Key Takeaways
- SK Telecom and Nvidia unveiled plans for a gigawatt-scale AI cloud facility in South Korea, with initial operations targeted for 2027.
- A multi-year strategic partnership between Nvidia and SK Hynix focuses on advanced memory chip development to tackle worldwide AI infrastructure constraints.
- Korean tech giant Naver will deploy Nvidia’s DSX platform to build sovereign AI infrastructure, beginning at 55 megawatts with gigawatt-level expansion planned.
- Doosan Group partnered with Nvidia on robotics-focused AI initiatives.
- These agreements materialized following CEO Jensen Huang’s Seoul weekend visit, which included meetings with industry executives and a television appearance.
When Nvidia’s CEO Jensen Huang arrived in Seoul last Friday with promises of “surprises,” he wasn’t exaggerating. During a whirlwind weekend that featured Korean barbecue dinners and media appearances, Huang finalized multiple AI infrastructure agreements covering cloud computing, semiconductor technology, and robotic systems.
The most significant announcement involves a collaboration between Nvidia and SK Telecom. These companies are developing a gigawatt-scale AI cloud infrastructure in Korea, with the inaugural AI factory scheduled to launch in 2027. The initiative aims to deliver “sovereign, physical and agentic AI services” to Korean businesses, with expansion plans throughout the Asian region. Financial terms of the agreement remain undisclosed.
Advanced Memory Technology Takes Center Stage
SK hynix, another subsidiary within the SK Group conglomerate, simultaneously revealed a distinct “multi-year technology partnership” with Nvidia. This collaboration emphasizes advanced memory solutions — critical components for operating large-scale AI systems that are experiencing significant global supply constraints.
The agreement encompasses joint development of memory technologies for Nvidia’s hardware platforms spanning AI infrastructure, consumer AI applications, and physical AI sectors. SK Group’s chairman, Chey Tae-won, recently committed to doubling silicon wafer manufacturing capacity, though he cautioned that semiconductor shortages may continue through 2030.
Huang underscored the supply challenge quite visibly. During last week’s Computex conference in Taipei, he autographed a memory chip exhibit at the SK hynix pavilion with a direct message: “Please make more.”
Nvidia’s market capitalization has surpassed $5 trillion — exceeding the gross domestic product of nations like Japan or India — as governments and corporations invest hundreds of billions in AI infrastructure development.
Naver and Doosan Partnerships Round Out Seoul Agreements
In addition to the SK collaborations, Nvidia revealed an AI infrastructure partnership with Naver, Korea’s dominant internet company. Naver will leverage Nvidia’s DSX platform for developing and scaling comprehensive AI platforms, beginning with 55 megawatts and targeting gigawatt-scale expansion.
“Useful AI has arrived, and demand for AI factories is extraordinary,” Huang stated in a press release accompanying the Naver announcement.
Nvidia additionally disclosed a robotics-centered partnership with Doosan Group, completing what proved to be an eventful Monday in the South Korean capital.
Prior to his South Korea visit, Nvidia introduced a high-performance laptop processor for Windows systems targeting the consumer AI PC segment.
Naver’s infrastructure buildout utilizing the Nvidia DSX platform represents one of the most substantive commitments emerging from the Seoul trip, featuring a specified initial capacity and well-defined expansion trajectory already established.


