The industry's best economics.
Period.

Our compute is priced appropriately, at an average of ⅓ less than our competitors. We don’t hand out credits and ask our paying clients to foot the bill.

CoreWeave’s infrastructure is built with your workloads and business in mind, ensuring that you maximize performance while optimizing your cloud spend. After adjusting for performance, our clients typically spend 50-80% less with CoreWeave than our competitors.

We commit to true usage based billing, which starts when your applications are ready to serve, not while instances are loading.

Compute Optimized GPU Instances

GPU Type VRAM vCPU RAM (GB) Ephemeral Storage (GB) Network $/GPU/Hour
NVIDIA V100 for NVLINK 16GB HBM2 4 32 450 10 Gbps $1.00
NVIDIA V100 for PCIe 16GB HBM2 3 20 64 1 Gbps $0.60
NVIDIA P100 for NVLINK 16GB HBM2 4 32 450 10 Gbps $0.75
NVIDIA Quadro RTX 6000 24GB GDDR6 8 60 450 10 Gbps $1.35
NVIDIA Quadro RTX 5000 16GB GDDR6 8 60 450 10 Gbps $0.95
NVIDIA Quadro RTX 4000 8GB GDDR6 3 16 125 1 Gbps $0.35

*additional CPU cores with 8GB of RAM can be added to any GPU for $0.07/hr

Virtual Workstation Optimized Instances

Virtual Workstation VRAM vCPU/GPU RAM (GB)/GPU Ephemeral Storage (GB) Network Instance Price
NVIDIA Quadro RTX 6000 24GB GDDR6 8 60 450 10 Gbps $1.85
NVIDIA Quadro RTX 5000 16GB GDDR6 8 60 450 10 Gbps $1.45
NVIDIA Quadro RTX 4000 8GB GDDR6 5 30 125 2.5 Gbps $0.95

*includes hourly, usage-based Teradici license cost

Compute Optimized CPU Instances

CPU Type RAM per vCPU (GB) Instance Storage (GB)/vCPU Network $/vCPU/Hour
AMD Epyc 4 10 10 Gbps $0.01
Intel Xeon 3 10 10 Gbps $0.005

Peripheral Compute

Offering Rate
Block Storage (HDD) $0.02 per GB
Block Storage (SSD) $0.07 per GB
Bandwidth $0.04 per GB

Ready to get started?

Our support engineers are here to help!