COMPUTE / GLOBAL MARKETS
NVIDIA H100
Hopper architecture · Training · HPC. Compare actual provider offers.
Market pulseLoading observations
LOWEST OBSERVED—USD / GPU-hour
MEDIAN OBSERVED—Across provider observations
AVAILABLE GPUS—Only explicitly reported capacity
PRICE DISPERSION—Range divided by median
PRICE DISCOVERY
H100 observed median / USD per GPU-hour
+History starts with the first observation.
No historical prices have been collected yet.
Actual observations only. Missing periods are not interpolated into a historical market.
Available offers
Waiting for a market connection
Provider credentials unlock current prices and capacity. No inventory or historical prices are simulated.
View provider connectionsHardware profile
- Architecture
- Hopper
- Reference memory
- 80 GB
- Use cases
- Training · HPC
- Benchmarks
- Available only when reported by the provider