MENU
NVIDIA has unveiled its Vera Rubin accelerated computing platform, positioning it as a new generation of AI-enabled supercomputing infrastructure for scientific research, industrial simulation and high-performance computing (HPC). Announced at ISC High Performance 2026, the platform combines traditional double-precision computing with large-scale AI capabilities in a single architecture.
For eeNews Europe readers, the announcement highlights how AI and HPC are increasingly converging into unified systems. It also provides insight into the next wave of liquid-cooled, rack-scale computing platforms that will underpin scientific research, energy exploration and advanced engineering workloads over the coming years.
The Vera Rubin platform integrates NVIDIA Rubin GPUs and Vera CPUs connected through NVLink-C2C, ConnectX-9 SuperNICs and BlueField-4 DPUs within a direct liquid-cooled architecture.
According to NVIDIA, a fully configured Vera Rubin system can deliver more than 7 exaflops of AI performance for scientific workloads alongside 5 petaflops of native FP64 computing performance. With support for up to 144 GPUs per rack, the company claims a single rack can provide performance comparable to systems found on the TOP500 list of the world’s most powerful supercomputers.
The platform is aimed at computationally intensive workloads including climate modelling, computational fluid dynamics, quantum chemistry and energy exploration. By supporting simulation, AI model training, data analytics and real-time processing on a common infrastructure, NVIDIA is targeting research organisations seeking to combine traditional numerical methods with emerging agentic AI techniques.
“Scientific discovery is now a race between the complexity of the world’s greatest challenges and the computing systems built to solve them,” said Jensen Huang, founder and CEO of NVIDIA. “NVIDIA Vera Rubin is a new instrument for science — a rack-scale supercomputer that brings simulation, AI and data processing together to help researchers and industries design and discover faster than ever.”
Several major research centres have already selected Vera Rubin for future systems.
The Leibniz Supercomputing Centre (LRZ) in Germany plans to deploy the Blue Lion supercomputer, based on the Vera Rubin platform and second-generation exascale-class HPE Cray technology. Scheduled for 2027, the system is expected to deliver around 30 times the computing performance of LRZ’s current infrastructure.
In the US, the National Energy Research Scientific Computing Center will use Vera Rubin in the Doudna supercomputer at Lawrence Berkeley National Laboratory. The Dell Technologies-built system is intended to support large-scale HPC workloads, AI training and data-intensive scientific research.
Meanwhile, Los Alamos National Laboratory has selected Vera Rubin technology for its upcoming Mission, Vision and Veritas systems. The systems will address national security, open science and AI-driven scientific discovery workloads.
NVIDIA is also expanding its partner ecosystem around the platform. Bull, Dell Technologies, GIGABYTE, HPE and Supermicro are all developing direct liquid-cooled Vera Rubin NVL4 systems designed for deployment in research institutes, national laboratories and enterprise environments.
The company said Vera Rubin NVL4-based systems are expected to become available from global system manufacturers in the fourth quarter of this year, marking the start of a broader rollout of rack-scale AI and HPC infrastructure.
All material on this site Copyright © 2026 European Business Press SA. All rights reserved.
Leave a Reply