Skip to main content

Frequently Asked Questions (FAQs)

How do I rent a GPU on NovaPod?​

Sign up with email or WebAuthn Passkey, load credits into your personal deposit address using USDT, USDC, SOL, ETH, or BNB, select your desired GPU and pre-built template, and click Launch. Your instance will be online and accessible in under 15 seconds.

Is there a free trial?​

We do not offer a free trial due to GPU abuse prevention, but you can get started with as little as $5.00 in cryptocurrency. Our per-second metering ensures you only pay for the exact seconds your instance is executing.

How am I charged when my instance is paused?​

When you pause a container pod or Root KVM VM, GPU compute fees immediately drop to $0.00. You pay only a minimal standby storage retention fee of $0.00014 per GB per hour (~$0.10/month for 100GB NVMe) to keep your disk, installed packages, and checkpoints intact.

What is the difference between a Docker Pod and a Root KVM VM?​

A Docker Pod is an optimized container sharing the host kernel with direct GPU PCI-passthrough—it boots in under 15 seconds and is ideal for PyTorch notebooks and inference serving. A Root KVM VM is a full virtual machine with root access from minute one, its own kernel, persistent system drive, and static public IPv4—ideal for Docker-in-Docker, k3s, and custom OS modules.

Yes. You can launch 8-GPU Hopper (H100/H200) or Blackwell (B200/B300) nodes connected via high-bandwidth NVLink and Quantum-2 InfiniBand RDMA directly from the dashboard or CLI.