GPU Cloud Pricing
Get access to world-class cloud resources and the best GPUs on the market.
Our infrastructure is built from the ground up to give AI teams the powerful compute they need to build and deploy cutting-edge GenAI applications.
We aim to provide first-to-market access to the latest and greatest NVIDIA GPUs—accelerating and streamlining your workloads for superior price-to-performance ratios.
Superpower your GenAI workloads today and build on CKS.
Compute
Get models to market faster with the latest and greatest NVIDIA chips.
On-demand GPU instances
Utilize on-demand GPU instances for additional workloads when you’re not looking for long-term capacity commitments. Quickly spin-up burst capacity on-demand—with the flexibility of great on-demand pricing.
On-demand CPU instances
We know AI training and serving workloads don’t exist in a vacuum. Get CPUs on-demand to support GPUs during the overall lifecycle of model training jobs.
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
AMD Genoa (9454)
AMD Genoa (9454)
AMD Genoa (9454)
AMD Genoa (9454)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
AMD Turin (9655P)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Reserved capacity
CoreWeave offers up to 60% discounts over our On-Demand prices for committed usage. To learn more about these options, please reach out to us today.
Storage
CoreWeave’s high-performance storage is priced to be simple to understand and budget, with no limitations on scale, throughput, or IOPS.
Without extraneous charges for egressing data or for IOPS that drive up costs on other providers, our customers find storage pricing to be straightforward and transparent.
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Networking
Networking on CoreWeave is built to empower scale computational workloads.
Plus, if your workload requires significant throughput, we will work with our partners to structure IP transit or peering agreements that work for your business.
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Distributed File Storage
Distributed File Storage
Distributed File Storage
Distributed File Storage
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Managed Kubernetes
Our managed Kubernetes environment is purpose-built for building, training, and deploying AI Applications.
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
Slurm on Kubernetes (SUNK)
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
CoreWeave Kubernetes Service Control Plane
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer between CoreWeave and internet
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Data transfer within CoreWeave
Distributed File Storage
Distributed File Storage
Distributed File Storage
Distributed File Storage
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Egress and Input Output Operations per Second
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Intel Emerald Rapids (8562Y+)
Looking for CoreWeave Classic pricing?


Start building today.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.
- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.
- LabelH100_PCIE
- Networking100Gbps
- GPU Memory80GB HBM2e
- Graphics BusPCIe
- FP32 Performance51 TFlops
- NVIDIA Tensor Cores456
- Max Power Consumption350W
- Cuda Cores14592


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.
- LabelRTX_A4000
- Networking40 Gbps Ethernet
- RTX-OPS37.4T
- VR ReadyYes
- GPU Memory16 GB GDDR6
- Graphics BusPCI Express 4.0 x 16
- NVIDIA RT Cores48
- FP32 Performance19 TFLOPS
- NVIDIA Tensor Cores192
- Max Power Consumption140W
- Cuda Parallel-Processing Cores6


- GPU MemoryThis is some text inside of a div block.
- LabelA100_PCIE_80GB
- Networking40 Gbps Ethernet
- GPU Memory80GB HBM2e
- Graphics BusPCIe
- FP32 Performance19.5 TFLOPS
- NVIDIA Tensor Cores432
- Max Power Consumption250W
- Cuda Parallel-Processing Cores6912


- GPU MemoryThis is some text inside of a div block.
- LabelA100_NVLINK
- Networking100 Gbps Ethernet + 800 Gbps HDR InfiniBand with SHARP
- GPU Memory80GB HBM2e
- Graphics BusSXM4
- FP32 Performance19.5 TFLOPS
- NVIDIA Tensor Cores432
- Max Power Consumption400W
- Cuda Parallel-Processing Cores6912


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.


- GPU MemoryThis is some text inside of a div block.
- LabelQuadro_RTX_4000
- Ethernet10 Gbps Ethernet
- RTX-OPS43T
- VR ReadyYes
- Rays Cast6 Giga Rays/Sec
- GPU Memory8 GB GDDR6
- Graphics BusPCI Express 3.0 x 16
- NVIDIA RT Cores36
- FP32 Performance7.1 TFLOPS
- NVIDIA Tensor Cores288
- Max Power Consumption160W
- Cuda Parallel-Processing Cores2304


- GPU MemoryThis is some text inside of a div block.
- LabelQuadro_RTX_5000
- Ethernet10 Gbps Ethernet
- RTX-OPS62T
- VR ReadyYes
- Rays Cast8 Giga Rays/Sec
- GPU Memory16 GB GDDR6
- Graphics BusPCI Express 3.0 x 16
- NVIDIA RT Cores48
- FP32 Performance11.2 TFLOPS
- NVIDIA Tensor Cores384
- Max Power Consumption265W
- Cuda Parallel-Processing Cores3072


- GPU MemoryThis is some text inside of a div block.
- LabelRTX_A6000
- Networking40 Gbps Ethernet
- RTX-OPS43T
- VR ReadyYes
- GPU Memory48 GB GDDR6
- Graphics BusPCI Express 4.0 x 16
- NVIDIA RT Cores84
- FP32 Performance38TFLOPS
- NVIDIA Tensor Cores336
- Max Power Consumption300W
- Cuda Parallel-Processing Cores10752


- GPU MemoryThis is some text inside of a div block.
- LabelTesla_V100_NVLINK
- Networking10 Gbps Ethernet
- GPU Memory16 GB HBM2
- Graphics BusSXM2
- FP32 Performance15.7 TFLOPS
- NVIDIA Tensor Cores640
- Max Power Consumption300W
- Cuda Parallel-Processing Cores5120


- GPU MemoryThis is some text inside of a div block.
- LabelRTX_A5000
- Networking40 Gbps Ethernet RTX-OPS
- VR ReadyYes
- GPU Memory24 GB GDDR6
- Graphics BusPCI Express 4.0 x 16
- NVIDIA RT Cores64
- FP32 Performance28 TFLOPS
- NVIDIA Tensor Cores256
- Max Power Consumption230W
- Cuda Parallel-Processing Cores8192


- GPU MemoryThis is some text inside of a div block.
- LabelA40
- Networking40 Gbps Ethernet
- RTX-OPS43T
- VR ReadyYes
- GPU Memory48 GB GDDR6
- Graphics BusPCI Express 4.0 x 16
- NVIDIA RT Cores84
- FP32 Performance38TFLOPS
- NVIDIA Tensor Cores336
- Max Power Consumption300W
- Cuda Parallel-Processing Cores10752


- GPU MemoryThis is some text inside of a div block.
- LabelA100_PCIE_40GB
- Networking40 Gbps Ethernet
- GPU Memory40GB HBM2e
- Graphics BusSXM4
- FP32 Performance19.5 TFLOPS
- NVIDIA Tensor Cores432
- Max Power Consumption250W
- Cuda Parallel-Processing Cores6912


- GPU MemoryThis is some text inside of a div block.
- LabelA100_NVLINK
- Networking100 Gbps Ethernet + 400 Gbps HDR InfiniBand
- GPU Memory40GB HBM2e
- Graphics BusSXM4
- FP32 Performance19.5 TFLOPS
- NVIDIA Tensor Cores432
- Max Power Consumption400W
- Cuda Parallel-Processing Cores6912