Nvidia and South Korea’s SK Group just announced a partnership worth over $500 billion to build AI data centers and next-generation memory infrastructure.
The initiative, announced on July 24, represents one of the largest single commitments to AI infrastructure ever made.
What the deal actually includes
At its core, SK hynix, the memory chip giant within SK Group, will co-develop high-bandwidth memory (HBM) designed specifically for Nvidia’s AI platforms.
SK Telecom, another arm of the conglomerate, plans to construct a 2-gigawatt AI data center powered by Nvidia’s Vera Rubin chips and SK hynix’s next-generation HBM4 memory. The first facility is expected to come online in 2027.
The deal builds on previous agreements struck in June 2026 between Nvidia, SK Telecom, and SK hynix. Those earlier pacts laid the groundwork for what’s now being described as a multi-gigawatt AI factory expansion across South Korea.
There’s also a separate but related effort involving Naver, South Korea’s dominant search and tech platform, alongside Brookfield, the Canadian infrastructure giant. Nvidia is working with both to expand an existing AI data center in the country.
South Korea’s strategic positioning
South Korea is already home to Samsung and SK hynix, two of the world’s three major memory chip manufacturers. Adding multi-gigawatt AI data centers to that ecosystem creates a vertically integrated AI supply chain that few nations can match.
For Nvidia, the partnership extends its strategy of locking in long-term customers and supply chain partners across Asia. The company has struck similar deals in Japan, the UAE, and Saudi Arabia in recent months, building a global network of AI factories that depend on its chips and software ecosystem.