commitment to sustainability. Bull and NVIDIA work together from silicon to system integration to software optimisation, delivering a complete, future-ready intelligent computing platform. 

Bull and NVIDIA combine deep system architecture expertise with industry-leading accelerated computing technology. Together, we deliver fully integrated AI and HPC solutions, from silicon to software to sustainable infrastructure, optimised for performance, efficiency and scalability. 

Bull offers a comprehensive portfolio of AI and HPC systems powered by NVIDIA Grace CPUs and NVIDIA GPU-accelerated computing, spanning from enterprise AI platforms to exascale-class supercomputers:

Exascale & HPC Systems

• BullSequana XH3500 modular supercomputing platform integrating NVIDIA GH200 Grace Hopper Superchips and advanced Direct Liquid Cooling, designed for large-scale HPC, AI training and exascale workloads.

• BullSequana XH3000 architecture underpinning top-ranked energy-efficient supercomputers (Top 3 Green500-ranked systems), powered by NVIDIA GH200 Grace Hopper Superchips.

• JUPITER Booster (BullSequana XH3000 architecture) European Exascale system leveraging NVIDIA accelerated computing to support large-scale scientific simulation, AI and advanced research programmes.

AI Platforms (Mid-Range to Large-Scale AI)

• BullSequana AI 600 series GPU-accelerated AI servers supporting NVIDIA GPUs for enterprise AI training, inference and data analytics workloads.

• BullSequana AI 620 series, including configurations such as AI 621 and AI 622 platforms powered by NVIDIA L40S and NVIDIA H100 GPUs, designed for scalable enterprise AI deployments.

• BullSequana AI 640 series – Including AI 641 configurations integrating NVIDIA HGX platforms with H100 or H200 GPUs, delivering high-density AI acceleration for large-scale model training.

• Gen AI in a Box solutions. Pre-integrated AI infrastructure combining NVIDIA GPUs, NVIDIA AI software stack and Bull hardware, enabling rapid deployment of production-ready AI environments.

HPC & AI Optimisation Services

• Centre for Excellence in Performance Programming (CEPP) supporting application modernisation and CPU-to-GPU porting using NVIDIA CUDA and NVIDIA libraries to unlock accelerated performance across industry verticals.

Enterprise Platforms

• BullSequana SA series enterprise-class 1- and 2-socket servers that support NVIDIA GPU acceleration for AI, virtualisation and data-intensive workloads, enabling organisations to integrate accelerated computing into business-critical environments.

This portfolio demonstrates how Bull integrates NVIDIA Grace CPUs, GH200 Superchips and advanced GPU platforms across its accelerated computing systems, delivering scalable performance, energy efficiency, and AI innovation from enterprise deployments to exascale supercomputing.

The partnership enables organisations to train AI models faster, run complex simulations more efficiently and scale workloads confidently. By combining NVIDIA GPUs with Bull’s HPC architecture and Direct Liquid Cooling, customers achieve high performance with improved energy efficiency. 

Bull’s advanced Direct Liquid Cooling and high-density system design, integrated with NVIDIA accelerated computing platforms, reduce energy consumption while maintaining peak performance. This approach supports both performance leadership and environmental responsibility. 

Bull and NVIDIA collaborate on large-scale programmes such as JUPITER, combining advanced architecture, GPUs and optimisation expertise to deliver Exascale-class performance for scientific discovery and industrial innovation. 

Industries including healthcare, climate science, automotive, manufacturing, financial services and public research benefit from accelerated AI training, faster simulations and improved energy efficiency.