ThinkPatternGet the app
Story
TECHNOLOGY · NOV 16, 2026

NVIDIA Unveils Rubin HPC GPU for Scientific Discovery

NVIDIA announced the Rubin HPC GPU and Nemotron for Science at SC26, introducing hardware and AI tools to accelerate complex scientific simulations by 2028.

NVIDIA introduced the Rubin HPC graphics processing unit and a suite of scientific research tools at the SC26 high-performance and supercomputing conference in Chicago. Scheduled for release in 2028, the Rubin HPC chip features over 50,000 CUDA cores and 7th Gen NVLink interconnect technology. The hardware includes a sparse data acceleration feature designed to optimize simulations in astrophysics and biology, and will be integrated into servers with Vera CPUs to potentially pack over 200,000 cores per tray.

Alongside the hardware, the company launched Nemotron for Science, an AI bundle based on the Nemotron 3.5 Super VL model, and several CUDA-X libraries. These include cuPhoton.xBeam for black hole data analysis, cuPhoton.xLens for fusion reactor plasma simulation, and the open-source cuSpa for cell polygon calculations. Ian Buck stated that these tools provide researchers with the models and data needed to build specialized agents that reason through scientific problems.

To advance quantum computing, NVIDIA is partnering with IBM Corp. and the U.S. Department of Energy's Fermilab. The collaboration aims to develop a modular quantum codesign platform that combines DGX B200 systems with IBM quantum hardware to improve software and error correction algorithms.


Reported across 5 outlets
Actors
NVIDIA CorporationIan BuckIBM CorporationFermilab

Keep reading in the app

The full story and every source, free in the app.