CoreWeave Cloud Pricing

CoreWeave's pricing is designed for flexibility. Instances are highly configurable, giving you the freedom to customize GPU, CPU, RAM, and storage requests when scheduling your workloads.

Our 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.

CoreWeave GPU Cloud Pricing

CoreWeave Cloud GPU instance pricing is highly flexible, and meant to provide you with ultimate control over configuration and cost. Pricing below is a la carte, where the total instance cost is a combination of a GPU component, the number of vCPU, and the amount of RAM allocated. To keep things simple, CPU and RAM cost are the same per base unit, and the only variable is the GPU chosen for your workload or Virtual Server.

A valid GPU instance configuration must include at least 1 GPU, at least 1 vCPU and at least 2GB of RAM. When deploying a Virtual Server, the GPU instance configuration must also include at least 40GB of root disk NVMe tier storage.

GPU Model
VRAM (GB)
Max vCPUs per GPU
($0.01/hr)
Max RAM (GB) per GPU
($0.005/hr)
GPU Component
Cost Per Hour
80
48
256
$2.21
80
48
256
$2.21
40
48
256
$2.06
40
48
256
$2.06
48
48
256
$1.28
48
48
256
$1.28
24
36
128
$0.77
16
36
128
$0.61
16
36
128
$0.57
8
36
128
$0.24
16
36
128
$0.80
A100 80GB NVLINK
$2.21 / Hour
80 VRAM
48 Max CPUs
256 Max RAM
A100 80GB PCIe
$2.21 / Hour
80 VRAM
48 Max CPUs
256 Max RAM
A100 40GB NVLINK
$2.06 / Hour
40 VRAM
48 Max CPUs
256 Max RAM
A100 40GB PCIe
$2.06 / Hour
40 VRAM
48 Max CPUs
256 Max RAM
A40
$1.28 / Hour
48 VRAM
48 Max CPUs
256 Max RAM
RTX A6000
$1.28 / Hour
48 VRAM
48 Max CPUs
256 Max RAM
RTX A5000
$0.77 / Hour
24 VRAM
36 Max CPUs
128 Max RAM
RTX A4000
$0.61 / Hour
16 VRAM
36 Max CPUs
128 Max RAM
Quadro RTX 5000
$0.57 / Hour
16 VRAM
36 Max CPUs
128 Max RAM
Quadro RTX 4000
$0.24 / Hour
8 VRAM
36 Max CPUs
128 Max RAM
Tesla V100 NVLINK
$0.80 / Hour
16 VRAM
36 Max CPUs
128 Max RAM
*compute instances on CoreWeave Cloud are configurable. For more info, please refer to our Resource Based Pricing Documentation.

CoreWeave CPU Cloud Pricing

CPU only instance pricing is simplified and is driven by the cost per vCPU requested. Pricing scales linearly with the vCPU count, and RAM is included in the per vCPU price.

CPU Model
RAM per vCPU
Cost Per vCPU
AMD EPYC Milan
$0.035 / Hour
4 RAM GB
AMD EPYC Rome
$0.03 / Hour
4 RAM GB
4 RAM GB
Intel Xeon v4
$0.02 / Hour
4 RAM GB
Intel Xeon v3
$0.0125 / Hour
4 RAM GB
*compute instances on CoreWeave Cloud are configurable. For more info, please refer to our Resource Based Pricing Documentation.

Storage

CoreWeave Cloud’s high-performance network attached storage is priced to be simple to understand and budget, with no limitations on scale, throughput or IOPS. Without extraneous charges for transferring data that drive up costs on other providers, our clients find storage pricing to be straightforward and transparent.

Offering
Rate
Block Storage (HDD)
$0.04 per GB / month
Block Storage (NVMe)
$0.07 per GB / month
Object Storage
$0.03 per GB / month
Block Storage (HDD)
$0.04 per GB / month
Block Storage (NVMe)
$0.07 per GB / month

Networking

Networking on CoreWeave is built to empower scale computational workloads. We don’t charge for ingress and egress, and if your workload requires significant throughput we will work with our partners to structure IP transit or peering agreements that work for your business.

Offering
Rate
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
$4.00 per IP / month
Burstable IP Address
$10.00 per IP / month
VPC
$20.00 per IP / month

GPU Cloud Pricing FAQ

How is CoreWeave up to 80% less expensive?

At CoreWeave, we focus on delivering the best performance-adjusted cost in the market. To do this, we offer the lowest on-demand prices of any large-scale cloud provider, the industry’s broadest range of NVIDIA GPUs so you can always find the best GPU for your workload, the fastest & most flexible infrastructure for performance and efficiency, and exceptional support to ensure that your infrastructure is optimized for performance and cost.

Do you offer discounts?

The vast majority of our clients enjoy the benefits of CoreWeave’s on-demand pricing with no contract or commitment. For workloads with relatively predictable usage patterns, we offer two types of volumetric discounts: Reserved Instances (up to 60% for 24/7 committed usage) and Bulk Credits (up to 25% for large-scale, burst compute use cases consumed on-demand). Reach out to our Sales team to learn more about how these discounts can help your business.

Do you charge for bandwidth/data transfer?

No! Gross. Other cloud providers “tax” you for each GB you transfer, making the cost to run at scale in the cloud prohibitive. At CoreWeave, we don’t charge for things like region-to-region transfers, workstation data, or egress in the vast majority of use cases. If your bandwidth needs are extraordinary, we’ll work to understand your needs before negotiating flat bandwidth rates with our suppliers on your behalf.

Do you offer “spot” or interruptible instances?

No, we do not. You can think of our on-demand pricing as “interruptible pricing for uninterruptible instances”. Our clients benefit from economics that allow them to scale, with peace of mind knowing that when they spin-up instances on CoreWeave Cloud, it’s theirs until they choose to spin it down once their workloads are complete.

Does CoreWeave have a free credits program?

We don’t have a standard free credits program. Other cloud providers hand out a ton of free credits, and subsidize the cost by charging clients extremely high prices once those credits run out. By asking all of our clients to pay an appropriate cost for compute, we’re able to deliver economics that scale for years to come.

CoreWeave Cloud
Pricing FAQs


How is CoreWeave up to 80% less expensive?

At CoreWeave, we focus on delivering the best performance-adjusted cost in the market. To do this, we offer the lowest on-demand prices of any large-scale cloud provider, the industry’s broadest range of NVIDIA GPUs so you can always find the best GPU for your workload, the fastest & most flexible infrastructure for performance and efficiency, and exceptional support to ensure that your infrastructure is optimized for performance and cost.

Do you offer discounts?

The vast majority of our clients enjoy the benefits of CoreWeave’s on-demand pricing with no contract or commitment. For workloads with relatively predictable usage patterns, we offer two types of volumetric discounts: Reserved Instances (up to 60% for 24/7 committed usage) and Bulk Credits (up to 25% for large-scale, burst compute use cases consumed on-demand). Reach out to our Sales team to learn more about how these discounts can help your business.

Do you charge for bandwidth/data transfer?

No! Gross. Other cloud providers “tax” you for each GB you transfer, making the cost to run at scale in the cloud prohibitive. At CoreWeave, we don’t charge for things like region-to-region transfers, workstation data, or egress in the vast majority of use cases. If your bandwidth needs are extraordinary, we’ll work to understand your needs before negotiating flat bandwidth rates with our suppliers on your behalf.

Do you offer “spot” or interruptible instances?

No, we do not. You can think of our on-demand pricing as “interruptible pricing for uninterruptible instances”. Our clients benefit from economics that allow them to scale, with peace of mind knowing that when they spin-up instances on CoreWeave Cloud, it’s theirs until they choose to spin it down once their workloads are complete.

Does CoreWeave have a free credits program?

We don’t have a standard free credits program. Other cloud providers hand out a ton of free credits, and subsidize the cost by charging clients extremely high prices once those credits run out. By asking all of our clients to pay an appropriate cost for compute, we’re able to deliver economics that scale for years to come.

recent blog posts
Introducing LGA1: CoreWeave’s Cutting Edge Data Center in NYC
Read more →
The Evolution of Large Language Models and How They Stack Up
Read more →
Best NVIDIA GPUs for Serving Inference on CoreWeave
Read more →

Ready to get started?