SEOUL, June 09 (AJP) - Nvidia CEO Jensen Huang concluded his official four-day visit to South Korea Monday evening, securing a comprehensive, long-term memory and foundry supply roadmap with Samsung Electronics to power the next generation of artificial intelligence hardware.
Huang wrapped up his multi-industry tour with a private meeting with Jun Young-hyun, Vice Chairman and Head of Samsung Electronics’ Device Solutions (DS) Division, at the Shilla Hotel in Seoul. The high-stakes meeting finalized a localized supply chain for high-bandwidth memory (HBM), which acts as the critical bottleneck for scaling advanced AI accelerators.
Following the meeting, Vice Chairman Jun expressed strong confidence in the partnership’s trajectory.
"I had a lot of good conversations with Jensen, and honestly, although we have been cooperating for a long time, I think we had the best conversation so far today," Jun told reporters. "We comfortably had many good conversations, and we discussed how we will collaborate in the short term on HBM4 and foundry cooperation."
The two leaders mapped out both immediate and future supply lines. According to Jun, Samsung is prioritizing the delivery of its 6th-generation HBM4 and Server Low-Power Memory Modules (SOCAMM) to meet current demand.
"In the short term, starting this year, we must sufficiently supply HBM4 and SOCAMM," Jun said. "Then, from next year, we talked a lot about long-term cooperation, such as HBM4E, the foundry business, and HBM5."
Samsung has already shipped samples of its 7th-generation HBM4E to Nvidia. Beyond memory, Jun confirmed that Samsung and Nvidia are actively expanding their foundry (contract manufacturing) partnerships.
"Currently, we are collaborating on autonomous driving chips required for 4-nanometer and 8-nanometer processes, and Nvidia’s accelerator chip called the Groq chip," Jun said, adding that "discussions on next-generation cooperation are also underway."
The Samsung meeting served as the capstone to a broader Nvidia initiative to leverage South Korea’s "physical AI" ecosystem. During his doorstep press conference earlier that evening, Huang emphasized the necessity of massive infrastructure and robust supply chains to transition from software-based AI to physical robotics.
"The next wave of AI is physical AI where AI can interact with the physical world, Robotics," Huang said, praising the country's unique manufacturing and electronics capabilities. "Korea is in a very unique position in the world of physical AI," Huang said, praising the country's unique manufacturing and electronics capabilities. "Manufacturing, heavy industries, electronics, and software in AI must unite, must fuse together into Robotics."
Huang also addressed market concerns about the AI boom, firmly stating that the demand for AI infrastructure is only beginning.
"Intelligence is a commodity, which mainly means intelligence plus character is not a commodity," Huang told reporters, crediting the Korean national character—the "ability to suffer" and "resilience"—as a primary driver for his investment in the country. "If you combine that with AI, incredible, incredible things. It is the reason why I'm here today."
When asked about Huang’s earlier comment designating SK Hynix as Nvidia’s "largest memory partner," Jun offered a measured, results-oriented response regarding Samsung's competitive position.
"We will work hard on our tasks," Jun said. "We will show it through results later."
Following the private meeting and press conference, Jun joined other top Samsung semiconductor executives, including Foundry Business President Han Jin-man, at Nvidia's 'Korea AI Ecosystem' reception. The reception, attended by leaders from SK Hynix and LG Electronics, marked the official end of Huang’s visit before his departure.
Copyright ⓒ Aju Press All rights reserved.