Jensen Huang Visits South Korea, Meets SK, LG and Other Chaebols — Robotics and Physical AI Emerge as Core Topics
N.R. Finch
Jensen Huang landed in Seoul this week for back-to-back meetings with the heads of SK, LG, Hyundai Motor, and Naver, pushing the partnership agenda beyond HBM supply into robotics and autonomous driving — signaling Nvidia is repositioning Korea as a physical-AI proving ground, not just a chip supplier.
Two visits in seven months — what is Huang after?
Huang arrived at Seoul's Gimpo Airport on Friday by private jet and dined that evening with SK Chairman Chey Tae-won, Hyundai Motor Executive Chair Euisun Chung, LG Chairman Koo Kwang-mo, and Naver founder Lee Hae-jin.
This is his second Korea visit in seven months; with Chey alone it is the fourth meeting in half a year — a frequency virtually unmatched in the industry.
This reflects a strategic upgrade: Korea is no longer just where Nvidia buys HBM — it is being cast as a co-development partner for next-generation AI hardware.
Physical AI — what does the core agenda actually mean?
Physical AI — embedding AI directly into robots, vehicles, and factory equipment so they act autonomously in the real world — was the top topic of the trip.
SK, LG, and Hyundai are all pushing factory automation and robotics, making Nvidia a shared critical-technology supplier across all three conglomerates.
KB Securities analyst Jeff Kim called Korea "the most ideal proving ground" for physical AI. This means → Korea's dense manufacturing base plus chaebol execution speed can move Nvidia's AI frameworks from lab to production line fastest.
The Doosan case — how does a robot go from concept to product?
Doosan Robotics is co-building an agentic operating system for next-gen robots with Nvidia, targeting commercial smart-robot solutions in 2027 and industrial humanoid robots in 2028.
Doosan Electronics BG has simultaneously entered Nvidia's next-gen AI-accelerator supply chain, providing CCL (copper-clad laminate — the base material for printed circuit boards).
In plain terms = Doosan buys Nvidia's "brain" for its robots while supplying "body parts" for Nvidia's chips — a two-way lock-in, no longer a simple buyer-seller relationship.
HBM and memory — what is new on the old agenda?
Chey said at Computex he plans to double memory capacity over five years to address a shortage lasting through 2030.
Nvidia's HBM (high-bandwidth memory — ultra-fast memory designed specifically for AI chips) supply partnerships with Samsung and SK hynix are expected to deepen further.
Nvidia committed last October to supply Korea with over 260,000 advanced AI chips. This means → the chips-for-memory reciprocity is hardening into a structural lock-in on both sides.
Gaming and startups — what else is on the agenda?
Huang is expected to meet NCSoft CEO Kim Taek-jin and Krafton Executive Director Chang Byung-gyu to discuss AI-gaming collaboration. Krafton this year launched robot company Ludo Robotics — the line between gaming and robotics is blurring.
He will also visit Seoul National University's AI research institute and hold his first-ever closed-door session with Korean robotics startup founders.
One industry insider flagged the key question: most of these companies are still buyers of Nvidia solutions — whether the relationship can move beyond procurement into genuine AI investment partnerships is the deeper test of this trip.
Content is for reference only, not financial advice.