Deploy a new GPU virtual machine

The world's cheapest hourly cloud GPUs, pro-rated to the microsecond. Available in dozens of locations worldwide. From P4's to H100s, we have you covered. Contact support if you encounter any issues — we're here to help!

You can also become a host to monetize your own servers.

Don't need a GPU? Check out our CPU-only instances, optimized for batch parallel computing.

1. Choose a graphics card model

AI GPUs for Deep Learning by NVIDIA

Powerful GPUs for the most demanding training workloads. Available in secure data centers. Scroll horizontally for more options.

H100 SXM5 80GB

From $2.52/hour

A100 SXM4 80GB

From $0.98/hour

V100 SXM2 16GB

From $0.18/hour

Professional GPUs for Rendering by NVIDIA

Affordable GPUs for the most demanding rendering and image generation workloads. Scroll horizontally for more options.

RTX 6000 Ada 48GB

From $0.72/hour

L40S 48GB

From $0.82/hour

L40 48GB

From $1.07/hour

A40 PCIE 48GB

From $0.42/hour

L4 24GB

From $0.27/hour

RTX A6000 48GB

From $0.23/hour

RTX 5000 Ada 32GB

From $0.37/hour

Quadro RTX 8000 48GB

From $0.27/hour

RTX A5000 24GB

From $0.22/hour

RTX 4500 Ada 24GB

From $0.33/hour

RTX 4000 Ada 20GB

From $0.18/hour

RTX A4000 16GB

From $0.11/hour

NVIDIA A2 16GB

From $0.12/hour

Consumer-grade GPUs by NVIDIA

Cost-effective, gaming GPUs available in large quantities for distributed computing and cloud gaming. Scroll horizontally for more options.

RTX 4090 24GB

From $0.32/hour

RTX 3090 24GB

From $0.21/hour

RTX 4080 Super 16GB

From $0.22/hour

RTX 4080 16GB

From $0.22/hour

RTX 3080 Ti 12GB

From $0.23/hour

RTX 3080 10GB

From $0.15/hour

2. GPU Count

3. RAM (GB)

4. CPU Count

vCPUs

5. NVMe SSD

6. Select a Location

Available servers

With at least the requested resources

Alternative in-stock GPUs

Scroll to view more...

7. Select an operating system

Operating system templates updated on July 7, 2023.

Ubuntu + TensorFlow ML / AI

Jupyter, TensorFlow, Keras, CUDA, 30GB size
Requires 40GB+ disk

Included

Ubuntu + PyTorch ML / AI

Jupyter, PyTorch, CUDA, 30GB size
Requires 40GB+ disk

Included

Ubuntu + Everything ML / AI

Jupyter, TensorFlow, PyTorch, Keras, fastai, CUDA, 53GB size
Requires 60GB+ disk

Included

Ubuntu 20.04 LTS

NVIDIA drivers preinstalled, 5GB size
Requires 20GB+ disk

Included

Ubuntu 22.04 LTS

NVIDIA drivers preinstalled, 5GB size
Requires 20GB+ disk

Included

Windows 10

NVIDIA drivers preinstalled, 30GB size
Requires 90GB+ disk

Bring Your Own License

8. Configure details

Administrator username

Administrator password

At least 8 characters long and contains at least 1 uppercase letter, symbol, or number. If you've set an SSH public key in your organization settings, leave this field blank to use SSH authentication. You can set an SSH public key here.

Name

Configure port forwards

You may forward up to 0 ports. The external port is where requests will enter; the internal port is where you set the requests to be forwarded to. Please remember to include an SSH port (a port forwarded to port 22) or an RDP port (a port forwarded to port 3389) so that you will be able to access your instance once created.

External port

Internal port

Click for more advanced options

Cloudinit Script

Interruptible instance (documentation)

/hr

Apply Bid

9. Payment options

Choose between on-demand and subscription billing.

Payment type

Your Server

Pricing details will appear once you select a
location.

Email us at sales@tensordock.com if you are
interested in saving 40% with a long-term contract

Share Your Server

Share your server's specs using a custom link.
Pressing the button auto-copies the link.

Generate and Copy Link