SK Hynix Inc., a major supplier of Nvidia Corp., said on Friday that it has completed the development of the world’s first HBM4 and is ready to begin mass production of the most advanced AI chip.
The announcement comes about six months after the South Korean chipmaker said it has begun delivering samples of the 12-layer HBM4 to its key customers and planned to get ready for its mass production within the second half of this year.
The sixth-generation high-bandwidth memory (HBM) delivers significantly faster data processing with improved power efficiency by more than 40% compared to HBM3E, the fifth-generation model in the HBM series.
“When customers integrate this HBM4 with their systems, they can improve AI service performance by up to 69%,” said SK Hynix, adding that the technology will alleviate data bottlenecks and significantly reduce data center power consumption.
By doubling the number of I/O terminals, or input/output connectors, to 2,048 compared to the previous generation, HBM4 achieves twice the bandwidth, according to the company.
Bandwidth refers to the maximum amount of data that can be transferred between the memory and processor per second.
“Completion of HBM4 development will be a new milestone for the industry,” Cho Joohwan, head of HBM development at SK Hynix, who has led the development, said in a press release.
“By supplying the product that meets customer needs in performance, power efficiency and reliability in a timely manner, the company will fulfill time to market and maintain competitive position,” he said.
HBM4 boats an operating speed of 10 gigabit per second (Gbps), surpassing the global standard of 8Gbps set by the Joint Electron Device Engineering Council (JEDEC).
A Gbps refers to a unit of data transfer rate representing one billion bits of data transmitted per second.
The latest high-performance memory is expected to be fitted in Nvidia’s next-generation graphics architecture Vera Rubin, the successor of the Blackwell AI chip.
In November 2024, SK Group Chairman Chey Tae-won said that Nvidia Corp.’s Chief Executive Jensen Huang has asked SK to supply 12-layer HBM4 chips, six months earlier than SK’s schedule of early 2026.
By Yeonhee Kim
yhkim@hankyung.com
Yeonhee Kim edited this article.
