Nvidia Chief Says He Will Announce $500 Billion SK Group Partnership

Before the summit, Reuters reported that Samsung Electronics and SK Hynix would announce large, long-term memory-chip supply contracts with global technology partners. Kim Yong-beom, Lee's policy adviser, told reporters that global technology companies would also announce strategic investments in AI data centers, without naming the companies or supplying financial details.

Those discussions are material to AI infrastructure because high-bandwidth memory is a central input for training and serving large GPU clusters. SK Group includes SK Hynix, one of the major suppliers of HBM used alongside AI accelerators. However, neither the Korea Times report nor the Reuters report establishes whether the reported $500 billion Nvidia-SK Group partnership is a memory-supply agreement, a data-center investment program, a memorandum of understanding, or a broader collection of commercial commitments.

Nvidia also announced a joint AI research lab with the Korea Advanced Institute of Science and Technology at KAIST's Kim Jaechul Graduate School of AI in Seoul. According to Nvidia, the lab will focus on agentic AI and combine Nvidia's full-stack AI technologies, open models, and AI Cloud partner computing with KAIST research.

For ML infrastructure teams, the central item to watch is the eventual disclosure of contract scope: accelerator allocation, HBM generations and volumes, data-center capacity, delivery dates, and whether announced amounts represent firm purchases or nonbinding investment targets. Industry reporting on comparable AI infrastructure agreements often combines chip supply, cloud capacity, and long-dated capital commitments under a single headline figure, making those distinctions consequential for assessing near-term hardware availability.