NVIDIA H100 Hopper
The NVIDIA H100 is the most powerful GPU for AI and HPC workloads, featuring the revolutionary Hopper architecture with Transformer Engine for accelerating large language models.
Memory
80GB
Bandwidth
3.35 TB/s
Architecture
H100
Use Cases
5+
Common Use Cases
Large-scale AI trainingLLM fine-tuningHigh-performance computingScientific simulationsGenerative AI
H100 80GB Pricing & Availability
Compare pricing across datacenters worldwide
| Datacenter | GPU | Interconnect | $/hr | Location | Stock | Type | Deployment | Link |
|---|---|---|---|---|---|---|---|---|
| Hyperstack | H100_80 | SMX | $2.04 | North America | Contact | Reserved | Bare Metal | |
| Hyperstack | H100_80 | SMX | $2.04 | Europe | Contact | Reserved | Bare Metal | |
| Hyperstack | H100_80 | NVLink | $1.37 | North America | Contact | Reserved | Bare Metal | |
| Hyperstack | H100_80 | NVLink | $1.37 | Europe | Contact | Reserved | Bare Metal | |
| Hyperstack | H100_80 | — | $1.33 | North America | Contact | Reserved | Bare Metal | |
| Hyperstack | H100_80 | — | $1.33 | Europe | Contact | Reserved | Bare Metal |
← swipe →