NVIDIA to host AI Day in Seoul to showcase agentic and physical AI

A focus on agentic and physical AI

NVIDIA has announced it will host NVIDIA AI Day Seoul 2026 at the COEX convention center. The two-day event, scheduled for November 9 and 10, aims to bring together a diverse group of 2,000 participants, including developers, startups, and researchers. As reported by 파이낸셜뉴스, the primary goal is to demonstrate how Korean companies are applying advanced technologies to real-world problems.

The program is structured to highlight three specific pillars: agentic artificial intelligence, physical AI, and AI infrastructure. Agentic AI refers to systems capable of taking actions to achieve specific goals, while physical AI focuses on the intersection of software and hardware, such as robotics. These technologies are moving beyond simple text generation into active participation in industrial workflows.

During the opening plenary session on November 9, NVIDIA leaders will discuss the broader strategies for these technologies. They plan to address the specific requirements for building AI factories and deploying edge AI. The event serves as a showcase for how these tools can be integrated into the existing economic and industrial fabric of South Korea.

Industrial applications from Korean leaders

Several of Korea's largest technology and manufacturing firms will present their specific use cases during the event. Samsung Electronics and SK hynix are set to demonstrate how they use NVIDIA PhysicsNeMo to accelerate simulations for semiconductors. This software helps them analyze complex physics and device processes more efficiently during the manufacturing stage.

In the realm of robotics, LG Electronics will showcase a system for training and validating robots using NVIDIA Isaac GR00T. They plan to provide specific examples of how this technology is being applied within factory environments. Additionally, Doosan Robotics will present methods for simulating physical phenomena like deformation and contact, which are critical for realistic robot movement.

Other participants will focus on the data and infrastructure side of the AI boom. Superb AI will discuss the use of synthetic data in robot development, while NAVER Cloud will share its experience with open models and large-scale training. These collaborations show how the hardware provided by NVIDIA is being customized for diverse industrial needs across the country.

Infrastructure and AI factories

A significant portion of the event is dedicated to the massive infrastructure required to run modern AI. SK Telecom (SKT) will address the logistical challenges of building AI factories, which involve everything from power grids to semiconductor requirements. This highlights the physical reality of AI, where energy and hardware availability are major bottlenecks.

NAVER Cloud will provide a technical deep dive into operating clusters made up of thousands of graphics processing units (GPUs). They will specifically discuss how they manage and respond to outages in these massive systems. This information is vital for enterprises looking to maintain high uptime for their own AI services.

Coupang Intelligent Cloud will also join the infrastructure track to introduce their specific architecture. They will showcase how they use the NVIDIA GB300 NVL72 for both training and inference. By focusing on these large-scale systems, the event emphasizes that the next phase of AI growth depends on reliable, high-capacity computing environments.

Hands-on training and research

The event is designed to be interactive, offering more than just presentations for the 2,000 expected attendees. On the first day, a program called 'Build-a-Claw' will take place in the lobby. This allows participants to build their own AI agents using specific tools like OpenClaw and NVIDIA NemoClaw, providing a direct way to test the technology.

The second day will shift toward deeper technical education through the NVIDIA Deep Learning Institute (DLI). These paid workshops will teach developers how to optimize large-scale AI inference and robotics simulation. They will also cover the development of safe agents based on large language models (LLMs), which is a key safety concern for enterprise adoption.

Beyond the workshops, the event will feature an investment-matching session specifically for physical AI startups. This is intended to connect early-stage companies with the capital and expertise needed to scale. Korean researchers will also present new AI science research, ensuring that the latest academic breakthroughs are shared with the commercial sector.

Accessibility and next steps

NVIDIA is making several provisions to ensure that the knowledge shared at the event reaches a wider audience. While registration for the physical event is closed, the major sessions will be available to watch on demand through the NVIDIA On-Demand platform. This allows those who could not attend in person to still benefit from the insights shared by Korean industry leaders.

For those attending in person, the first-floor lobby will serve as a consultation hub. Attendees can speak directly with NVIDIA experts regarding AI infrastructure, LLMs, and robotics. This direct access provides a rare opportunity for developers to get specific advice on their own projects and technical hurdles.

The event also includes a practical certification component for those looking to formalize their skills. Attendees will have the opportunity to take an NVIDIA certification exam at the associate level for free. By combining high-level strategy, industrial use cases, and hands-on certification, the event aims to solidify Korea's position in the global AI landscape.