Nvidia asks SK hynix to pull forward chip deliveries

Nvidia CEO Jensen Huang has urged South Korea’s SK hynix Inc. to expedite the delivery of its upcoming HBM4 (High Bandwidth Memory) chips by six months, according to SK Group’s chairman on Monday.

Nvidia asks SK hynix to pull forward chip deliveries

This request reflects Nvidia's critical need for advanced memory chips to support AI workloads amid a global shortage of cutting-edge semiconductor technology. The Icheon-based SK hynix, the world’s second-largest memory chip maker, is a key supplier of HBM chips, which are essential for handling the vast data demands required for training AI models. Nvidia, the dominant player in the AI hardware market, heavily relies on these chips and has collaborated with SK hynix and Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC) to meet surging demand.

Rapid pace of HBM technology development

Speaking at an AI summit in Seoul via video link, Huang commended the rapid pace of HBM technology development but stressed that AI applications demand even higher-performance memory. “The current pace of HBM memory technology development and product launches is impressive, but AI still requires higher-performance memory,” he stated. SK hynix announced last month that it expects to deliver 12-layer HBM4 chips by the second half of 2025. Additionally, on Monday, the company revealed it would ship samples of the world’s first 16-layer HBM3E chips by early 2025, further strengthening its foothold in the AI chip market.