GPUs and cloud rental prices
Each published GPU model has its own reference price, provider comparison, regional offerings and price history. Every price below is for one GPU for one hour.
How the reference price is calculated| GPU | Memory and architecture | Reference price | Published month | Providers publishing |
|---|---|---|---|---|
| NVIDIA H100 80GB SXM The H100 is the GPU most large language models were trained on between 2023 and 2025, and it remains the reference unit for AI compute pricing. It is one of the two GPUs CME Group and Silicon Data plan to list rental-index futures on. | 80GB · Hopper | $11.66 | Sep 2026 | AWS, Azure, Google Cloud, Oracle |
| NVIDIA B200 180GB The B200 is the first Blackwell-generation datacenter GPU to reach on-demand cloud rental, at roughly twice the hourly list price of an H100. It is the second of the two GPUs CME Group and Silicon Data plan to list rental-index futures on. | 180GB · Blackwell | $15.74 | Sep 2026 | AWS, Google Cloud, Oracle |
| NVIDIA H200 141GB SXM The H200 is an H100 with 141GB of faster HBM3e memory, which lifts throughput on models too large to sit comfortably in 80GB. It reached the clouds through 2024 and 2025, so its price series is short by construction rather than by omission. | 141GB · Hopper | $11.05 | Sep 2026 | AWS, Azure, Google Cloud, Oracle |
| NVIDIA A100 80GB SXM The A100 80GB is the previous training generation and carries the longest price series GPUQuant holds. It is the clearest available record of what happens to cloud rental prices as a GPU generation ages. | 80GB · Ampere | $4.97 | Sep 2026 | AWS, Azure, Google Cloud, Oracle |