SEOUL (Reuters) – South Korea’s Samsung Electronics said on Friday it is in “advanced negotiations” to supply its next-generation high-bandwidth memory (HBM) chips, or HBM4, to North American Nvidia.
Samsung, which plans to commercialize the new chip next year, has not specified when it plans to launch the latest version of its HBM chip, a key component of artificial intelligence chipsets.
SK Hynix, Nvidia’s main HBM chip supplier, announced on Wednesday that it intends to start shipping its latest HBM4 chips in the fourth quarter and expand sales next year.
FREE TOOL
XP simulator
Find out in 1 minute how much your money can yield
In a statement announcing the cooperation with Samsung, Nvidia said it is a “fundamental collaboration in the supply of HBM3E and HBM4”, without giving further details.
In a separate deal, Samsung said it will buy 50,000 next-generation chips from Nvidia to build an artificial intelligence-powered semiconductor factory, aiming to improve the speed and yield of chip production.
Samsung has been slower to capitalize on the boom in AI-driven memory chips, which led to weaker financial performance and a restructuring of its chip division last year. Its profits rebounded last quarter on demand for conventional memory chips.
Continues after advertising
The launch of the HBM4 chips will be an important test of Samsung’s ability to regain its market advantage, according to analysts.
HBM – a type of dynamic random access memory (DRAM) standard first produced in 2013 – involves vertical stacking of chips to save space and reduce power consumption, helping to process the large volumes of data generated by complex AI applications.
Investors are watching to see whether Samsung’s HBM4 technology can chip away at SK Hynix’s lead in the advanced memory chip market.
(Reporterem by Heekyong Yang in Hunjoo Jin)
