CoreWeave vs RunPod
for GPU capacity.
CoreWeave sells 8-GPU HGX instances on Kubernetes, aimed at clusters. RunPod rents single GPUs by the second, plus serverless workers and instant clusters.
Facts checked against each provider's own pages
The short answer
Choose CoreWeave for multi-GPU clusters on managed Kubernetes or Slurm[1]: an 8× HGX H100 instance is $49.24/hour on demand[2], with free internet egress[2]. Choose RunPod for single GPUs and short jobs billed per second[3], with H100 PCIe pods from $1.99/hr on Community Cloud[4] and serverless workers that scale to zero[3].
CoreWeave vs RunPod on the facts
| CoreWeave | RunPod | |
|---|---|---|
| Category | GPU cloud | GPU cloud |
| Pricing model | Per instance-hour, on demand or spot; storage per GB-month; committed pricing via sales[2] | Pay as you go: Pods, Serverless workers and Clusters billed by usage; storage per GB-month[4] |
| Entry paid plan | GH200 (1 GPU) $6.50/hr; 8× L40 $10.00/hr; 8× HGX H100 $49.24/hr on demand[2] | No plans; cheapest Secure Cloud pod listed: RTX A5000 $0.27/hr; H100 PCIe $2.89/hr[4] |
| Free tier | — | — |
| Compute billing granularity | Priced per instance-hour; finer metering not stated[2] | Per second for Pods and Serverless (Serverless rounded up to the nearest second)[3] |
| Internet egress (first paid tier) | Free: data transfer between CoreWeave and the internet[2] | No fees for data ingress or egress on Pods[6] |
| Included transfer | No ingress, egress or transfer fees[2] | Not metered on Pods[6] |
| Regions / locations | 52 data centers across North America and Europe[5] | 31 global regions (also described as 30+)[4] |
| Deploy model | CoreWeave Kubernetes Service (managed Kubernetes on bare metal) and SUNK (Slurm on Kubernetes)[1] | Docker images: Pods from templates; Serverless workers run your image with a Python handler[7] |
| GPUs | Yes: GB300/GB200 NVL72, HGX B300/B200, H200, H100, RTX PRO 6000, GH200, L40S, L40, A100[2] | Yes: B300, B200, H200, H100 (PCIe/SXM/NVL), A100, L40S, L40, RTX 6000 Ada, L4 and consumer cards[4] |
| Managed databases | — | — |
| Preview environments | — | — |
| Persistent volumes | Yes: Distributed File Storage volumes via PVCs ($0.070/GB/mo)[2] | Yes: volume disks ($0.10/GB/mo running) and network volumes ($0.07/GB/mo under 1 TB)[8] |
| Scale to zero | — | Serverless flex workers scale to zero (default idle timeout 5 s); active workers run 24/7[3] |
| Config file | — | — |
Which to pick, by use case
| If you need… | Pick | Why |
|---|---|---|
| Multi-node training | CoreWeave | 8-GPU HGX instances and SUNK (Slurm on Kubernetes)[1]; RunPod also offers Instant Clusters with no commitment[4]. |
| One GPU for a few hours | RunPod | Per-second pods with no commitment[3]; CoreWeave prices most GPUs as 8-GPU instances[2]. |
| Scale-to-zero inference | RunPod | Serverless flex workers[3]; CoreWeave doesn't list a scale-to-zero product. |
| Moving lots of data | Either | CoreWeave lists internet transfer as free[2]; RunPod has no egress fees on Pods[6]. |
Per-GPU prices side by side
Dividing CoreWeave's 8-GPU instance prices by eight gives $6.155/GPU-hour for H100, $6.305 for H200, $8.60 for B200 and $2.70 for A100[2]. RunPod's Secure Cloud pods list H100 SXM at $4.39/hr, H200 at $5.29/hr and B200 at $7.99/hr[4]. The full table, including the hyperscalers, is on the GPU cloud pricing page.
Run CoreWeave and RunPod together
You don't have to pick one. AW3 Host connects both your CoreWeave and RunPod accounts and deploys to either from one dashboard and API. A reserved CoreWeave pool can carry baseline inference while RunPod absorbs bursts, with weighted routing between them.
Weighted routing splits traffic between them by percentage, canary and blue-green releases promote or roll back a version in one step, and failover driven by health checks every 60 seconds moves traffic to the healthy side. Daily cost sync and budget policies show both providers' spend side by side; secrets are encrypted per environment. Free tier: 500 credits a month; Pro: $29/month.
Frequently asked questions
- Is CoreWeave or RunPod cheaper for H100s?
- Per GPU-hour on list prices, RunPod: H100 SXM pods are $4.39/hr on Secure Cloud and $2.69/hr on Community Cloud. CoreWeave's 8× HGX H100 instance is $49.24/hr, which is $6.155 per GPU-hour.
- Does CoreWeave charge for egress?
- CoreWeave's pricing page lists data transfer between CoreWeave and the internet as free, with no ingress, egress or transfer fees on storage.
- How do you deploy on CoreWeave versus RunPod?
- CoreWeave runs workloads on CoreWeave Kubernetes Service (managed Kubernetes on bare metal) or SUNK (Slurm on Kubernetes). RunPod runs Docker images as Pods from templates, or as Serverless workers with a Python handler.
Sources
- CoreWeave Docs: CKS. https://docs.coreweave.com/products/cks (checked 2026-11-16)
- CoreWeave: Pricing. https://www.coreweave.com/pricing (checked 2026-11-16)
- RunPod Docs: Serverless pricing. https://docs.runpod.io/serverless/pricing (checked 2026-11-16)
- RunPod: Pricing. https://www.runpod.io/pricing (checked 2026-11-16)
- CoreWeave: AI data centers. https://www.coreweave.com/ai-data-centers (checked 2026-11-16)
- RunPod Docs: Pod pricing. https://docs.runpod.io/pods/pricing (checked 2026-11-16)
- RunPod Docs: Serverless overview. https://docs.runpod.io/serverless/overview (checked 2026-11-16)
- RunPod Docs: Network volumes. https://docs.runpod.io/storage/network-volumes (checked 2026-11-16)
Your next deploy
goes everywhere.
Connect a provider, point a service at it, and ship. Add the second provider when you're ready.
No card needed · Free tier · Pay as you go after that