Nvidia and South Korea’s SK Group signed letters of intent on July 24 for a partnership worth more than $500 billion, spanning new AI data centers and a multi-year memory-chip supply agreement, the companies said in coordinated announcements.
A 2-Gigawatt AI Factory
SK Telecom, SK Group’s telecom subsidiary, will build a 2-gigawatt AI factory in South Korea running on Nvidia’s DSX full-stack platform and its next-generation Vera Rubin accelerators, with the first phase targeted to come online in 2027. Nvidia CEO Jensen Huang said South Korea has “all the ingredients to become a global AI powerhouse — world-class networks and data centers, leadership in chip technology and vast industrial scale.”
SK Hynix Locks In as Memory Supplier
The deal’s second pillar addresses a bottleneck that has increasingly constrained AI buildouts worldwide: memory supply. SK Hynix, an SK Group subsidiary that recently raised $26.5 billion in a record Nasdaq debut, will supply HBM4 chips for Nvidia’s Vera Rubin accelerators and co-develop future generations of the technology with Nvidia. High Bandwidth Memory stacks DRAM directly beside AI processors so they can be fed data fast enough to keep up with model training and inference — and it has been in tight supply as Nvidia and rivals race to build out AI data centers.
SK Group Chairman Chey Tae-won said the company would “leverage SK Hynix’s AI memory and SK Telecom’s AI infrastructure capabilities” to build what he called a world-class AI factory in partnership with Nvidia.
Part of a Wider Race
The agreement follows Samsung’s own $648 billion commitment to AI chip and infrastructure investment and Nvidia’s similar national AI factory deal with Japan, underscoring how South Korea and its neighbors are competing to lock in Nvidia’s newest chip generations and the memory that powers them. The companies said the partnership remains subject to definitive agreements following the signed letters of intent.