Neocloud
CoreWeave
The flagship neocloud: Kubernetes-native GPU fleets built for training clusters, now public company.
Grew from Ethereum mining to a NASDAQ-listed AI cloud running some of the largest H100/H200/GB200 fleets outside the hyperscalers. Sells mostly via reserved contracts; on-demand exists but reserved is where pricing shines. SUNK (Slurm on Kubernetes) targets serious training teams.
GPU pricing on CoreWeave
On-demand list per GPU-hour, last reviewed 2026-09. Methodology.
| GPU | $/GPU-hr | Notes |
|---|---|---|
| L40S | $1.28 | |
| A100 80GB | $2.06 | |
| H100 SXM | $3.35 | |
| H200 | $4.15 | |
| B200 | $6.48 | HGX B200 on-demand; reserved much lower |
Strengths
- Frontier-scale InfiniBand clusters
- First to new NVIDIA silicon
- Kubernetes/Slurm-native ops
Watch for
- Oriented to committed contracts
- On-demand selection thinner than reserved