Simple, transparent pricing

Pay only for what you use. No subscriptions, no hidden fees, no minimum commitment.

RTX 4090

$1.98
per hour
VRAM24 GB
CUDA Cores16,384
Per Second$0.00055
Perfect for ML training, inference, and rendering workloads

L40S

$4.00
per hour
VRAM48 GB
CUDA Cores18,176
Per Second$0.00111
Enterprise-grade performance for demanding AI workloads

A100

$2.34
per hour
VRAM40 GB
CUDA Cores6,912
Per Second$0.00065
Tensor Core acceleration for AI training and HPC

H100

$8.00
per hour
VRAM80 GB
CUDA Cores14,592
Per Second$0.00222
Next-generation performance for large-scale AI models

What's included

Per-second billing

Pay only for the exact time you use. Stop your instance and stop paying immediately.

Instant deployment

Cold starts in under 60 seconds. Your custom Docker images supported.

Full SSH access

Direct SSH access to your instance with root permissions. Install anything you need.

Port forwarding

Expose any port to the internet. Perfect for APIs, Jupyter notebooks, and more.

Crypto payments

Pay with Ethereum, Base, Polygon, Arbitrum, Solana, or Bitcoin via x402 protocol.

No minimums

No hourly minimums, no monthly fees, no subscription required. Pure pay-as-you-go.

Ready to launch?

Get your GPU instance running in under a minute