LATEST

← Back to GPN · All articles

NVIDIA Announces Rubin HPC, a GPU With More Than 50,000 CUDA Cores

NVIDIA Announces Rubin HPC, a GPU With More Than 50,000 CUDA Cores

OFFICIAL Written by GPN · Nov 16, 2026, 1:00 PM ET · NVIDIA Rubin HPC · Tech

NVIDIA has announced Rubin HPC at SC26, a new GPU in the Vera Rubin platform designed to accelerate simulation and data processing for science, engineering and analytics work.

The GPU carries a record number of CUDA cores — more than 50,000 on a single GPU — along with sparse data acceleration to keep those cores fed with data, and seventh-generation NVLink for GPU-to-GPU connectivity. The design balances data movement and compute together rather than increasing compute alone, so the cores spend their time computing instead of waiting on data.

The pitch is aimed at the agents now helping scientists and engineers: as agents explore more candidate materials, molecules and designs, they lean on simulations to test those ideas and guide their next steps. Faster simulation means faster feedback, shortening the loop of hypothesis, simulation and refinement. Ian Buck, NVIDIA's vice president of hyperscale and high-performance computing, said: "For agents tackling science and engineering, faster simulation means faster discovery."

Rubin HPC is expected to be available in 2028. NVIDIA's CUDA-X libraries offer optimized implementations of the algorithms agents use for simulation and analytics, and paired with Rubin HPC they give agents the fastest path to reasoning outcomes, according to NVIDIA.

Facts gathered from

More on this