HBM2e
HBM2e (High Bandwidth Memory 2e (Enhanced)) is a JEDEC HBM memory generation from 2020, with 460 GB/s of bandwidth per stack.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
HBM2e (High Bandwidth Memory 2e (Enhanced)) is a JEDEC HBM memory generation from 2020, with 460 GB/s of bandwidth per stack.
Basic
Enhanced HBM2 used in NVIDIA A100, Google TPU v4, Intel Gaudi 2, and various HPC accelerators. Mature technology with declining demand as the industry transitions to HBM3/3e. Still in production for legacy deployments. Status in the dataset: mature.
Deep
Interface width is 1,024 bits at 1.2 V. Capacity per stack: 8, 16 GB, up to 8-high stacks. Suppliers: Samsung (volume), SK hynix (volume) and Micron (eol). Main buyers: NVIDIA, Google and Intel.
Expert
The dataset prices it at about $8.5 per GB (-22% year over year), trend declining. AI accounts for about 60% of its demand in the dataset. All figures are dated with the curated memory dataset; see the memory page for sources.
Depending on why you're here
- ·460 GB/s per stack · 1,024-bit interface
- ·JEDEC HBM · 2020
- ·Capacity per stack: 8, 16 GB
- ·Compare generations on /memory/hbm2e
- ·Suppliers: Samsung, SK hynix and Micron
- ·Price trend: declining
- ·HBM2e is a type of memory chip
- ·It sits right next to AI chips and feeds them data very fast
Frequently Asked Questions
Read the primary sources
- Samsung HBM2E product briefsemiconductor.samsung.com