GPU guide
Rent an H100 in Canada: PCIe, NVLink or SXM, by the hour
Updated on
Short answer
At GPU Cloud, an NVIDIA H100 PCIe (80 GB) rents by the hour from CA$3.35/h, and 8-GPU H100 servers (H100 NVLink and H100 SXM) go up to CA$34.27/h, taxes extra. The servers are in Montreal, billed in Canadian dollars, and an H100 out of stock shows the closest equivalent machine in stock.
Three H100 versions: PCIe, NVLink and SXM
| Version | Memory per GPU | Link between GPUs | GPUs per server | From, per hour |
|---|---|---|---|---|
| NVIDIA H100 PCIe | 80 GB HBM2e | PCIe Gen5 | 1, 2, 4 and 8 | $3.35 |
| NVIDIA H100 NVLink | 80 GB HBM2e | PCIe Gen5 + NVLink (600 GB/s) | 8 | $27.84 |
| NVIDIA H100 SXM | 80 GB HBM3 | NVLink 4.0 (900 GB/s) | 8 | $34.27 |
Price of the whole server, in Canadian dollars, taxes extra, read from the catalog when the page is built.
The H100 PCIe rents with 1 to 8 GPUs per server. The NVLink and SXM versions link the GPUs of a server with a faster link than PCIe: they target training spread over several GPUs. The H100 SXM and H100 PCIe comparison puts their specs side by side.
Price and availability of every H100 machine
| GPU | GPUs per server | Server GPU memory | On demand, per hour | Availability |
|---|---|---|---|---|
| NVIDIA H100 PCIe | 1 | 80 GB | $3.35 | |
| NVIDIA H100 PCIe | 2 | 160 GB | $6.69 | |
| NVIDIA H100 PCIe | 4 | 320 GB | $13.39 | |
| NVIDIA H100 PCIe | 8 | 640 GB | $26.77 | |
| NVIDIA H100 NVLink | 8 | 640 GB | $27.84 | |
| NVIDIA H100 SXM | 8 | 640 GB | $34.27 |
Prices in Canadian dollars, taxes extra, read from the catalog when the page is built. Availability is read live; a machine out of stock shows the closest equivalent machine in stock.
When an H100 is worth it, and when another card is enough
- The H100 (Hopper architecture, with FP8 support) targets training and fine-tuning of large models, and inference when compute limits more than memory.
- If your job is limited by memory rather than compute, an A100 PCIe offers 80 GB per GPU for CA$1.81/h. See the H100 and A100 comparison.
- For inference of a 7 to 13 billion parameter model, an RTX A6000 with 48 GB at CA$0.67/h is often enough.
If the H100 you want is out of stock
H100s are often out of stock. When a machine of the table is out of stock, the closest equivalent machine in stock is shown first, with its hourly price and a button to deploy it: at least as much GPU memory, the same number of GPUs when possible, then the closest price.
You can also ask for an availability alert, or reserve an H100 for a given date with the number of GPUs and the duration you need.
Cost examples
- An 8-hour work day on 1 H100 PCIe: $26.80, or $21.44 on Spot (interruptible).
- 24 hours on a 8-GPU H100 PCIe server (640 GB of GPU memory): $642.48.
- 24 hours on a 8-GPU H100 NVLink server (640 GB of GPU memory): $668.16.
- 24 hours on a 8-GPU H100 SXM server (640 GB of GPU memory): $822.48.
Formula: hourly rate × number of hours, rounded to the cent, taxes extra.
Good to know before you launch
- Billing is hourly, at the server’s hourly rate, from prepaid credit. There is no per-minute billing, and the first hour is charged when you deploy.
- A stopped server is still billed at the full hourly rate. To pay less, hibernate it ($0.02 per hour (about $14.60 per month), disk kept) or delete it. A Spot machine cannot be hibernated: delete it when you are done.
- The minimum top-up is CA$25.00, paid by credit card.
- Servers and their disks stay in Canada (region canada-montreal, in Montreal). Account information (identity, billing, e-mails) may be processed outside Quebec, as the privacy policy explains.
Frequently asked questions
How much does an H100 cost per hour?
From CA$3.35/h for a 1-GPU H100 PCIe, CA$2.68/h on Spot, taxes extra. 8-GPU H100 servers go up to CA$34.27/h.
Can I rent a single H100?
Yes, the H100 PCIe rents with 1 GPU. The H100 NVLink and H100 SXM versions rent as 8-GPU servers.
What is the difference between the H100 PCIe and the H100 SXM?
Both have 80 GB of memory per GPU. The SXM links its 8 GPUs with NVLink and draws more power; the PCIe rents from 1 GPU, at a lower price.
What if the H100 is out of stock?
Deploy the equivalent machine offered first, in stock, or ask for an availability alert, or reserve an H100 for a given date.
Where are the H100s hosted?
In Montreal, Canada. The servers and their disks stay in Canada, and the invoice is in Canadian dollars.
Need GPUs? We’ve got you.
Launch an NVIDIA GPU server by the hour in Canada, paid in Canadian dollars by credit card, or reserve a GPU for a given date.
Other guides
- Your AI says you need a GPU? Run your Python script on a GPU in minutes
- CUDA out of memory: causes, fixes, and when to move to a GPU with more VRAM
- torch.cuda.is_available() returns False: causes and fixes (Mac, PC without NVIDIA, Docker)
- How much does a GPU cost per hour in Canada, in Canadian dollars?
- OpenAI API too expensive? Host your LLM on a GPU by the hour, in Canada
- Rent a GPU by the hour in Canada: prices, steps and billing
- GPU cloud in Montreal, Canada: NVIDIA GPU servers hosted in the country
- Which GPU to fine-tune an LLM? Memory, machine and cost
- Cheap GPU cloud: how to pay as little as possible for a GPU