Choi Tae-won, Chairman of SK Group, and Jensen Huang, CEO of NVIDIA, reaffirmed their global artificial intelligence (AI) memory collaboration, solidifying their strategic alliance toward building an 'AI factory.' Their goal is to combine NVIDIA's AI chip design technology with SK Group's memory infrastructure to capture the next-generation AI data center market.
According to industry sources on June 3, Choi and Huang met the previous day at Computex 2026 in Taiwan, where they toured the SK Hynix exhibition booth and examined key AI memory technologies and products.
After the tour, Huang left a witty yet urgent message on a seventh-generation high-bandwidth memory (HBM4E) wafer, stating, "Please make more." This underscores the necessity of SK Hynix's memory infrastructure for the successful implementation of NVIDIA's next-generation AI accelerator designs.
Huang also drew attention by writing "I love LPCAMM" on SK Hynix's 192GB low-power memory product. Speaking with South Korean reporters, he remarked, "Korea is the most important part of our AI semiconductor ecosystem. We have much to do together beyond just supplying chips and DRAM, including in science, robotics, and the AI factory sector." This indicates NVIDIA's intention to collaborate with Korean companies to create a next-generation AI hub that integrates AI infrastructure and software.
However, Huang expressed concerns about the memory supply situation, stating, "We have secured supply chains in all areas, including HBM4, post-packaging, and silicon photonics, but we are still facing shortages."
In response, Choi immediately pledged significant infrastructure investments and outlined his vision. He diagnosed that the expansion of AI infrastructure would lead to memory bottlenecks and shortages continuing until 2030.
Choi declared, "We will double SK Hynix's total wafer production capacity within the next five years." Considering that building a new semiconductor fab takes at least three to five years, this proactive investment aims to fully absorb NVIDIA's explosive demand.
Choi also actively embraced Huang's 'AI factory' agenda, indicating a transformation within SK Group. He emphasized, "Currently, we are merely a parts supplier producing memory chips for AI, but in the future, we want to directly challenge the production of AI factories that refine and generate intelligence." An AI factory refers to a next-generation data center that processes raw data to deliver advanced intelligence services.
This reflects SK Hynix's ambition to evolve from a simple hardware supplier to a comprehensive AI infrastructure partner supporting NVIDIA's AI designs.
However, Choi clearly stated that overcoming structural obstacles related to funding, power, and equipment supply over the next decade will be a challenge. To address this, SK Hynix plans to strengthen its collaboration within the AI memory ecosystem, including local supply chains such as TSMC and Foxconn.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.