NVHBM is a custom high-bandwidth memory (HBM) architecture developed by NVIDIA as part of the NVLink Fusion ecosystem. It is designed to increase memory bandwidth, reduce HBM power, and save area on the XPU compute die.
NVHBM targets custom XPUs integrated into NVIDIA's infrastructure via NVLink Fusion. Amazon's Annapurna Labs is the first to work with NVIDIA on NVHBM.
High-bandwidth memory (HBM) · serves as: Connectivity / data interconnect.
Which group NVHBM belongs to and how it is built
A subcategory covering memory components — including high-bandwidth memory (HBM) — critical to the performance of AI accelerators and compute systems.
A component type covering high-bandwidth memory (HBM) — used in AI accelerators to deliver massive bandwidth while optimizing power draw and silicon area.
