country_code

From Government to Gaming, AI Is ‘Strengthening Korea’s Digital Foundation,’ NVIDIA Leader Says at AI Day Seoul

Editor’s note: This post is part of a blog series highlighting NVIDIA AI Days across the globe.

Following packed events in Ho Chi Minh City and Tokyo last month, the tour of NVIDIA AI Days — bringing together AI enthusiasts, developers, researchers and startups in different pockets of the world — made its latest stop in Seoul, South Korea.

Attendees check in at NVIDIA AI Day Seoul.

Last week, more than 1,000 attendees joined NVIDIA AI Day Seoul to learn about sovereign AI — including breakout sessions on agentic and physical AI, hands-on workshops and a startup reception for members of the NVIDIA Inception program, all offering insights into the current and future landscape of the AI ecosystem in Korea. 

Andy Lee, Inception program manager at NVIDIA, gave opening remarks at the Inception and Venture Capital reception, highlighting this year’s Inception Grand Challenge in Korea. 

Five of 100 Korean startups that applied to the challenge advanced to the finals at AI Day Seoul, where the winning team received an award from the Korean Ministry of SMEs and Startups. All five finalists were recognized with certificates from NVIDIA for reflecting the momentum of Korea’s AI startup community.

Andy Lee, Inception program manager at NVIDIA, welcomes attendees to the Inception and VC reception at AI Day Seoul.

“We were thrilled to be named a finalist for the Inception Grand Challenge, providing an opportunity to share the successful collaboration forged with the NVIDIA technical teams,” said Jaeho Oh, CEO of NVIDIA Inception member PYLER. “The strong support from the Inception program enabled us to become the first in South Korea to adopt an NVIDIA DGX B200 system, as we continue to advance our globally leading, safe and reliable video understanding AI.”

Why It Matters

The latest AI trends in Korea showcase how the technology is driving advancements across robotics, industrial manufacturing and government

Major Korean conglomerates like Samsung, SK Group, Hyundai Motor Group and NAVER Cloud are engaged in deep AI integration.

“In the public sector, AI is already helping process documents, policies and regulatory information at scale, enabling citizens to receive the answers they need instantly,” said Shilpa Kolhatkar, global head of AI nations at NVIDIA. “Adoption is also accelerating across AI avatars, speech recognition technologies, gaming, music and the K-pop industry — innovations that together are strengthening Korea’s digital foundation.”

The numbers help showcase Korea’s expanding AI ecosystem:

Infographic that says there are 700+ NVIDIA Inception startups in Korea, 11 Korean sovereign LLM projects, 332,000 developers in Korea using NVIDIA technologies, 49,000 NVIDIA Deep Learning Institute enrollments in Korea and 13 supercomputers operating in Korea.

Event Highlights

In addition to four NVIDIA Deep Learning Institute training sessions, the event featured breakout sessions by topic. And at the NVIDIA Inception reception, Korean AI startups and venture capital firms came together to connect and collaborate.

An NVIDIA Deep Learning Institute workshop at AI Day Seoul.

At an “Ask the NVIDIA Tech Experts” area, specialists from across AI infrastructure, high-performance computing, large language models (LLMs), robotics and automotive connected with Korean developers for Q&As about real-world technological challenges and solutions.

The “Ask the NVIDIA Tech Experts” area at AI Day Seoul.

“The best part of AI Day Seoul was definitely coming to the Q&A counter to directly ask questions and communicate with NVIDIA technical experts,” said Useong Kim, principal engineer at Samsung Electronics. “It’s invaluable to ask what you’re curious about and get immediate feedback. I strongly felt the value of in-person communication beyond just absorbing online information, and I received a lot of practical help.”

Ankit Patel, senior director of developer marketing at NVIDIA, presented an all-conference session that explored agentic AI through the lens of scaling laws. He emphasized that reasoning models — the foundation of intelligent decision-making — are at the heart of the latest generation of AI.

Ankit Patel, senior director of developer marketing at NVIDIA, discusses NVIDIA open models, data and libraries at AI Day Seoul.

In addition to Patel’s talk, all-conference plenary sessions featured Marc Hamilton, vice president of global solutions architecture and engineering at NVIDIA, who discussed how AI factories are converging with physical AI.

And Kolhatkar shared how Korea is building sovereign AI models and modern AI factories tailored to national priorities through government initiatives and global partnerships.

Shilpa Kolhatkar, global head of AI nations at NVIDIA, presents at AI Day Seoul.

Industry in Motion

Korea-based NVIDIA partners are helping fuel the next industrial revolution driven by AI. 

“To build technological sovereignty and a sustainable AI foundation for the era of sovereign AI, NAVER Cloud is deepening its collaboration with NVIDIA — from GPU infrastructure to AI models and agent technologies,” said Yong-Jae Kwak, chief technology officer at NAVER Cloud, at AI Day Seoul.

At the event:

  • NAVER Cloud presented on its development of an agentic AI system, detailing the process of deriving optimization insights using the NVIDIA NeMo Agent Toolkit.
  • LG AI Research showcased advanced LLM training using the NVIDIA NeMo framework, tapping into FP8 training recipes to achieve over 20% faster training and applying reasoning-aligned techniques to enhance reasoning performance to reach new levels of training efficiency. 
  • Coupang Intelligent Cloud discussed how it is building Coupang’s AI factory — built with NVIDIA DGX systems based on the NVIDIA Hopper and Blackwell architectures — which will deliver AI-powered solutions for demand forecasting, logistics route optimization, ad personalization and more.
Kweonwoo Jung, senior machine learning engineer at NAVER Cloud, presents on AI agents within the NAVER ecosystem.

What’s Next

Announced at the APEC Summit last month, Korean developers across enterprises, startups and research institutes will soon be equipped to harness the power of 260,000 GPUs with NVIDIA’s full stack of accelerated libraries and software development kits.

“NVIDIA AI Day was a meaningful opportunity for Korean companies and startups to see cutting-edge AI technologies up close and explore new possibilities for collaboration,” said Kyungwon Cho, director of the Korean Ministry of SMEs and Startups. “The 260,000-GPU AI infrastructure announced at the recent APEC Summit will play a major role in advancing the next phase of Korea’s AI ecosystem. We expect this to help startups move faster in developing their technologies and spark new waves of innovation across industries. The Ministry of SMEs and Startups will continue working with NVIDIA to create even more opportunities for Korean startups to grow and compete globally.”

Next up: Join NVIDIA AI Day Sao Paulo, running in January, to discover the latest AI breakthroughs, connect with peers and experts, and more.

Explore more content from AI Days around the globe.

Related News