NVIDIA A100 80GB cloud prices
A100 80GB on-demand cloud prices run from $2.32 to $4.00 per GPU-hour across 3 providers gpu.gl tracks (median $3.43). The lowest is DataCrunch.
Lowest price by provider
| Configuration | Where | Fetched | Source | |||
|---|---|---|---|---|---|---|
| DataCrunch | On-demand | $2.32 | 1A100.22V · 1, 2, 4, 8 GPUs | Not stated per region | DataCrunch ↗ | |
| AWS | On-demand | $3.43 | p4de.24xlarge · 8 GPUs | 4 regions | AWS ↗ | |
| OCI | On-demand | $4.00 | B95907 (Compute - GPU - A100 - v2) · 1 GPU | All commercial regions | OCI ↗ | |
| DataCrunch | Spot / interruptible | $1.16 | 1A100.22V · 1, 2, 4, 8 GPUs | Not stated per region | DataCrunch ↗ | |
| AWS | Reserved 1 yr | $2.18 | p4de.24xlarge · 8 GPUs | 4 regions | AWS ↗ | |
| AWS | Reserved 3 yr | $1.46 | p4de.24xlarge · 8 GPUs | 4 regions | AWS ↗ |
Price history
History starts 2026-11-08. gpu.gl stores one snapshot a day from launch and never back-fills, so a chart appears here once there are two daily snapshots with on-demand prices for this GPU (1 so far).
A100 80GB specifications
The A100 80GB has 80 GB of HBM2e and 2.039 TB/s of memory bandwidth. That is about 153 dense FP16 TFLOPS per TB/s of bandwidth; workloads that stream large weights, like LLM inference at small batch sizes, tend to be limited by the bandwidth figure, while large-batch training leans on the compute figure.
| Architecture | Ampere |
|---|---|
| Memory | 80 GB HBM2e |
| Memory bandwidth | 2.039 TB/s |
| FP16 / BF16 tensor (dense) | 312 TFLOPS |
| FP8 tensor (dense) | Not supported |
| Max power | 400 W |
| Form factor | SXM4 |
Source: NVIDIA spec sheet. Tensor figures are dense (without sparsity). Figures for the SXM version; the PCIe version has 1.935 TB/s and a 300 W limit.
Every configuration
Each instance size or listing, at its cheapest region.
| Provider | Price type | $/GPU-hr | Instance | GPUs | Instance price | Cheapest region | Fetched | Source |
|---|---|---|---|---|---|---|---|---|
| DataCrunch | On-demand | $2.32 | 1A100.22V | 1 | $2.32/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | On-demand | $2.32 | 2A100.44V | 2 | $4.63/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | On-demand | $2.32 | 4A100.88V | 4 | $9.26/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | On-demand | $2.32 | 8A100.176V | 8 | $18.53/hr | Not stated per region | DataCrunch ↗ | |
| AWS | On-demand | $3.43 | p4de.24xlarge | 8 | $27.45/hr | US East (N. Virginia) +3 more | AWS ↗ | |
| OCI | On-demand | $4.00 | B95907 (Compute - GPU - A100 - v2) | 1 | $4.00 per GPU-hour | All commercial regions | OCI ↗ | |
| DataCrunch | Spot / interruptible | $1.16 | 1A100.22V | 1 | $1.16/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | Spot / interruptible | $1.16 | 2A100.44V | 2 | $2.32/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | Spot / interruptible | $1.16 | 4A100.88V | 4 | $4.63/hr | Not stated per region | DataCrunch ↗ | |
| DataCrunch | Spot / interruptible | $1.16 | 8A100.176V | 8 | $9.26/hr | Not stated per region | DataCrunch ↗ | |
| AWS | Reserved 1 yr | $2.18 | p4de.24xlarge | 8 | $17.40/hr | US East (N. Virginia) +3 more | AWS ↗ | |
| AWS | Reserved 3 yr | $1.46 | p4de.24xlarge | 8 | $11.72/hr | US East (N. Virginia) +3 more | AWS ↗ |