SK Hynix announced the start of mass production of the world’s fastest DRAM memory, and it is none other than the HBM3. The announcement comes just seven months after the company was the first in the industry to announce that it had completed development of HBM3 memory in October.

“With advances in cutting-edge technologies such as artificial intelligence and the acceleration of big data, the world’s leading technology companies are looking for ways to quickly process ever-increasing volumes of data. HBM, which is very competitive in terms of speed of data processing and performance compared to traditional DRAM, to capture industry attention and see increased adoption.”

HBM3 by SK Hynix for Nvidia

The first company that will have access to these chips will be NVIDIA, pending its release in its NVIDIA H100which is based on the Hopper architecture and manufactured by TSMC using the manufacturing process of 4nm. This graphic charter turned towards the professional world proposes a configuration of 18432 CUDA cores accompanied by none other than 80 GB of HBM3 memorywith which it will be able to offer a bandwidth of up to 3 TB/sit’s a 50 percent more compared to its predecessor, the Nvidia A100.

NVIDIA recently completed its performance evaluation of SK Hynix’s HBM3 samples. SK Hynix will supply HBM3 memory for NVIDIA systems expected to ship from the third quarter of this year. SK hynix will increase HBM3 memory volume in H1 according to NVIDIA schedule.

Kevin (Jongwon) Noh, President and Chief Marketing Officer of SK Hynix, says the company has ensured top-notch competitiveness in the high-end DRAM market through its close collaboration with NVIDIA. “Our goal is to become a solution provider that deeply understands and responds to our customers’ needs through open and ongoing collaboration,” he said.