[Edaily Reporter SOYEON KIM ] SK hynix announced on the 4th that it has unveiled the first standard specifications for High Bandwidth Flash (HBF), a next-generation storage technology, in collaboration with SanDisk.

SK hynix is unveiling the standard specifications for the first time in conjunction with “FMS (Future of Memory and Storage) 2026,” taking place from the 4th to the 6th (local time) at the Santa Clara Convention Center in California, U.S. The company plans to introduce HBF as a core technology to resolve memory bottlenecks in the age of artificial intelligence (AI) through keynote speeches and panel discussions at the event.

HBF is a new memory tier positioned between HBM and SSDs. Like HBM, it enables fast data transfer while significantly increasing capacity thanks to its NAND-based architecture. As AI inference becomes more widespread and the volume of data to be processed increases rapidly, HBF is gaining attention as a technology capable of simultaneously addressing both bandwidth and capacity scalability.

This specification is the result of efforts that began last August with a collaboration between SanDisk and the HBF standardization consortium, and was released approximately six months after the consortium was launched in February of this year. Capacity is defined up to a maximum of 512 GB (gigabytes) based on two stack configurations: 8-layer and 16-layer NAND dies. Bandwidth is divided into three grades (Grade 1–3) and is specified to enable data transfer speeds ranging from 0.4 TB (terabytes) per second to 3.0 TB.

The method of connecting HBF to processors is particularly noteworthy. By adopting the industry-standard Chiplet Connection Interface (UCIe), the technology enables flexible integration of HBF with various types of processors, such as graphics processing units (GPUs) and central processing units (CPUs).

The specification was released through the Open Compute Project (OCP), the world’s largest open data center technology consortium. It is an open standard available for use by the entire industry. The company plans to use this as an opportunity to accelerate the development of the HBF ecosystem in the AI storage market. Currently, Google and Tenstorrent are participating in the HBF consortium.

On the first day of FMS 2026, Kim Cheon-sung, Head of Solution Development (Senior Vice President), and Kang Wook-sung, Head of Next-Generation Product Planning (Senior Vice President), will deliver a joint keynote speech. The topic is “The Era of Agent AI: Strategies for Efficiently Implementing AI Infrastructure Based on Hierarchical Memory.” The two speakers will present “hierarchical memory systems” as a key strategy for resolving data bottlenecks. They will explain that the memory tiers can be implemented using technologies such as 3D-stacked DRAM for xPUs, high-bandwidth flash (HBF), and CXL-based memory pooling.

On the 6th, Lim Ui-cheol (Solution AT) from SK hynix, Xiaoyu Ma, a researcher at Google DeepMind, and Rajeev Nagabhirava, Vice President at SanDisk, will gather for a panel discussion. This event brings together three global leaders in the memory, storage, and AI sectors, with the theme “Breaking the Memory Wall with High Bandwidth Flash.” The three speakers will highlight the fundamental shifts in memory architecture occurring across AI infrastructure and focus on the role and potential of HBF as a solution to overcome these challenges.

SK hynix will operate an exhibition booth during FMS 2026. At this event, SK hynix will unveil for the first time the wafers and products of its 10th-generation (V10) 375-layer 4D NAND, which is currently in active development. The 375-layer 4D NAND features a 2.5-fold increase in performance per watt compared to the previous generation. It is optimized for AI infrastructure environments, such as data centers, that demand both power efficiency and performance. The company plans to begin mass production of high-performance, high-capacity eSSDs based on this technology early next year. In addition, it will showcase a product lineup applicable to various fields, including mobile, PC, automotive, and robotics.

Vice President Kim Cheon-sung stated, “As the use of AI rapidly expands, we have reached a point where a redesign of the entire data processing architecture is required,” adding, “Through HBF, we will contribute to the implementation of a new architecture that blurs the boundaries between memory and storage and enhances the efficiency of the entire system.”