Specifications

Vendor

Micron Technology

Product

HBM4

Capacity per Stack

36 GB

Stack Height

12-Hi

Interface Width

2048-bit

DRAM Process

1-gamma (1y) with EUV lithography

Status

Final customer samples delivered; capacity expansion underway

Capacity Target

Approximately 15,000 HBM4 wafers per month by end of 2026 (reported)

Application

Next-generation AI accelerators (Rubin-generation class)

Packaging

Advanced 2.5D/3D with logic base die

One-gamma EUV is Micron's differentiator

Micron is the third of the three HBM suppliers and the one that leans hardest on EUV. Its HBM4 DRAM is built on a 1-gamma node with EUV lithography, which Micron has been the most aggressive adopter of among the HBM makers. The engineering argument for that choice is die-density and overlay precision at the bit densities HBM4 requires, which translates into yield and, ultimately, cost per stack once the node matures.

Micron has stated that its HBM4 met the revised NVIDIA specification for the Rubin-generation accelerators and that final customer samples have been delivered, with capacity being expanded toward roughly 15,000 HBM4 wafers per month by the end of 2026. As with all HBM programmes, allocation follows the accelerator programme rather than open distribution — buyers plan against accelerator timelines and the HBM vendor relationship, not against catalogue stock.

Micron products are supplied by QS Compute with B2B pricing, global shipping and short sample lead times for AI data center, edge deployment and industrial integration projects.

Request a Quote — Micron HBM4 36GB 12-Hi High Bandwidth Memory — 2048-bit Interface (Micron HBM4 36GB)

QS Compute — Global B2B AI computing hardware supply, 15-day sample lead time.

Get Your Quote →