Specifications
Capacity
36GB (12-High) / 48GB (16-High)
Pin Speed
>11 Gb/s
Bandwidth
>2.8 TB/s per stack
Power
+20% efficiency vs HBM3E
Bandwidth Gain
2.3× vs HBM3E
Stack
12-High (HVP) / 16-High (sampling)
Status
High-volume production (36GB)
Overview
Micron's HBM4 36GB 12-High entered high-volume production in Q1 2026 and is designed into NVIDIA Vera Rubin. It delivers more than 11 Gb/s pin speed and more than 2.8 TB/s of bandwidth — a 2.3× improvement and more than 20% better power efficiency over Micron HBM3E.
Micron has also shipped samples of a 48GB 16-High HBM4 stack to customers, raising capacity per HBM placement by 33% over the 36GB 12-High product. QS Compute supplies Micron HBM4 for GPU board, accelerator, and CoWoS packaging builds.
Request a Quote — Micron HBM4 36GB / 48GB
QS Compute — Global B2B AI computing hardware supply, 15-day sample lead time. Price available on quote; stock available to order.
Get Your Quote →