This week All our GPUs are sold at cost price — zero margin on the server. See the cost sheets

This week Every GPU sold at cost price

Getting started

Getting started

From an empty account to a root shell on a dedicated GPU server in four steps: sign up, top up in crypto, order, connect.

4 min read · updated 2 September 2026

Four steps, no forms beyond an email address, no approval queue. Most people are logged in to their first server within a quarter of an hour, and most of that is waiting for a blockchain.

1. Create an account

Go to /signup, enter an email address and a password twice. That is the whole form. No name, no phone, no documents, no card. The account opens at once and you land in the console.

Keep your password. We send no email and hold nothing that identifies you, so there is no “forgot password” link and no way for us to reopen an account whose password is lost. Store it in a password manager before you go any further.

2. Top up your balance

Open Billing → Top up, type an amount (minimum $40), pick a coin, and send the exact amount shown to the address on the invoice. The balance is credited when the network confirms; the invoice page refreshes on its own and tells you how many confirmations are needed. USDT on TRON is the cheapest and fastest option; Monero the most private. Details in Balance and top-ups.

How much? The first term of the server you want. A single RTX 4090 for one month is $139; a single H100 SXM is $1,134. Every price is on the catalogue and none has a setup fee.

3. Order the server

Servers → Order walks you through six short steps: the card and how many, the region, the operating system and template (with a model to pre-load if you pick a serving template), storage and network options, the term and your SSH keys, and a review with the exact amount due. The whole term is taken from your balance and provisioning starts. A single-card server is online in under 5 minutes.

4. Connect

The server page shows the public IPv4, the hostname and either your SSH keys or a generated root password. Then:

ssh [email protected]
nvidia-smi          # every card, the whole card
df -h /             # the local NVMe

Templates come up listening on their port (Jupyter on 8888, vLLM on 8000, ComfyUI on 8188…); the server page lists the ports and the first command to run for the template you chose. See Templates and ports.

What happens next

  • The term renews automatically from your balance at the end of the paid period, at the same price. Terms and renewals.
  • Support opens in the console as soon as a server is online.
  • An API token from API gives you the same data as the console, as JSON. API reference.

Try it on a real server.

One month, at cost, root in under 5 minutes. Support opens with the first server.