NVIDIA details physical AI work with Samsung, Kakao Mobility and LG in Korea
NVIDIA used AI Day Seoul to present physical AI projects with Korean manufacturers, covering semiconductor engineering, robotics and autonomous vehicles. Samsung cut a chip-scale thermal and mechanical analysis from about 18 hours on a 32-core CPU to 41 seconds on a GPU, roughly a 1,600x speedup, using AI-predicted material properties, a coarser grid and GPU acceleration. Kakao Mobility is targeting a Level 4 autonomous driving system, and LG Electronics is training factory robots. Samsung also uses cuISS and cuDSS for TCAD and reports up to 5x faster quantum chemistry runs with cuEST.
