SK hynix and Sandisk publish first HBF specs for AI memory up to 512GB
firstadopter · x · 2026-08-04
SK hynix and Sandisk publish the first HBF standard specs
SK hynix says it has unveiled the first standard specifications for High Bandwidth Flash (HBF) with Sandisk, and will present the technology at FMS 2026 in Santa Clara.
What HBF is
- HBF is positioned as a memory layer between HBM and SSDs.
- It aims to provide HBM-like bandwidth with much higher capacity, using NAND technology.
- The company frames it as a way to address the bandwidth and capacity bottlenecks created by growing AI inference workloads.
Key details from the spec
- Capacity specs go up to 512GB.
- The standard covers 8-high and 16-high NAND stack configurations.
- Bandwidth is divided into three grades, ranging from roughly 0.4 TB/s to 3.0 TB/s.
- HBF uses UCIe for processor interconnects, so it can connect across CPUs and GPUs.
Why it matters
- The announcement follows the formation of the consortium earlier this year and an initial standardization partnership with Sandisk.
- SK hynix argues HBF is meant for the new AI era, where inference systems need both speed and scale.
- The spec was disclosed through the Open Compute Project, signaling an open-industry standard rather than a proprietary one.
More from Infra
- AMD pre-call read says the real test is 70% server CPU growth and MI450 timing — tengyanAI · 2026-08-04
- OpenAI's $51M Chip Deal with Altman-Backed Rain AI Faces Uncertainty — suchenzang · 2026-08-04
- OpenAI once lined up $51M for Rain AI chips as Altman personally invested — tinyfool · 2026-08-04
- Inference engineering is becoming a new middle-layer market in AI — dotey · 2026-08-04
- FlashAttention 2, SGLang and DeepSeek v3 named as modern AI’s most important open-source projects — hyhieu226 · 2026-08-04
- Fluidstack is hiring across dozens of data center roles as it builds gigawatt-scale AI infrastructure — MxMnr · 2026-08-04