2 weeks Ago
Nvidia’s CEO Jensen Huang had asked SK Hynix to bring forward the supply of HBM4 memory chips by six months, according to SK Group Chairman Chey Tae-won. Last month, memory chip maker SK Hynix said it aimed to supply its next-generation high-bandwidth memory chips to customers in the latter half of 2025. However, this timeline was faster than the initial target, a SK Hynix spokesperson said on Monday (November 4), according to a Reuters report.
Nvidia’s CEO Jensen Huang’s request for faster delivery shows the high demand for new iterations of Nvidia’s GPUs, which are crucial for AI and will have next-generation HBM chips. Tech giant Nvidia, the market leader in AI chips, is set to replace struggling chipmaker Intel on the Dow Jones Industrial Average, ending Intel’s 25-year presence on the blue-chip index. SK Hynix is one of the major companies racing to meet the massive demand for HBM chips.
SK Hynix started mass producing 12-layer HBM3E in September, with plans to supply the latest products in the ongoing quarter. HBM4 is expected to offer system designers very high bandwidth capabilities and optimal power efficiency, according to an article in Semiconductor Engineering . For AI training, HBM4 builds on a strong track record of HBM3 and HBM3E, which were implemented in state-of-the-art AI accelerators.
.
