NVIDIA H100 SXM5
NVIDIA · Hopper
Shipping (allocation constrained)
traininginferenceflagship
Memory
80GB HBM3
FP8 TFLOPS
3958
Power
700W
Released
Oct 1, 2022
Technical Specifications
| Manufacturer | NVIDIA |
|---|---|
| Architecture | Hopper |
| Memory | 80GB HBM3 |
| Memory Bandwidth | 3.35 TB/s |
| FP16 TFLOPS | 1979 |
| FP8 TFLOPS | 3958 |
| Interconnect | NVLink4 (900GB/s) |
| Power (TDP) | 700W |
| Released | Oct 1, 2022 |
| Status | Shipping (allocation constrained) |
Availability
Constrained — major allocations to hyperscalers and AI labs
Pricing
Retail (estimated)
30000-40000
Cloud Pricing ($/hr)
| aws_p5_xlarge | $12.99/hr |
|---|---|
| azure_nd96asr_v4 | $11.91/hr |
| gcp_a3_highgpu | $11.8/hr |
| coreweave | $2.23/hr |
Key Specs
Memory: 80GB HBM3
Bandwidth: 3.35 TB/s
FP8: 3958 TFLOPS
Power: 700W
Datacenters Using This GPU
- Stargate Abilene Campus
- xAI Memphis Supercluster
- Google The Dalles AI Campus
- Crusoe West Texas AI Campus
- Microsoft Quincy Data Center Campus
- Microsoft San Antonio Data Center
- Microsoft West Des Moines AI Campus
- Microsoft Boydton Data Center
- Google Council Bluffs Data Center
- Google Douglas County Data Center
- Google Mayes County Data Center
- Google New Albany Data Center
- AWS Ashburn Data Center Complex
- AWS Boardman Data Center
- Meta Eagle Mountain Data Center
- Meta DeKalb Data Center
- Meta Sarpy County Data Center
- Meta Huntsville Data Center
- CoreWeave New Jersey AI Campus
- Lambda Cloud GPU Infrastructure
- Together AI Training Cluster
- Crusoe North Dakota Digital Flaring Campus
- South Korea AI Infrastructure Cluster
- Sakura Internet Ishikari AI Campus
- Yotta D1 AI Data Center
- Jean Zay Supercomputer Expansion
- UK AI Research Resource (AIRR)
- Canada AI Compute Infrastructure (Toronto)
- Singapore AI Compute Cluster
- NVIDIA Israel-1 Supercomputer