HBM Industry Revenue Growth
Total quarterly revenue across all HBM (High-Bandwidth Memory) suppliers. Up roughly 20x since early 2022 because every AI accelerator now ships with 80–192 GB of stacked DRAM — making HBM the highest-margin memory product in the industry.
Total HBM industry revenue (all suppliers combined) in $ billions per quarter. Revenue surged 20x from early 2022 driven by AI accelerator demand. Q1-Q2 2025 are estimates. Sources: TrendForce, company earnings, analyst reports.
HBM Generation Comparison
How each HBM (High-Bandwidth Memory) generation compares on bandwidth, capacity per stack, layer count, and which AI chips use it. HBM3E is the volume product today (NVIDIA H200, B100); HBM4 begins shipping in late 2025 with roughly 2x the bandwidth.
| Generation | Year | Bandwidth | Capacity/Stack | Layers | TSV Pitch | Used In | Status |
|---|---|---|---|---|---|---|---|
| HBM2 | 2018 | 256 GB/s | 8 GB/stack | 8-Hi | ~40 μm | NVIDIA V100AMD MI60 | Legacy |
| HBM2E | 2020 | 460 GB/s | 16 GB/stack | 8-Hi | ~36 μm | NVIDIA A100Intel Gaudi 2 | Mature |
| HBM3 | 2022 | 665 GB/s | 24 GB/stack | 12-Hi | ~30 μm | NVIDIA H100AMD MI300X | Mass Production |
| HBM3E | 2024 | 1.18 TB/s | 36 GB/stack | 12-Hi | ~26 μm | NVIDIA B200AMD MI325X | Mass Production |
| HBM4 | 2025–2026 | 2.0 TB/s | 48 GB/stack | 16-Hi | ~20 μm | NVIDIA Rubin (expected)AMD MI400 (expected) | Sampling |
| HBM4E | 2026–2027 | 3.25 TB/s | 64 GB/stack | 16-Hi | ~18 μm | Next-gen AI accelerators | Development |
HBM specifications by generation. Bandwidth figures are per-stack maximums. HBM4/4E specs are preliminary based on JEDEC standards and manufacturer previews. Sources: JEDEC, SK Hynix, Samsung, Micron product announcements.