Amazon Web Services vs GPU.ai
As of , GPU.ai lists the lower on-demand rate on 9 of the 9 GPUs both offer on-demand. Not counted: 2 more shared GPUs, on community capacity or a marketplace floor. On the priciest shared card both list on-demand, NVIDIA B300: Amazon Web Services $17.80 (us-east-1) vs GPU.ai $7.89 per GPU-hour on-demand — GPU.ai is $9.91 (56%) cheaper.
On-demand, per GPU-hour
| GPU | Amazon Web Services | GPU.ai | Difference | Cheaper |
|---|---|---|---|---|
| NVIDIA B300 median $7.99 · 21 providers · lowest Deep Infra $4.89 | $17.80 us-east-1 captured Deploy ↗ | $7.89 from captured Deploy ↗ | $9.91 (56%) | GPU.ai |
| NVIDIA B200 median $6.89 · 32 providers · lowest Deep Infra $3.69 | $14.24 us-east-1 captured Deploy ↗ | $6.79 from captured Deploy ↗ | $7.45 (52%) | GPU.ai |
| NVIDIA H200 median $4.54 · 42 providers · lowest Beam $2.09 from | $7.91 us-east-1 captured Deploy ↗ | $4.59 from captured Deploy ↗ | $3.32 (42%) | GPU.ai |
| NVIDIA H100 SXM median $3.49 · 31 providers · lowest Cudo Compute $1.79 GPU only | $6.88 us-east-1 captured Deploy ↗ | $3.49 from captured Deploy ↗ | $3.39 (49%) | GPU.ai |
| NVIDIA A100 80GB median $1.80 · 46 providers · lowest Deep Infra $0.89 | $3.43 us-east-1 captured Deploy ↗ | $1.07 from captured Deploy ↗ | $2.36 (69%) | GPU.ai |
| NVIDIA RTX PRO 6000 Blackwell median $2.17 · 34 providers · lowest Hostinger $0.65 from | $3.36 us-east-1 captured Deploy ↗ | $1.99 from captured Deploy ↗ | $1.37 (41%) | GPU.ai |
| NVIDIA V100 median $2.48 · 9 providers · lowest Verda $0.20 | $3.06 us-east-1 captured Deploy ↗ | $0.09 from captured Deploy ↗ | — | not the same market |
| NVIDIA A100 40GB median $1.99 · 17 providers · lowest SynpixCloud $0.79 | $2.74 us-east-1 captured Deploy ↗ | $0.41 from captured Deploy ↗ | — | not the same market |
| NVIDIA RTX PRO 4500 Blackwell median $0.82 · 6 providers · lowest Edgevana $0.39 from | $2.52 us-east-1 captured Deploy ↗ | $0.72 from captured Deploy ↗ | $1.80 (71%) | GPU.ai |
| NVIDIA L40S median $1.21 · 33 providers · lowest Beam $0.76 from | $1.86 us-east-1 captured Deploy ↗ | $0.80 from captured Deploy ↗ | $1.06 (57%) | GPU.ai |
| NVIDIA L4 median $0.80 · 19 providers · lowest Seeweb $0.43 | $0.80 us-east-1 captured Deploy ↗ | $0.49 from captured Deploy ↗ | $0.31 (39%) | GPU.ai |
How this is compared
Cloud GPU rental prices of Amazon Web Services and GPU.ai, compared per GPU-hour on the GPUs both list. Where a provider has several tiers or regions, its cheapest on-demand rate is used and labelled. Community capacity (RunPod Community Cloud) and marketplace floors are shown only where a provider has nothing else, and never decide which is cheaper.
GPU by GPU
- NVIDIA B300: Amazon Web Services $17.80 (us-east-1) vs GPU.ai $7.89 per GPU-hour on-demand — GPU.ai is $9.91 (56%) cheaper.
- NVIDIA B200: Amazon Web Services $14.24 (us-east-1) vs GPU.ai $6.79 per GPU-hour on-demand — GPU.ai is $7.45 (52%) cheaper.
- NVIDIA H200: Amazon Web Services $7.91 (us-east-1) vs GPU.ai $4.59 per GPU-hour on-demand — GPU.ai is $3.32 (42%) cheaper.
- NVIDIA H100 SXM: Amazon Web Services $6.88 (us-east-1) vs GPU.ai $3.49 per GPU-hour on-demand — GPU.ai is $3.39 (49%) cheaper.
- NVIDIA A100 80GB: Amazon Web Services $3.43 (us-east-1) vs GPU.ai $1.07 per GPU-hour on-demand — GPU.ai is $2.36 (69%) cheaper.
- NVIDIA RTX PRO 6000 Blackwell: Amazon Web Services $3.36 (us-east-1) vs GPU.ai $1.99 per GPU-hour on-demand — GPU.ai is $1.37 (41%) cheaper.
- NVIDIA V100: Amazon Web Services $3.06 (us-east-1) vs GPU.ai $0.09 (from) per GPU-hour — not the same market, not counted.
- NVIDIA A100 40GB: Amazon Web Services $2.74 (us-east-1) vs GPU.ai $0.41 (from) per GPU-hour — not the same market, not counted.
- NVIDIA RTX PRO 4500 Blackwell: Amazon Web Services $2.52 (us-east-1) vs GPU.ai $0.72 per GPU-hour on-demand — GPU.ai is $1.80 (71%) cheaper.
- NVIDIA L40S: Amazon Web Services $1.86 (us-east-1) vs GPU.ai $0.80 per GPU-hour on-demand — GPU.ai is $1.06 (57%) cheaper.
- NVIDIA L4: Amazon Web Services $0.80 (us-east-1) vs GPU.ai $0.49 per GPU-hour on-demand — GPU.ai is $0.31 (39%) cheaper.
Not compared: GPUs only one of the two lists, reserved/committed tiers, and anything either provider prices only on request. See each provider's page for its full list.
Data updated: