NVIDIA GPU server hosted in Canada
NVIDIA A100 PCIe in the cloud, in Canada
Rent NVIDIA A100 PCIe servers (80 GB HBM2e per GPU) hosted in Canada: 1, 2, 4 or 8 GPUs per server, from CA$2.11/h, no commitment.
GPU specifications
- Memory per GPU
- 80 GB HBM2e
- CUDA cores
- 6,912
- Tensor cores
- 432
- Interconnect
- PCIe Gen4
- Power (TDP)
- 300W
Sizes and prices
Price per server (all its GPUs) in Canadian dollars, taxes extra. Hourly billing, no commitment.
| Size | vCPU | RAM | Storage | Hourly price | Spot (interruptible) | Order |
|---|---|---|---|---|---|---|
| 1 × GPU | 28 | 120 GB | 100 GB + 750 GB local | $2.11/h | $1.69/h | |
| 2 × GPU | 60 | 240 GB | 100 GB + 1.5 TB local | $4.21/h | $3.37/h | |
| 4 × GPU | 124 | 480 GB | 100 GB + 3.2 TB local | $8.40/h | $6.72/h | |
| 8 × GPU | 252 | 960 GB | 100 GB + 6.5 TB local | $16.78/h | $13.43/h |
Spot machines may be interrupted without notice and are billed hourly only.
One-click ready machines
Your GPU application, already installed
Pick a template when you deploy: the server arrives with the application installed and protected by credentials generated for you.
JupyterLab with PyTorch (CUDA)
JupyterLab notebooks with PyTorch and CUDA, protected by a token generated for you.
Deploy with this templateOllama and Open WebUI
Chat with open models in your browser. The admin account is created for you before access opens.
Deploy with this templateOpenAI compatible vLLM server
OpenAI compatible API (/v1) serving the model you choose, protected by a generated API key.
Deploy with this templateComfyUI
Node based image generation, behind an access protected by a generated username and password.
Deploy with this templatePyTorch environment (CUDA)
PyTorch and CUDA ready in /opt/pytorch, activated at SSH login. No port opened besides SSH.
Deploy with this template
No interface is exposed without authentication. Automatic installation at first boot, usually a few minutes (indicative, not guaranteed).
Typical use cases
- Model training and fine-tuning
- High-performance LLM inference
- Scientific computing (HPC)
Frequently asked questions
How much does a A100 PCIe server cost?
From CA$2.11/h for a 1-GPU server, billed hourly with no commitment, in Canadian dollars, taxes extra. The price of every size is shown in the sizes table.
How many A100 PCIe GPUs can a server have?
1, 2, 4 or 8 GPUs per server. Each size has its own vCPUs, RAM and storage, listed in the table.
How much memory does a A100 PCIe GPU have?
Each A100 PCIe GPU has 80 GB HBM2e.
What is a A100 PCIe Spot machine?
A Spot machine may be interrupted without notice and is billed hourly only. It is available from CA$1.69/h for 1 GPU.
Where are the A100 PCIe servers hosted?
In Canada. Prices and billing are in Canadian dollars (CAD), with the applicable Canadian taxes.
Guides to go further
Other GPU models
Indicative prices, reviewed every hour according to the market and the exchange rate.