What if one system could run AI, simulations, and data workloads together? A new platform aims to do exactly that—at supercomputer scale.
NVIDIA has introduced the Vera Rubin platform, a new accelerated computing system designed to combine AI, simulation, and data analytics for scientific and industrial research. By integrating NVIDIA’s hardware, software, CUDA-X libraries, and AI technologies into a single platform, Vera Rubin turns each rack into a powerful supercomputing system capable of handling complex scientific workloads.
The platform is built to support applications such as climate modeling, computational fluid dynamics, quantum chemistry, energy exploration, and other data-intensive research tasks. It combines high-precision computing with AI capabilities, allowing researchers to run simulations, train AI models, analyze large datasets, and perform real-time analytics on the same system.
A Vera Rubin-based supercomputing system can provide more than 7 exaflops of AI computing performance, 5 petaflops of native FP64 performance, and support configurations with up to 144 GPUs. NVIDIA says this enables performance comparable to some of the systems listed in the TOP500 ranking of the world’s most powerful supercomputers, helping organizations run larger models and reduce time to results.
The platform combines NVIDIA Rubin GPUs and Vera CPUs connected through NVIDIA NVLink-C2C, ConnectX-9 SuperNICs, and BlueField-4 DPUs within a direct liquid-cooled architecture. It provides native FP64 support for high-accuracy simulations while also delivering AI performance for scientific foundation models, surrogate models, and AI-assisted analysis.
System manufacturers including Dell Technologies, HPE, GIGABYTE, Supermicro, and Bull are preparing direct liquid-cooled AI and HPC racks based on the Vera Rubin NVL4 architecture. These systems are intended to help research institutions, national laboratories, and enterprises deploy large-scale accelerated computing platforms for simulation, AI, data processing, and visualization.
“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.”
Click here for the original announcement.
