[Edaily Reporter Han Kwangbeom ] NVIDIA will host “NVIDIA AI Day Seoul 2026,” an artificial intelligence (AI) developer event, in Seoul for two days starting on the 9th. Major Korean companies, including SamsungElectronics, SK hynix, LGELECTRONICS, Naver Cloud, and SKTelecom, will participate to share use cases ranging from AI model development to applications in robotics, semiconductors, and data centers.

NVIDIA announced on the 2nd that it will hold “NVIDIA AI Day Seoul 2026” at COEX in Seoul on the 9th and 10th. Approximately 2,000 people, including developers, startups, corporate representatives, and researchers, are expected to attend. The event will feature technical presentations and demonstrations focused on “agent-based AI,” where AI autonomously performs various tasks, “physical AI,” applied to robots operating in the real world, and large-scale AI infrastructure.

On the first day of the event, the 9th, there will be keynote sessions by key NVIDIA executives, as well as 15 technical sessions across three categories. The following day will feature hands-on training for developers and special programs for researchers and startups. Registration for the event has already closed.

◇Samsung and SK hynix Innovate Semiconductor Design and Manufacturing with AI

On the morning of the first day, Deepu Tala, Vice President of Robotics and Edge AI at NVIDIA, will deliver a presentation on the current state and future of Physical AI. He plans to introduce how NVIDIA technologies—including the robotics development platform “Isaac” and the world model “Cosmos”—are being utilized in Korea’s manufacturing, automotive, and semiconductor industries.

Jonathan Cohen, Vice President of Applied Research at NVIDIA, will explain the company’s strategy for AI agents and the open model ecosystem. Jeong Gu-hyung, Team Manager of Solution Architects at NVIDIA, and Jesse Clayton, Senior Product Marketing Manager for AI Infrastructure, will discuss NVIDIA’s “DSX” technology for building AI factories and the power and infrastructure challenges faced by Korean companies.

In the afternoon, the focus will be on real-world use cases from Korean companies. SamsungElectronics(005930)will present its approach to accelerating fluid, thermal, and structural analysis using “PhysicsNeMo,” NVIDIA’s physics-based AI development tool, and automating design exploration through AI agents. SK hynix will introduce methods for accelerating semiconductor device and process simulations using NVIDIA’s accelerated computing library “CUDA-X” and PhysicsNeMo.

Naver (NAVER(035420)) Cloud will share its experience with NVIDIA’s open AI model, “Nemotron,” and training large-scale AI models. LG Corp. AI Research, SKTelecom(017670), and Upstage will participate in the “K-AI Frontier Builders” panel to discuss their experiences using NVIDIA technology in the development and deployment of domestic sovereign AI models.

Technologies for the safe operation of AI agents will also be discussed. Korean AI company AIM Intelligence will introduce methods for controlling AI agents’ access permissions and execution environments using NVIDIA’s “Nemo Guardrail,” “OpenShell,” and “NemoClow.”

◇From LGELECTRONICS’ robot training to Naver’s AI infrastructure operations

In the field of physical AI, LGELECTRONICS(066570)will showcase a training system utilizing NVIDIA’s robot AI model, “Isaac GR00T.” The company plans to explain the process—from data collection to training, simulation, and on-site validation—along with real-world factory application examples. Doosan Robotics will introduce how to use NVIDIA’s physics simulation technology to replicate complex contact scenarios—such as robots grasping or deforming objects—in a virtual environment.

Domestic startups will also participate. N.DotLight will present technology that converts design data, images, and text into 3D digital assets for simulation, while Superb AI will present a method for generating synthetic data using NVIDIA Cosmos. RLWRLD will share use cases for benchmarks that evaluate a robot hand’s object manipulation capabilities.

In the AI infrastructure sector, technical challenges in the construction and operation of data centers will be a key agenda item. SKTelecom will discuss power issues in AI factories—from power grids to semiconductor chips—and present a plan for building large-scale data centers based on NVIDIA DSX. Naver Cloud will share its experience in design, fault response, and stabilizing AI model training, accumulated through the stable year-round operation of clusters comprising thousands of GPUs.

Coupang Intelligent Cloud will explain the training and inference architecture of its rack-level infrastructure—based on NVIDIA’s latest AI system, the “GB300 NVL72”—and its serverless AI platform. Vessel AI will demonstrate an AI-based operations platform that intelligently manages GPU resource allocation and the operational status of AI factories.

◇ Hands-On Experience: Building Agent AI and Training Robots

NVIDIA is also organizing programs that allow developers to experience AI technology firsthand. On the 9th, the “Build-a-Claw” hands-on program will be held in the COEX lobby. Participants can use OpenClaw and NVIDIA Nemoclaw to build and run their own AI agents. Demonstrations of always-on AI agents running on NVIDIA hardware and physical AI will also take place.

On the 10th, paid hands-on training sessions hosted by the NVIDIA Deep Learning Institute (DLI) will be held. NVIDIA solution architects will conduct training in Korean on topics such as the deployment and optimization of large-scale AI inference systems, Isaac-based robot learning and validation, and the development of safe AI agents.

At the “AI for Science Forum” taking place on the same day, domestic researchers—including Kim Hyun-woo, Director of the NVIDIA-KAIST Joint Research Center—will present case studies of scientific research utilizing AI. The K-Moonshot Project and the “Hangang” supercomputer from the Korea Institute of Science and Technology Information (KISTI) will also be discussed.

In addition, the event will feature an investment and collaboration program involving 10 domestic physical AI startups, developer community events, and one-on-one Q&A sessions with NVIDIA experts. A total of 20 ecosystem partners, including SamsungElectronics, SK hynix, LGELECTRONICS, Naver Cloud, SKTelecom, Coupang Intelligent Cloud, HyundaiEngineering&Construction, and Hancom, will participate in this event.

NVIDIA plans to release key presentation videos from the event via “NVIDIA On-Demand” following the event.