Skip to main content

Resource Based Pricing

Customize your hardware configurations with à la carte pricing

CoreWeave's entire infrastructure is purpose-built for compute-intensive workloads, and everything from our servers to our storage and networking solutions are designed to deliver best-in-class performance that are up to 35x faster and 80% less expensive than generalized public Clouds.

Hardware selection is significantly flexible, allowing for custom configurations of CPU, GPU, RAM, and storage requests when scheduling workloads. Resources are scheduled using provided configurations, granting both savings and simplicity on top of legacy Cloud instance type selection.

Instance pricing

While CoreWeave allows scheduling and provides pricing based on standard instances, all CoreWeave Cloud instances are highly configurable. All billing is à la carte; priced by the hour, and billed by the minute.

All billing is based on the greater of resources requested in an instance. If the instance is configured to be burstable, billing is based on the actual resources consumed during any minute billing window.

NVIDIA GPU instance pricing

The following components are configurable in GPU based instances.

DescriptionLabelCost per hourVRAM
H100 HGX (80GB)H100_NVLINK_80GB$4.7680GB HBM3
H100 for PCIe (80GB)H100_PCIE$4.2580GB HBM2e
A100 HGX (80GB)A100_NVLINK_80GB$2.2180GB HBM2e
A100 HGX (40GB)A100_NVLINK$2.0640GB HBM2e
A100 for PCIe (80GB)A100_PCIE_80GB$2.2180GB HBM2e
A100 for PCIe (40GB)A100_PCIE_40GB$2.0640GB HBM2e
V100 for NVLINKTesla_V100_NVLINK$0.8016GB HBM2
A40A40$1.2848GB GDDR6
RTX A6000RTX_A6000$1.2848GB GDDR6
RTX A5000RTX_A5000$0.7724GB GDDR6
RTX A4000RTX_A4000$0.6116GB GDDR6
Quadro RTX 5000Quadro_RTX_5000$0.5716GB GDDR6
Quadro RTX 4000Quadro_RTX_4000$0.248GB GDDR6

CPU component pricing

The following CPU components are also configurable in GPU based instances. For CPU-only instances, see CPU-only instance pricing.

Node typeLabelCost per hour
AMD Epyc Milanamd-epyc-milan$0.010
AMD Epyc Romeamd-epyc-rome$0.010
Intel Xeon Ice Lakeintel-xeon-icelake$0.010
Intel Xeon Scalableintel-xeon-scalable$0.010
Intel Xeon v4intel-xeon-v4$0.010

CPU-only instance pricing

Instances without a GPU attached are configurable as combinations of vCPU and system RAM.

For these instances, system RAM is included in the vCPU price. Combinations can be configured in multiples of the following.

CPU TypeLabelRAM per vCPUCost per vCPU per Hour
AMD Epyc Milanamd-epyc-milan4GB$0.035
AMD Epyc Romeamd-epyc-rome4GB$0.030
Intel Xeon Ice Lakeintel-xeon-icelake4GB$0.035
Intel Xeon Scalableintel-xeon-scalable4GB$0.030
Intel Xeon v4intel-xeon-v44GB$0.020
Intel Xeon v3intel-xeon-v34GB$0.0125


RAM may be specified as an additional component on its own.

ComponentLabelCost per hour
System RAM per GBmemory$0.005


Storage may be specified as an additional component on its own.

Storage typeLabelPrice per GB per month
HDD Block Storageblock-hdd-<region>$0.04
NVMe Block Storageblock-nvme-<region>$0.07
Standard HDD Shared Filesystemshared-hdd-<region>$0.04
Standard NVMe Shared Filesystemshared-nvme-<region>$0.07
Premium NVMe Shared Filesystemshared-vast-<region>$0.07
Object StorageN/A$0.03


See the Networking documentation to learn more about these options.

Public IP Address$4.00 per IP / month
Burstable IP Address$10.00 per IP / month
VPC$20.00 per VPC / month

Public IP address pricing

IP Addresses are billed at $4.00 per IP per month.

For periods of use less than one month, this charge is prorated by the minute just as all other billing. If a public IP address is assigned to an instance but the instance is not running, billing will continue to accrue for this reserved Public IP Address.

Example configurations

Below is an example configuration provided in an excerpt of a YAML manifest, in which 6 AMD Epyc Rome vCPUs with 24GB of RAM are requested.


In this example, the node type is chosen using an affinity. For more information on using affinities to select hardware, see Node Types and Advanced Label Selectors.

- name: epyc-example
cpu: 6
memory: 24Gi
cpu: 6
memory: 24Gi

- matchExpressions:
- key:
operator: In
- amd-epyc-rome

In this example, the requested resources are:

CPU amountInstance typeUnit cost per hourSystem RAM
6AMD Epyc Rome$0.0324Gi

The total per hour cost of this instance would be $0.03 x 6, equaling $0.18 per hour.

In this example, a hardware configuration of 4 Tesla V100 NVLINK GPUs with 32 Intel Xeon Scalable vCPU and 128GB of RAM is requested:

- name: v100-example
cpu: 32
memory: 128Gi 4
cpu: 32
memory: 128Gi 4

- matchExpressions:
- key:
operator: In
- Tesla_V100_NVLINK

In this example, the requested resources are:

Instance typeInstance countUnit cost per hour
Tesla V100 NVLINK4$0.80
vCPU countvCPU cost per hour
RAM amountRAM cost per hour

The per hour cost of these resources equals ($0.80 x 4) + ($0.01 x 32) + ($0.005 x 128), or $4.16 per hour.

Billing periods

All CoreWeave Cloud billing periods cover the calendar month.

For example, a billing period that begins on 1 January 12:00am UTC ends on 1 February 12:00am UTC.