Nvidia CEO Jensen Huang (right) shares fried chicken and beer with Samsung Electronics Chairman Lee Jae-yong (left) and Hyundai Motor Group Executive Chair Chung Euisun at a Kkanbu Chicken outlet in Gangnam-gu, Seoul, in October, at a gathering that Korea nicknamed the "kkanbu summit." (Yonhap) Nvidia CEO Jensen Huang is set to land in Seoul next Friday for a round of meetings with the heads of Korea's largest conglomerates, his first return since the October "kkanbu summit" that captivated the country.This time, the agenda has moved up the value chain: from the memory chips that feed Nvidia's processors to the machines those processors will power.According to industry sources on Friday, SK Group Chairman Chey Tae-won, LG Group Chairman Koo Kwang-mo and Naver founder Lee Hae-jin have confirmed the June 5 date. Hyundai Motor Group Executive Chair Chung Euisun is leaning toward attending and finalizing his schedule.Samsung Electronics Chairman Lee Jae-yong, a central figure at the original gathering, is reportedly unable to join because of overseas commitments.Huang will fly to Korea straight after the main agenda at GTC Taipei 2026, Nvidia's annual AI conference running June 1-4, where he will deliver the opening keynote speech. His last trip to Korea ended over fried chicken and beer at a Samsung-dong outlet of Kkanbu Chicken, where he, Lee Jae-yong and Chung gathered without aides. As the franchise is named after the Korean word for ally, popularized by the drama "Squid Game," the unscripted outing became shorthand for the alliance binding Nvidia chips to Korean manufacturing.The focus this time is physical AI, the field that fuses artificial intelligence with robots, vehicles and factory hardware. The Koo-Huang meeting will be their first, building on talks last month between LG Electronics CEO Lyu Jae-cheol and Madison Huang, the CEO's daughter and an Nvidia senior director, on pairing LG's CLOiD home robot with Nvidia's Isaac robotics software. The agenda also covers a robot foundation model, the "brain" for humanoids that LG AI Research is developing.Hyundai is positioning Boston Dynamics' Atlas humanoid as a core future business, while Naver, which signed a physical AI platform pact with Nvidia at last year's Gyeongju APEC summit, aims to advance joint work tied to Nvidia's Omniverse simulation platform. Naver was allotted 60,000 of the 260,000 GPUs Nvidia pledged to Korea at that summit.Chey, who meets Huang frequently and is due to see him again at Computex 2026 in Taiwan, would reconnect in Seoul on memory. The timing sharpens the point as Samsung said Friday it had begun the world's first sample shipments of seventh-generation HBM4E memory, three months after starting HBM4 mass production.An Nvidia source said the venue is still being chosen.