VLSI KoreaNVIDIA Hopper Architecture In-Depth
NVIDIA specifies H100 SXM5 with 80 GB HBM3 across five HBM3 stacks.
- Published
- 2022-03-22
- Retrieved
- 2026-09-08
Intelligence Graph
The semiconductor ecosystem, one documented connection at a time.
Browse the entity collection
- Microncompany
- HBM3ememory technology
- Yosyseda tool
- Micron Singapore HBM packaging facility projectfacility
- Micron HBM3E 24GB 8Hproduct
- Intelcompany
- NVIDIA fifth-generation NVLinkinterface
- SK hynixcompany
- AMD Instinct MI300Xproduct
- HBM2memory technology
- Blue Cheetah Analog Designcompany
- Icheon-si, Gyeonggi-do, Republic of Korearegion
- HBM3memory technology
- NVIDIA A100 Tensor Core GPUproduct
- NVIDIA Blackwell GPUs (2024 platform)product
- NVIDIA Hopper architecturearchitecture
- AMDcompany
- NVIDIA GA100 GPUchip
- TSMCcompany
- Phoenix, Arizona, United Statesregion
Filters & dates1 step
Unknown validity is excluded when a validity date is selected. Knowledge time records what VLSI had observed; it is separate from the relationship’s valid time.
4 entities and 3 documented connections shown.
EXPLORING
HBM3
memory technology · 4 entities · 3 connections · 1 relationship step
Knowledge snapshot: · Relationship dates may differ
Drag the map or use pan buttons to move. Tab to a node or connection and press Enter to inspect its evidence. Arrows follow the actual claim direction.
CompanyChip / productTechnologyPlace / facility
Select a connection for its evidenceWhat this graph establishes
Relationships may describe different dates. A path does not establish simultaneous activity.
Bounded source-backed collection. Missing or filtered connections do not establish absence. Supported does not mean independently verified.
SUPPORTED records a source-backed assertion. VERIFIED requires an identified human review. This map does not measure supplier concentration or prove that an unrecorded relationship is absent.
THE EVIDENCE
Documented relationships
| Subject | Relationship | Object | Status & dates | Evidence |
|---|---|---|---|---|
| NVIDIA H100 SXM | HBM3 | SUPPORTEDSource-reported snapshot: 2022-03-22 | Why are these connected? · 1 source(s)Variant-specific: NVIDIA also describes PCIe H100 configurations with HBM2e. | |
| AMD Instinct MI300X | HBM3 | SUPPORTEDSource-reported snapshot: 2023-12-06 | Why are these connected? · 1 source(s)No memory-chip vendor is asserted. AMD Delivers Leadership Portfolio of Data Center AI Solutions with AMD Instinct MI300 SeriesAMD specifies MI300X with 192 GB of HBM3 memory.
| |
| AMD Instinct MI300A | HBM3 | SUPPORTEDSource-reported snapshot: 2023-12-06 | Why are these connected? · 1 source(s)No memory-chip vendor is asserted. AMD Delivers Leadership Portfolio of Data Center AI Solutions with AMD Instinct MI300 SeriesAMD specifies MI300A with 128 GB of HBM3 memory.
|
Entities in this view (4)
- AMD Instinct MI300XproductExpand →
- HBM3memory technologyExpand →
- NVIDIA H100 SXMproductExpand →
- AMD Instinct MI300AproductExpand →
TRACE A CONNECTION
Evidence pathfinder
Find paths from HBM3. Each step retains its own claim and evidence; a path does not establish a new supply relationship.