Feature

The Nano GPU platform enables seamless multi-tenant sharing, delivering efficient and flexible resource packages tailored for AI development, machine learning training, and related services. By leveraging compact, power-efficient GPU hardware, users can also process high-resolution data from sensors or streams with minimal overhead.

Sign Up & Add Credit

• Create an account at https://www.nanogpu.ai/ and verify your email address.
• After logging in, click Order New Services.
• Select your preferred package and container image.
• Complete the checkout process by paying the basic monthly fee.
• Navigate to Billing → Top Up Balance, then use a credit card (via Stripe) or Alipay to add funds.

Manage your GPU Services

• Click Services.
• Select your plan, such as GPU Container – A20 Normal.
• Review important information, including the IP address, port, root password, etc.
• Change the server image, if needed, by clicking Rebuild Server.
• Add the necessary port by clicking + Add Port.

Prepare to Connect

• Use SSH (Secure Shell), a protocol for remotely accessing Linux machines, to connect to the GPU service.
• Open an SSH client program, such as PuTTY.
• Enter the inbound IP address, port, and select SSH as the connection type.
• Click Open and select Yes to confirm the host identity.
• Log in using the username “root” and the root password to access the server.

Storage and Persistence Design

To maximize efficiency and resource flexibility, Nano GPU adopts a dedicated data persistence strategy:

• Persistent /data directory:
  All-important files, model weights, and datasets should be stored under the /data path. This directory is designed with persistence in mind, ensuring that your essential data remains intact even after a task is suspended or restarted.

• Clean and consistent runtime environment:
  System-level environments and externally installed dependencies are automatically refreshed whenever an instance is suspended and resumed. This ensures every session starts with a clean, standardized environment—free from accumulated conflicts or hidden errors.

With this design, you always benefit from a reliable, clean GPU runtime while keeping your critical data safe and persistent in /data.

Important Note

• The minimum deposit amount is USD$10.
• When your account balance falls below USD$1, your running instances will automatically stop. To prevent this, enable auto top-up on the Billing page. Set an auto-charge threshold higher than your average daily spend to ensure your card is automatically charged when your balance falls below that amount.

Other Questions Relating to Billing

Can you bill my card automatically so I don’t have to add credit in advance?

Part 1 – Monthly Fee

=> The monthly fee is automatically billed each month. This fee covers the basic cost of maintaining GPU server operations.

Part 2 – Pay-As-You-Go Fee

You can configure automatic billing by setting a balance threshold to keep your account balance above a specified amount. We recommend setting the threshold based on your estimated daily or weekly spending. You’ll receive an email notification if your credit balance falls below US$10, giving you time to top up before it reaches zero. Of course, you can also manually top up your balance by adding any desired amount.

NOTE: The currency is in USD.

How can I discontinue the service?

Your card will be charged at least the monthly fee, as active instances incur storage and bandwidth charges. To avoid ongoing charges, ensure you terminate all instances when they are no longer needed. If instances are not terminated, your card will continue to be charged periodically.

How does pricing work for the pay-as-you-go fee?

You are charged a base rate for active GPU usage, calculated per second while your instance is in an active or connected state. GPU usage prices may vary, so please check the current rates regularly. You are not charged for instances after they are terminated.

What is the billing frequency?

Account balances are updated approximately once per day.

Why should I trust LayerStack with my credit card information?

We do not store or process your credit card number or CVC/CVV code. All payment information is securely handled directly by Stripe, a trusted multinational financial services and software-as-a-service (SaaS) company. You can verify this in the JavaScript code used for payment processing.

Stripe: https://stripe.com/