“Whatever happens in the stock market, you should actually be happy,” Huang said, “because you can buy the stock more cheaply” . He called the market volatility a “chance to buy at a discount” and argued that the AI infrastructure build-out is still in its early stages
.
The sell-off itself was partly a story of profit-taking. Korean stocks that had rallied hard on the mere rumor of a Huang visit—companies like LG, Doosan, and Naver—reversed sharply as traders locked in gains . The KOSPI dropped roughly 5% on June 5 alone
. At the same time, broader AI-bubble fears were circulating in both the US and Asian markets, with some investors openly questioning whether AI spending could sustain its pace
. Huang’s reply treated those fears as short-sighted. “It is a foregone conclusion that the future of AI is very bright,” he said
.
Huang’s trip was not only about calming markets; it was about expanding Nvidia’s physical footprint in South Korea. The visit produced a series of concrete deals that extended Nvidia deeper into memory, AI data centers, robotics, and manufacturing infrastructure.
SK hynix – HBM technology partnership
A multi-year, multi-platform technology partnership will see SK hynix develop advanced memory for Nvidia’s AI infrastructure, personal AI, and physical AI platforms . Huang underscored the company’s importance by stating, “AI cannot advance without SK hynix,” and noted that Nvidia already purchases memory worth several billion dollars
.
SK Telecom – gigawatt-scale AI cloud
SK Telecom and Nvidia agreed to build a gigawatt-scale AI cloud in South Korea using Nvidia’s DGX platform. The first AI data center is expected to come online in 2027, with a goal of expanding the GPU-as-a-Service model into other parts of Asia .
LG Group – AI factory for robotics and autonomous driving
Nvidia and LG Group announced plans for a joint AI factory covering robotics, autonomous driving, data center technologies, and GPU cloud services. Huang met LG chairman Koo Kwang-mo while in Korea to accelerate the collaboration .
Naver, Doosan, Hyundai, and robotics startups
The trip also produced AI infrastructure cooperation agreements with internet giant Naver , AI and robotics collaboration with Doosan Group
, and physical AI and robotics platform discussions with Hyundai Motor, NC, and several domestic startups
.
Huang additionally hinted at a “surprise gift for Korea” and said Nvidia was willing to hold its flagship GTC conference in the country if Korea wanted it . He also revealed that Nvidia had begun hiring for an R&D center in Korea and would build a dedicated site once enough personnel were in place
.
Across his press appearances, Huang repeated a handful of themes that shaped the investment narrative around his visit.
First, that AI is now core global infrastructure, not a cyclical bet. “It has become an established fact that AI will serve as core global infrastructure,” he said . That thesis underpinned his dismissal of near-term stock volatility as noise.
Second, that the build-out has only just begun. Huang pointed to the Nvidia Vera Rubin platform, with mass production expected in the second half of 2026, as evidence that years of demand remain ahead . Nvidia’s growing focus on AI data centers and digital twin-based AI factories further supports the infrastructure thesis
.
Third, that robotics is South Korea’s next major sector. Huang framed Korea’s manufacturing base as a natural fit for physical AI applications. “Because Korea is a manufacturing centre of the world, we can apply the robotics technology, the physical AI technology that we invent here for the industry,” he told reporters upon landing in Seoul on June 5 . The manufacturing of semiconductors itself, he argued, would become increasingly robotics-driven
.
By the time Huang left Seoul, the Korean market’s early-week sell-off looked increasingly detached from the substance of the trip. The visit shifted Nvidia’s Korea relationship from a chip-supply and HBM-review dynamic into a full-stack AI infrastructure partnership that now touches memory, cloud, factories, robots, automobiles, and talent .
Comments
0 comments