CoreWeave vs GPU.ai
Two GPU specialist providers, priced against each other on the 5 GPUs both of them rent. Published per-GPU on-demand rates as each provider last stated them, with what their own pages say about billing, commitment and egress beside the numbers.
GPU.ai is cheaper on 5 of the 5 cards both of them rent, by a median of 43%.
- How they sell
- GPU specialist An operator whose whole business is renting these cards.
- GPU models
- 10 40 offers
- Registered in
- 🇺🇸 United States
- How they sell
- GPU specialist An operator whose whole business is renting these cards.
- GPU models
- 15 17 offers
- Registered in
- 🇦🇪 United Arab Emirates
The 5 cards both of them rent, cheapest published per-GPU on-demand rate each. Dearest first.
| GPU | CoreWeave | GPU.ai | Difference |
|---|---|---|---|
| NVIDIA H200 SXM5 141GB | $6.31/hr | $3.59/hr | 43% cheaper on GPU.ai |
| NVIDIA H100 SXM5 80GB | $6.16/hr | $3.49/hr | 43% cheaper on GPU.ai |
| NVIDIA A100 80GB | $2.70/hr | $1.07/hr | 60% cheaper on GPU.ai |
| NVIDIA L40S 48GB | $2.25/hr | $0.80/hr | 64% cheaper on GPU.ai |
| NVIDIA L40 48GB | $1.25/hr | $0.78/hr | 38% cheaper on GPU.ai |
Per GPU per hour, on-demand list prices as each provider last published them. Rates exclude tax and vary by region; a cheaper rate is not the same machine, the same network or the same support.
| CoreWeave | GPU.ai | |
|---|---|---|
| Billing granularityOn short jobs this outweighs the rate: an hour minimum on a 90-second run costs more than a 30% dearer per-second rate. | Hourly usage, invoiced monthly in arrears | Per second |
| Minimum commitment | not stated | $5 minimum deposit, no term |
| Payment model | Invoice on receipt; reserved and bulk credits per order form | Prepaid credit balance with burn rate and runway |
| Uptime SLA | 99.9% across regions, 99% single region | not stated |
| Regions | US and European data centres; General and Dedicated Access regions | 8 regions: US East/Central/West, Canada, EU West, EU East, India South, Asia-Pacific |
Neither site tells you whether your data trains their models, what egress costs, what network speed you get or how fast support answers. We looked on both and record the absence rather than filling it in — if you need any of it, ask before you commit.
- NVIDIA GH200 96GB$6.50/hr
- NVIDIA B200 SXM6 180GB$8.60/hr
- NVIDIA GB200 NVL72 192GB$10.50/hr
- NVIDIA RTX 2000 Ada 16GB$0.24/hr
- NVIDIA RTX A6000 48GB$0.40/hr
- NVIDIA L4 24GB$0.44/hr
- NVIDIA A40 48GB$0.49/hr
- NVIDIA RTX 6000 Ada 48GB$0.65/hr
- NVIDIA A100 SXM4 40GB$0.89/hr
- NVIDIA RTX PRO 6000 Blackwell 96GB$1.99/hr
- NVIDIA H100 PCIe 80GB$2.48/hr
- NVIDIA H100 NVL 94GB$3.19/hr
- NVIDIA H200 NVL 141GB$3.29/hr
Is CoreWeave cheaper than GPU.ai?
GPU.ai is cheaper on 5 of the 5 cards both of them rent, by a median of 43%.
Which GPUs do CoreWeave and GPU.ai both rent?
NVIDIA H200 SXM5 141GB, NVIDIA H100 SXM5 80GB, NVIDIA A100 80GB, NVIDIA L40S 48GB, NVIDIA L40 48GB — 5 in total, on published per-GPU on-demand rates.
What is the cheapest GPU at CoreWeave and at GPU.ai?
Of the cards both rent, CoreWeave starts at $1.25 per GPU-hour and GPU.ai at $0.78.