Written by Hyunju Jin and Joyce Lee
SEOUL (Reuters) – Nvidia CEO Jensen Huang has asked memory chip maker SK Hynix to bring forward supply of its next-generation high-bandwidth memory chip, called HBM4, by six months. SK Group Chairman Choi Tae-won said on Monday.
South Korea’s SK Hynix announced in October that it aims to supply chips to customers in the second half of 2025. An SK Hynix spokesperson said Monday that the schedule is faster than the original goal, but did not provide further details.
Huang’s call for faster delivery highlights the demand for larger, more energy-efficient chips used in Nvidia’s graphics processing units to develop AI technology. Nvidia controls more than 80% of the global AI chip market.
SK Hynix has led the global race to help process vast amounts of data to train AI technology and meet explosive demand for HBM chips, critical to Nvidia’s chipsets . However, it faces increasing competition from rivals such as Samsung Electronics and Micron.
SK Hynix plans to supply the latest 12-layer HBM3E to an unidentified customer this year, but also plans to ship samples of the more advanced 16-layer HBM3E early next year, CEO Kwak No-jeong told SK. Said at AI SUMMIT. Seoul in 2024.
Samsung announced last week that supply deals with unidentified key customers are progressing after being plagued by delays, and that it is in talks with key customers to produce “improved” HBM3E products in the first half of next year. He added that there is. Samsung also plans to produce next-generation HBM4 products in the second half of next year.
SK Hynix stock rose 5.1%, and Samsung stock rose 1%. The overall market rose 1.6%. (This story has been corrected to say “16-layer HBM3E” instead of “18-layer HBM3E” in paragraph 5)
(Reporting by Hyun-joo Jin and Joyce Lee; Editing by Ed Davis and Himani Sarkar)
