SK Hynix and Sandisk have officially joined forces to promote the global standardization of High Bandwidth Flash (HBF). This new storage technology is positioned as a key solution to fill the hierarchical gap between HBM and SSD in the era of AI inference. The collaboration between the two companies marks a substantial phase in the competitive landscape of next-generation storage architecture.On February 25, the two companies held the &#34;HBF Specification Standardization Alliance Launch&#34; event at Sandisk's headquarters in Milpitas, California, announcing the establishment of a dedicated working group under the Open Compute Project (OCP) framework to formally advance HBF standardization efforts. Ahn Hyun, President and Chief Development Officer of SK Hynix, stated that the standardization of HBF technology will help the company build a collaborative system, &#34;presenting an optimized storage architecture for customers and partners in the AI era, creating new value.&#34;In terms of commercialization expectations, Professor Kim Joungho of the Korea Advanced Institute of Science and Technology, known as the &#34;father of HBM,&#34; recently indicated that the pace of HBF implementation is significantly ahead of previous expectations. Samsung Electronics and Sandisk plan to integrate HBF into products from NVIDIA, AMD, and Google by the end of 2027 to early 2028, with the market scale expected to surpass HBM around 2038 in the long term.For the market, the launch of this standardization means that the strategic position of suppliers capable of providing full-stack storage solutions with both HBM and HBF will continue to rise. The industry anticipates that the demand for HBF-related storage solutions will enter a rapid growth phase around 2030.<h2>HBF Positioning: Filling the Storage Hierarchical Gap Between HBM and SSD</h2>The core value of HBF lies in constructing a new storage hierarchy between ultra-high-speed HBM and high-capacity SSD. In AI inference scenarios, as the user base of AI services rapidly expands, existing storage architectures face a structural contradiction in balancing high-capacity data processing and power efficiency—HBM offers excellent bandwidth but limited capacity, while SSDs provide ample capacity but insufficient read/write speeds.HBF is designed to bridge this gap by vertically stacking NAND flash memory, providing approximately 10 times the storage capacity of HBM while maintaining high bandwidth. In the system architecture, HBM is responsible for handling high-level bandwidth demands, while HBF serves as a supporting layer to undertake capacity expansion tasks, with both working together to meet the dual requirements of massive data processing and power efficiency in AI inference.SK Hynix points out that HBF not only enhances the scalability of AI systems but also has the potential to reduce total cost of ownership (TCO). In the AI inference market, the overall optimization of CPU, GPU, and memory at the system level is increasingly becoming the core determinant of comprehensive competitiveness, with the competition of single-chip performance giving way to the capability of full-stack solutions.<h2>Industrial Collaboration under the OCP Framework: Standardization Shifting from Bilateral to Multilateral</h2>Both SK Hynix and Sandisk possess design, packaging, and mass production experience in the fields of HBM and NAND, and this collaboration is based on this technical foundation to push for HBF standardization. According to previous reports, both Samsung Electronics and SK Hynix have signed memorandums of understanding with Sandisk to jointly promote HBF standardization, aiming to bring products to market by 2027As the world's largest open data center technology organization, OCP provides a widely recognized platform for promoting HBF standardization. The establishment of this exclusive working group signifies that the HBF standardization efforts are transitioning from bilateral agreements to broader industry collaboration.Ahn Hyun stated, &#34;The key to AI infrastructure lies in surpassing performance competition of individual technologies and achieving optimization of the entire ecosystem.&#34; SK Hynix also mentioned that establishing HBF as an industry standard will lay the foundation for the collective growth of the entire AI ecosystem.<h2>Market Outlook: Demand Acceleration Around 2030, Long-term Potential to Surpass HBM</h2>According to a previous article by Wall Street Insight, from a temporal perspective, Kim Joungho expects HBF to achieve widespread application during the HBM6 launch phase and to surpass HBM in market size around 2038. He also pointed out that, thanks to the process and design experience accumulated from HBM, the commercialization cycle of HBF will be significantly shorter than that of HBM's development history, which is one of the key reasons for the accelerated commercialization of this technology.The fundamental logic driving the demand for HBF is the continuous growth of AI workloads. Currently, the AI industry is accelerating its transition from the training phase focused on building large language models to the inference phase aimed at end users. The demand for fast, efficient, and high-capacity storage in inference scenarios is sharply rising, and HBF, with its capacity advantages brought by NAND stacking, precisely meets the core demands of this application scenario.In this trend, the strategic value of vendors that can provide full-stack storage solutions with both HBM and HBF is increasing. Companies that first gain standardized discourse power within the OCP framework are expected to seize opportunities in this emerging market, where demand is projected to significantly accelerate around 2030

SNDK

FTXL

SOXQ

SOXL

SOXX

SK Hynix and Sandisk are collaborating to promote the global standardization of High Bandwidth Flash (HBF), aiming to fill the hierarchical gap between HBM and SSD. HBF achieves a balance of high bandwidth and large capacity through stacked NAND, optimized specifically for AI inference. The two parties have established a working group under the OCP framework, planning for commercialization in 2027, with its market size expected to surpass HBM in the long term, becoming a key support for AI storage architecture

- SK Hynix and SanDisk are collaborating to advance the global standardization of High Bandwidth Flash (HBF) memory, positioned as a key solution bridging the gap between HBM and SSD.  
- The companies announced the formation of a dedicated working group under the Open Compute Project to expedite HBF standardization, aiming for market introduction by late 2027 or early 2028.  
- With rising demand projected around 2030, HBF could surpass HBM in market size by 2038, enhancing capacity and efficiency for AI workloads.

Wallstreetcn

Sandisk

First Trust Nasdaq Food & Semicon

Invesco Semiconductors ETF

Invesco PHLX Semiconductor ETF

Direxion Semicon Bull 3X

VanEck Semiconductor ETF

SPDR S&P Semicon

iShares Semiconductor ETF

SK Hynix collaborates with Sandisk to launch HBF standardization, with capacity crushing HBM by 10 times!