NVIDIA Collaborates with Korean Industry Leaders to Enhance Physical AI

9 November 2026 2 min read هدهد تك | Hudhud Tech

NVIDIA is collaborating with South Korean industry leaders to enhance the country’s physical AI ecosystem, leveraging its strengths in manufacturing and engineering. This partnership aims to apply accelerated computing, AI, and simulation across various sectors, including semiconductor manufacturing, robotics, and autonomous vehicles.

Samsung Electronics is expanding its collaboration with NVIDIA to drive semiconductor innovation through accelerated computing and AI physics. Their joint efforts encompass materials research, process development, and device engineering.

In technology computer-aided design, Samsung utilizes NVIDIA cuISS and NVIDIA cuDSS to expedite computations involving sparse linear systems. The collaboration also includes GPU acceleration for process and optical simulations. Samsung’s sMOL platform has achieved up to a fivefold increase in speed for selected quantum chemical calculations using NVIDIA cuEST.

Additionally, Samsung is exploring NVIDIA cuEquivariance to enhance tensor operations for atomistic AI models. The company is developing chip-scale thermal and mechanical analysis workflows with NVIDIA PhysicsNeMo, achieving a significant speedup in AI homogenization cases, reducing simulation time from 18 hours on a 32-core CPU to just 41 seconds on a GPU.

Looking ahead, the NVIDIA Agent Toolkit could integrate these simulation tools and AI models, enabling future engineering agents to conduct analyses and compare design options more efficiently.

Kakao Mobility is also collaborating with NVIDIA to advance its level 4 autonomous driving system. This initiative combines Kakao Mobility’s proprietary driving software with NVIDIA’s vehicle computing platform, including NVIDIA Hyperion, which serves as a reference platform for vehicle sensors and computing.

The initial development will focus on Kia’s purpose-built vehicles, integrating Kakao Mobility’s software with NVIDIA technologies to support future level 4 mobility services in Korea.

Meanwhile, LG Electronics has successfully demonstrated robotic manipulation and manufacturing tasks using NVIDIA Isaac GR00T-powered robots in a real manufacturing environment. This collaboration established a physical AI data flywheel that connects real-world data collection with NVIDIA Cosmos-based data augmentation and GR00T model training.

The teams are refining a repeatable process for training, evaluating, and improving robots for manufacturing, ensuring continued development and validation of LG’s CLOiD robotic capabilities.