Overview
This guide walks you through deploying your first GPU server on LightYear Cloud. By the end you will have a running instance you can SSH into and start using immediately.
Prerequisites
- A LightYear Cloud account (sign up at lightyear.host)
- A funded wallet (minimum $5.00)
- An SSH key pair (see Generating an SSH Key Pair)
Step 1 — Fund Your Wallet
LightYear Cloud uses a prepaid wallet model. Navigate to Billing → Add Funds and deposit at least $5.00 using a credit card or cryptocurrency.
Step 2 — Navigate to Deploy
Click Deploy Server in the top navigation bar or go directly to /deploy.
Step 3 — Choose a Region
Select the region closest to your end users or workload. For GPU workloads, Hong Kong (HKG) and Singapore (SGP) offer the lowest latency to Asia-Pacific.
| Region | Code | GPU Availability |
|---|---|---|
| Hong Kong | HKG | ✅ A100, H100 |
| Singapore | SGP | ✅ A100 |
| Los Angeles | LAX | ✅ A100, H100 |
Step 4 — Select a Plan
Choose a GPU plan from the Cloud GPU category. Plans are billed per hour and stopped instances still incur storage charges.
NVIDIA A100 80GB — $2.49/hr
NVIDIA H100 80GB — $3.99/hr
NVIDIA RTX 4090 — $0.99/hr
Step 5 — Configure Your Server
- Operating System — Choose Ubuntu 22.04 LTS (recommended) or any supported image.
- SSH Key — Select an existing key or paste a new public key.
- Label — Give your server a recognisable name, e.g.
ml-training-01. - Backups — Enable automated daily backups (+20% of plan cost).
Step 6 — Deploy
Click Deploy Now. Provisioning typically takes 60–90 seconds. You will receive an email confirmation once the server is active.
Step 7 — Connect via SSH
$ssh -i ~/.ssh/id_rsa root@<YOUR_SERVER_IP>Replace <YOUR_SERVER_IP> with the IP shown in the server detail page.
Verify GPU Access
$nvidia-smiExpected output:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 535.104.05 Driver Version: 535.104.05 CUDA Version: 12.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 NVIDIA A100-SXM... Off | 00000000:00:04.0 Off | 0 |
| N/A 34C P0 57W / 400W | 0MiB / 81920MiB | 0% Default |
+-----------------------------------------------------------------------------+