SK Hynix, the world’s second-largest memory chip maker, is set to commence mass production of its cutting-edge HBM3E 12-layer memory chips by the end of this month, according to a senior company executive. Justin Kim, president and head of the AI Infra division, made this announcement at the Semicon Taiwan industry forum in Taipei on Wednesday.
The South Korean company had previously outlined its plans to introduce the HBM3E chips in the fourth quarter of 2024, with the next-generation HBM4 chips slated for release in the latter half of 2025. High bandwidth memory (HBM) chips, known for their vertical stacking design, enhance performance by saving space and reducing power consumption, making them ideal for demanding applications like generative artificial intelligence (AI).
HBM chips are crucial for graphics processing units (GPUs) used in AI, as they handle large volumes of data generated by complex computational tasks. This advancement is expected to bolster SK Hynix's position in the market, where it competes with Micron and Samsung Electronics.
In May, SK Hynix CEO Kwak Noh-Jung revealed that the company's HBM chips were sold out for the remainder of this year and nearly depleted for 2025. The company has been a key supplier of HBM chips to Nvidia, having provided HBM3E chips to an undisclosed customer in late March.
As the demand for advanced memory chips continues to rise, SK Hynix’s new production capabilities are poised to play a significant role in meeting the needs of the growing AI and high-performance computing markets.
