HBF, or High Bandwidth Flash, is being touted as the next big thing in the ever-evolving memory universe. At its recent Investor Day, SanDisk argued that HBF can offer better capital efficiency than HBM. Can it?1
Memory is one of the most fascinating storylines of the current AI build. It used to be a perennially cyclical business. But data center demand has changed everything. There is a strong argument that memory demand is becoming more structural.
Cynics still argue that memory is, at the end of the day, a commodity, and that once capacity catches up, pricing will revert to normal or crater. A couple of years ago, DRAM and NAND diverged sharply.
Even the hardest cynics have more or less agreed that HBM (DRAM) has a technology moat that can extend the current cycle. That leaves NAND in a tough spot.
SK hynix and SanDisk began work on HBF standardization through the Open Compute Project. Google and Tenstorrent joined the workstream, and Meta joined around the August specification release.1,2
Read our white paper to learn how HBF could reshape memory markets, amid evolving data center demands.