NVIDIA expands CUDA-X libraries for scientific research
NVIDIA
· November 16, 2026
· ✓ verified
NVIDIA has announced expanded CUDA-X libraries for scientific computing and AI-assisted research across astrophysics, fusion, drug discovery, materials science, and biology.
- The update highlights cuPhoton, cuPotential, cuSpa, and related tools such as cuPhoton.xBeam, cuPhoton.xLens, Daviator, DAQIRI, and rapids-singlecell; NVIDIA says these tools help researchers process larger datasets and model more complex systems faster.
- Specific examples include Event Horizon Telescope black hole analysis dropping from over a month to eight minutes, cuPotential accelerating molecular simulation software by up to 3x, Daviator delivering up to 22x faster calculations, and cuSpa providing a 900x speedup versus CPUs; the article also cites adoption by groups including Pacific Fusion, Schrödinger, Cadence, Therna Biosciences, and 10x Genomics.