NVIDIA Expands CUDA-X Libraries to Accelerate Scientific Discovery
Reported by Forbes, NVIDIA, Business Insider and 1 more outlets
NVIDIA says expanded CUDA-X libraries, including cuPhoton, cuPotential and cuSpa, can speed data processing and simulations across fields from astrophysics to biology. The company cites performance gains in telescope data analysis, molecular simulation and single-cell research.
Key points
- NVIDIA says cuPhoton.xBeam cut Event Horizon Telescope data correlation from over a month to eight minutes
- NVIDIA says cuPotential speeds existing molecular simulation software by up to 3x
- NVIDIA says Daviator calculations can run up to 22x faster at up to 75% lower cost
- Newly open-source cuSpa reportedly delivered a 900x speedup for cell polygon calculations at Memorial Sloan Kettering
- NVIDIA says rapids-singlecell is available in a single agent-ready container
Sources
- ForbesRead ↗AI Raises The Stakes For Discovery Data
- NVIDIARead ↗NVIDIA CUDA-X Libraries Give Agents the Tools to Accelerate Scientific Discovery
- Business InsiderRead ↗NVIDIA Expands Vera CPU Platform to Accelerate Scientific Discovery
- Business InsiderRead ↗NVIDIA Introduces Nemotron for Science to Accelerate Discoveries With Agents
- StartupHub AIRead ↗NVIDIA CUDA-X libraries give agents real science jobs
Details
NVIDIA announced expanded CUDA-X libraries for scientific computing on Nov. 16. The tools include cuPhoton for processing scientific data, cuPotential for molecular simulations and cuSpa for cell polygon calculations. NVIDIA says they are intended to help researchers analyze larger datasets and model more complex systems.
For astrophysics, NVIDIA says cuPhoton.xBeam reduced Event Horizon Telescope observation correlation from more than a month to eight minutes. The company says the toolkit also correlated observations of black hole jets J1924 and 3C279 and can run the reconstruction pipeline on distributed multi-GPU systems or a DGX Spark. Pacific Fusion researchers used cuPhoton.xLens to examine plasma simulations as they ran; NVIDIA says a locally running Nemotron-powered agent can be coupled to the pipeline for interactive analysis.
NVIDIA says cuPotential accelerates existing molecular simulation software by up to 3x. The company is working with Schrödinger and communities behind AMBER, NAMD, GROMACS, LAMMPS and the Open Molecular Software Foundation to bring the acceleration to their workflows, and named Cadence, Fathom Therapeutics, Topos Bio, Genesis Molecular AI and Terray Therapeutics as adopters.
For materials research, NVIDIA says Daviator kernels for the Vienna Ab initio Simulation Package are coming to NVIDIA ALCHEMI, with quantum-mechanical calculations up to 22 times faster and at up to 75% lower cost. The company also reports a 900x CPU speedup for open-source cuSpa at Memorial Sloan Kettering, 200x faster data processing by Therna Biosciences with rapids-singlecell, and benchmark searches up to 84x faster with its GPU-accelerated HMMER library at Basecamp Research.
Written from coverage as of
Images
- NVIDIA
- Forbes
- StartupHub AI
- StartupHub AI
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh