Nvidia Strikes 3 South Korea AI Data Center Deals as SK Telecom Targets 2027 Launch
Updated
Updated · Reuters · Jun 7
Nvidia Strikes 3 South Korea AI Data Center Deals as SK Telecom Targets 2027 Launch
3 articles · Updated · Reuters · Jun 7
Summary
Nvidia unveiled agreements with SK Hynix, Naver and Doosan in South Korea to build AI data centers around its technology, expanding its push to sustain the AI boom.
SK Hynix paired that announcement with a multi-year technology partnership on next-generation memory for global AI data centers, aiming to secure supply despite long advanced-chip development cycles.
SK Telecom said it will build a gigawatt-scale AI cloud in South Korea using Nvidia technology, with its first AI data center scheduled to come online in 2027.
Naver and Doosan also plan to use Nvidia technology for AI data centers, while Doosan expects its energy solutions and physical AI tools to feed into Nvidia's data-center platforms.
The deals deepen Nvidia's ties in a key manufacturing hub where SK Hynix and Samsung dominate memory chips, a critical component for AI infrastructure.
Will the Nvidia-SK Hynix deal create an AI monopoly, locking out rivals from critical next-generation memory chips?
With a global chip shortage looming until 2028, can SK Hynix's new fabs truly satisfy the insatiable AI demand?
As AI drives memory prices up 480%, are high-end electronics about to become unaffordable for the average person?
The HBM4E Revolution: Nvidia, Samsung, and SK Hynix Drive the Next Generation of AI with Vera Rubin
Overview
Nvidia's next-generation AI platform, Vera Rubin, is rapidly advancing, with recent updates showing significant progress in its development. This platform is designed to drive the next wave of artificial intelligence by demanding unprecedented performance and data throughput. A key foundation of Vera Rubin, especially its high-end Ultra variant, is the advanced HBM4E memory, which is essential for handling the intensive data processing needs of future AI models. Samsung has already started shipping early samples of HBM4E, ensuring that the necessary memory components are moving through the development pipeline to support Vera Rubin’s ambitious capabilities.