Nvidia Secures Memory Supply from SK Hynix Amid $500 Billion AI Deal
Nvidia has secured a significant supply of high-bandwidth memory from SK Hynix, a move aimed at alleviating a global memory shortage and ensuring components for its AI systems.
Chipmaker Nvidia has secured a substantial supply of AI memory from South Korea's SK Hynix, a deal expected to run for several years and potentially worth hundreds of billions of dollars. The agreement addresses concerns about component availability for Nvidia's advanced processors and systems amid a global memory shortage driven by the AI boom.
As part of the arrangement, SK Hynix affiliate SK Telecom will establish a cloud business utilizing Nvidia's Vera Rubin systems. Nvidia plans to construct data centers requiring two gigawatts of power, indicating a large-scale buildout involving hundreds of thousands of GPUs.
High-bandwidth memory (HBM) is a critical component for Nvidia's graphics processing units (GPUs) and AI infrastructure. SK Hynix is identified as a leading producer of HBM, and this partnership aims to ensure a stable supply for Nvidia's demanding production needs.
The deal was announced at an AI summit in San Francisco, with the presence of South Korean officials. Additionally, Nvidia announced a $1 billion investment in Naver, a Korean cloud company, for its GPU-centric data center developments.