Nvidia CEO Jensen Huang has urged SK Hynix to accelerate the production of its next-generation high-bandwidth memory chips, HBM4. Initially, SK Hynix planned to begin supplying these advanced chips in late 2025, but it is now working towards an expedited timeline.
This request highlights the growing demand for high-capacity, energy-efficient memory chips essential for powering Nvidia's GPUs, which are critical for artificial intelligence advancements. Nvidia currently holds over 80% of the global AI chip market, increasing the urgency for faster HBM chip delivery.
SK Hynix is a significant player in the competition to meet the rising demand for HBM, necessary for processing large datasets for AI training. The company is also facing competition from Samsung Electronics and Micron.
In addition, SK Hynix plans to deliver its latest 12-layer HBM3E chips to an undisclosed customer this year and expects to showcase samples of a more advanced 16-layer HBM3E by early next year. Meanwhile, Samsung is making strides in securing supply agreements and aims to produce HBM4 by the second half of next year.