V100 cloud price per hour
We price 5 offers from 4 providers. The cheapest is $0.118 per GPU-hour at DataCrunch , the dearest $0.790 at Lambda. That is a 6.7× spread for the same card, so where you rent it matters more than anything else.
Every provider
Sorted by hourly rate. Spot and community tiers can be interrupted.
Cheapest nowDataCrunch (spot) at $0.12 /hr
| Source | ||||||
|---|---|---|---|---|---|---|
| DataCrunch Best value | spot | $0.118 | 16 GB | $0.0074 | $11.80 | ↗ |
| Vast.ai | community | $0.135 | 16 GB | $0.0084 | $13.46 | ↗ |
| RunPod | community | $0.190 | 16 GB | $0.012 | $19.00 | ↗ |
| DataCrunch | on-demand | $0.236 | 16 GB | $0.015 | $23.60 | ↗ |
| Lambda | on-demand | $0.790 | 16 GB | $0.049 | $79.00 | ↗ |
Prices last checked: . Methodology
Questions
Answered from the V100 price data on this page.
How much does an V100 cost per hour?
The cheapest we track is $0.118 per GPU-hour at DataCrunch on the spot tier, or $0.236 on demand at DataCrunch. Checked 28 Oct 2026.
What does a 100-hour training run on an V100 cost?
$11.80 at the cheapest rate we track, and $79.00 at the dearest. That spread is for the same card, so where you rent it matters more than anything else.
Is spot or community pricing worth it for an V100?
The cheapest interruptible rate is $0.118 against $0.236 on demand, a saving of 50%. Worth it for checkpointed training, not for serving.