GPU Server Pricing:
Rent a GPU for as low as $0.33/hr

Transparent pricing for GPU instances including RTX 4090, RTX 5090, RTX PRO 6000 and more. Use our cloud price calculator to estimate monthly costs for your AI workloads.

Launch ConsoleContact Us

Complete GPU Pricing Chart

Compare GPU server costs across NVIDIA and AMD models. Rent NVIDIA GPUs like RTX 4090, RTX 5090, RTX PRO 6000, H100, and H200 with flexible on-demand or reserved pricing. Find the cheapest GPU rental for your AI workloads.

  • RTX 4090

    NVIDIA
    On‑Demand$0.39
    1 mo$0.35
    3 mo$0.33
    96 GB VRAM500 GB StorageMax 10 GPUs

  • L40S

    NVIDIA
    On‑Demand$0.63
    1 mo$0.57
    3 mo$0.54
    48 GB VRAM500 GB StorageMax 8 GPUs

  • RTX 5090

    NVIDIA
    On‑Demand$0.65
    1 mo$0.59
    3 mo$0.55
    32 GB VRAM1000 GB StorageMax 8 GPUs

  • RTX PRO 6000

    NVIDIA
    On‑Demand$1.29
    1 mo$1.16
    3 mo$1.10
    96 GB VRAM1200 GB StorageMax 8 GPUs

  • H100

    NVIDIA
    On‑Demand—
    1 mo$1.40
    3 mo$1.33
    48 GB VRAM2000 GB StorageMax 8 GPUs

  • H200

    NVIDIA
    On‑Demand—
    1 mo$2.50
    3 mo$2.38
    141 GB VRAM2000 GB StorageMax 8 GPUs

  • MI300X

    AMD
    On‑Demand—
    1 mo$2.50
    3 mo$2.38
    192 GB VRAM1800 GB StorageMax 8 GPUs

  • MI350X

    AMD
    On‑Demand$3.90
    1 mo$3.51
    3 mo$3.32
    288 GB VRAM8000 GB StorageMax 8 GPUs

Prices can fluctuate. Availability varies by demand and region.

GPU Rental Calculator

Calculate GPU rental costs in real-time. Compare on-demand and reserved pricing for RTX 4090, RTX 5090, and RTX PRO 6000 to optimize your cloud GPU budget.

Save with Reservation Plans

Choose from flexible reservation options to optimize costs for AI training on NVIDIA GPUs and other deep learning workloads. 1 week, 1 month, and 3 month reservations can be selected directly in the console. For 1-year reservations with a an even higher discount, contact us to set up a custom plan.

Reach Out

Pricing FAQ

Everything you need to know about CloudRift pricing

Billing is per active runtime and stops when you shut down the instance.
On-demand is pay as you go. Reserved locks capacity and lowers the hourly rate.
Yes, reservations are paid upfront.
Yes. Use reserved nodes for steady workloads and burst with on-demand.
For reservations longer than 3 months and more than 2 nodes (i.e. 16 GPUs), contact us and include your target models, GPU count, and regions.
All prices include compute, storage, networking, and basic support.
For GPU rentals that aren't available in the console (such as NVIDIA H100 and H200), we may still be able to help. Reach out on Discord or via our contact form with your requirements and we'll work with you to find a solution.

Get in Touch

We're here to support your compute and AI needs. Let us know if you're looking to:

  • Find an affordable GPU provider
  • Sell your compute online
  • Manage on-prem infrastructure
  • Build a hybrid cloud solution
  • Optimize your AI deployment
hello@cloudrift.ai
CloudRift Inc., a Delaware corporation
PO Box 1224, Santa Clara, CA 95052, USA
+1 (831) 534-3437
Follow us on X

I'm interested in: