NewsBite analysis · AI, Cloud, Hardware
wccftech
SK hynix, In Collaboration With SanDisk, Unveils The New High Bandwidth Flash (HBF) Standard, Helping To Resolve AI Inference Bottlenecks, Targeting Up To 3TB/s Bandwidth
In a move to resolve the performance disparity between High Bandwidth Memory (HBM) and SSDs, SK hynix and SanDisk have unveiled the High Bandwidth Flash (HBF) standard. This new memory standard promises bandwidths reaching up to 3TB/s, positioning itself between HBM and traditional SSDs by utilizing NAND flash. The HBF standard aims to enhance data storage capacity for AI accelerators, particularly addressing the increasing demands of large language models and AI inference processing. By adopting HBF, companies can leverage its capability to support larger AI models efficiently.
Act nowConfidence: HighShort term horizon
Published 4 Aug 2026, 12:20 UTC2 min read
Visit original article ↗Lead image — wccftech
Summary
TL;DR
- HBF targets bandwidths between 0.4TB/s to 3TB/s.
- Utilizes NAND flash instead of DRAM, increasing storage capacity.
- Supports UCIe specification for flexible connections to CPUs and GPUs.
- Designed to complement, not replace, existing HBM solutions.
- Addresses growing needs for AI model storage and processing.
Interesting facts
- HBF can support capacities up to 512GB.
- The technology is expected to play a crucial role in the adoption of larger AI models.
Coverage gaps