Samsung Electronics is seeking to secure trademark rights related to high-bandwidth flash (HBF), a next-generation artificial intelligence (AI) memory technology.
According to industry sources, Samsung recently filed for the 'zHBF' trademark. This move is seen as a proactive measure to prepare for the future, while actual development is confirmed to be underway with the zNAND-O solution unveiled at FMS.
Unlike high-bandwidth memory (HBM), which stacks DRAM vertically, HBF utilizes stacked NAND flash to achieve both high capacity and fast data transfer speeds. It is gaining attention as a new memory tier that bridges the performance and capacity gap between HBM and solid-state drives (SSDs).
As the AI market expands from training to inference, the demand for high-capacity, high-bandwidth memory like HBF is increasing. The rapid growth in the size of AI models and the volume of data processed makes it challenging to meet memory capacity needs with the expensive HBM alone. HBF leverages the high-capacity characteristics of NAND while supplying data faster than traditional SSDs, alleviating memory bottlenecks for AI accelerators.
Samsung is also focusing on developing its next-generation NAND solution, zNAND-O. This product was first introduced at the FMS 2026 event held earlier this month in Santa Clara, California. It is designed to address the high-performance NAND requirements for processing large volumes of data during AI inference.
Additionally, Samsung is broadening its scope in next-generation AI memory technology. At FMS, the company showcased the concept of 'zHBM,' a new 3D memory that vertically stacks computational units and memory. This design aims to reduce physical distance between memory and accelerators, thereby minimizing latency and power consumption during data transfer.
However, the filing for the zHBF trademark does not immediately indicate the development or commercialization of a separate product. It reflects a strategic move to secure intellectual property rights in anticipation of future opportunities, as the technology and product names in the next-generation AI memory market are still being established.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.

