NVIDIA Brings Physical AI Into Korean Factories: Samsung Homogenization Case Drops From 18 Hours to 41 Seconds
Synopsis
NVIDIA and several Korean industry partners are pushing physical AI into production, spanning semiconductor engineering, autonomous driving and robot learning. In chip-scale thermal and mechanical analysis, Samsung Electronics combined AI-predicted effective material properties, coarser grids and GPU acceleration to cut one homogenization case from about 18 hours on a 32-core CPU fine-grid solve to 41 seconds on a GPU coarse-grid solve, roughly 1,600x faster; selected quantum chemical calculations on its sMOL platform ran up to 5x faster end to end.
Interpretation
Physical AI is moving from development into real-world production, with Korean manufacturing and engineering players applying accelerated computing, AI and simulation to semiconductor manufacturing, robotics and autonomous vehicles. Previously these efforts sat in development rather than real production. Samsung Electronics, Kakao Mobility and LG Electronics each disclosed collaborations or demonstrations in their own industrial settings, giving concrete cases of that shift.
Samsung Electronics is extending its collaboration to materials research, process development and device engineering, using GPU acceleration for sparse linear systems and for process and optical simulations, and speeding selected quantum chemical calculations on its sMOL platform. Previously such computations relied on CPU and fine-grid solves, making a single chip-scale analysis slow. Selected quantum chemical calculations on Samsung's sMOL platform ran up to 5x faster end to end with cuEST; in one AI homogenization case, AI-predicted effective material properties, coarser grids and GPU acceleration gave roughly 1,600x speedup, with simulation time falling from about 18 hours for a fine-grid solve on a 32-core CPU to 41 seconds for a GPU-accelerated coarse-grid solve.
Kakao Mobility plans to develop a level 4 autonomous driving system on NVIDIA's vehicle computing and safety technologies, while LG Electronics demonstrated NVIDIA Isaac GR00T-powered manipulation and manufacturing tasks with its CLOiD robot in an actual manufacturing environment. LG's demonstration took place in a real manufacturing environment rather than only in laboratory validation. Kakao Mobility plans to use the Hyperion reference platform powered by DRIVE AGX Thor and Halos, with initial development around Kia purpose-built vehicles; LG's physical AI data flywheel connects real-world data collection, Cosmos-based data augmentation, GR00T model training, Jetson AGX Thor onboard computing and simulation-based evaluation.
Perspective
The report covers what Samsung Electronics, Kakao Mobility and LG Electronics disclosed: Samsung's acceleration figures come from selected quantum chemical calculations and one AI homogenization case, Kakao Mobility's level 4 system and use of Hyperion and Halos are plans, and LG's demonstration concerns CLOiD robot manipulation and manufacturing tasks in an actual manufacturing environment.
Samsung's 1,600x speedup corresponds to one AI homogenization case and the 5x to selected quantum chemical calculations on the sMOL platform, not to whole workflows generally; Kakao Mobility's level 4 development and future mobility services are stated as plans; NVIDIA Agent Toolkit connecting simulation tools and AI models so future engineering agents can run analyses and compare design options remains a look-ahead.