4. Actions Taken When Credit Is Insufficient
If your credit balance runs out and you do not top up in a timely manner, the system will begin taking actions to reclaim unused resources:
Immediate Action Upon Credit Depletion:
Compute Resources: All active compute resources (e.g., notebook instances, training jobs, inference endpoints) will be automatically delete to prevent further usage charges.
Email Notification: You will receive an email alert notifying you that your compute resources have been stopped due to insufficient credit.
Action After 3 Days of Inactivity:
If no credit top-up has been made within 3 calendar days of the compute resource shutdown, the system will proceed to delete your network volumes.
Final Email Notification: Before deletion, a final email notification will be sent to inform you that your network volumes will be permanently deleted if no credit is added.
Important: Deleted network volumes cannot be recovered. Make sure to add sufficient credit before the 3-day window expires if you wish to retain your data.
B. Usage Reporting
Access to the usage report portal through the url: https://aiplatform.console.greennode.ai/usage-report
The Resource Usage Report is divided into four main views:
1. Summary View
- Total Estimated Cost: Displays the total estimated cost for the selected time period, broken down by resource type (GPU, CPU, and Storage).
- Resource Usage Breakdown: Provides a visual representation of the resource usage distribution across different resource types.
- Cost Breakdown: Shows a detailed breakdown of costs based on resource type and usage.
2. GPU View
- GPU Usage: Displays the total estimated GPU usage for the selected time period.
- GPU Utilization: Visualizes the utilization of individual GPU cards over time.
- Cost Breakdown: Shows the cost breakdown for each GPU card.
3. CPU View
- CPU Usage: Displays the total estimated CPU usage for the selected time period.
- Instance Usage: Visualizes the usage of different CPU instance types.
- Cost Breakdown: Shows the cost breakdown for different CPU instance types.
4. Storage View
- Storage Usage: Displays the total estimated storage usage for the selected time period.
- Storage Volume Usage: Visualizes the usage of different storage volumes types.
- Cost Breakdown: Shows the cost breakdown based on storage volume size and usage.
Querying the Report: The Resource Usage Report allows you to query usage data for specific time periods:
- Daily View: View detailed usage information for the last ten days.
- Hourly View: View hourly usage data for the last 72 hours.
- Monthly View: View monthly usage data for last 12 months
Tips for Effective Usage
- Regularly Monitor Usage: Keep track of resource consumption to identify trends and potential optimization opportunities.
- Optimize Resource Allocation: Adjust resource allocations based on usage patterns to maximize efficiency and minimize costs.
- Utilize Cost Optimization Tools: Leverage the platform's cost optimization tools to identify cost-saving measures.
- Consider Reserved Instances: Evaluate the use of reserved instances for consistent workloads to reduce costs.
- Monitor Performance Metrics: Track performance metrics like CPU utilization, memory usage, and network throughput to identify bottlenecks and optimize resource allocation.
By effectively utilizing the Resource Usage Report, you can gain valuable insights into your resource consumption and make informed decisions to optimize your costs and performance.

Please be aware that the data displayed in the Resource Usage Report is updated every 30 minutes. As a result, the estimated costs provided in the report may differ slightly from the final billing statement at the end of the month.
C. Billing Calculations
1. Compute Instances
Formula
- Billing Unit: Hour of usage
- Calculation:
- Usage time: Calculated from the moment you Start the instance until you Stop it.
- Duration: Calculated in hours (up to 8 decimal places, not rounded).
- Cost: Duration (hours) * Item price from the hourly pricing table.
- Amount on the bill without VAT: Up to 2 decimal places, not rounded.
Let's look at some specific examples:
a. Notebook Instance:
Example: You use an g5-standard-16x250-1h100 notebook instance priced at 0.1 USD/hour
for 2 hours and 35 minutes.
- Usage time is rounded up to 155 minutes.
- Duration: 155 / 60 = 2.58333333
- Cost:
2.5833333 * 0.1 USD/hour = 0.25833333 USD
- Amount on the bill without VAT:
0.25 USD
b. Training Job:
Example: You run a training job using 2 nodes of type g5-standard-16x250-1h100 at 3.06 USD/hour/node
.
- Node 1 runs for 1 hour and 20 minutes, Node 2 runs for 1 hour and 45 minutes.
- Total Duration:
(80 minutes + 105 minutes) / 60 = 3.08333333 hours
- Cost:
3.0833333 * 3.06 USD/hour = 9.43499998 USD
- Amount on the bill without VAT:
9.43 USD
c. Inferencing:
Example: You create an endpoint using 1 node of type g5-standard-16x250-1h100 priced at 0.1 USD/hour/node
. The endpoint is active for 5 hours and 12 minutes.
- Duration: 312 minutes / 60 = 5.2 hours
- Cost:
5.2 * 0.1 USD/hour = 0.52 USD
- Amount on the bill without VAT:
0.52 USD
Notes:- The above examples are for illustrative purposes only. Actual prices may vary depending on the instance type, availability zone, and other factors.
- Always refer to the detailed Greennode AI Platform pricing table .
2. Network Volume
Formula
Specific Example for Network Volume:
- You create a 100GB network volume priced at 0.10 USD/GB/month for 10 hours.
- Usage Time: 10 hours
- Duration in Months: 10 hours / (30 days/month * 24 hours/day) = 0.01388889 months
- Cost: 100GB * 0.10 USD/GB/month * 0.01388889 months = 0.01388889
- Amount on the bill without VAT: 0.01 USD
- You create a 100GB network volume priced at 0.10 USD/GB/month for 10 hours and update network volume size later
- Assumptions:
- Initial price: 0.10 USD/GB/month
- Initial volume size: 100GB
- Initial usage time: 10 hours
- New volume size: 150GB
- New usage time (since increasing the size): 20 hours
- To calculate the exact cost, we need to divide the usage into 2 phases:
Phase 1: Size with 100GB
Usage time: 10 hours
Cost calculation: As calculated in the previous example, the result is 0.01388889 USD.
- Phase 2: Size with 150GB
- Usage time: 20 hours
- Cost calculation:
- Convert time to months: 20 hours / (30 days/month * 24 hours/day) = 0.02777778 months
- Calculate cost: 150GB * 0.10 USD/GB/month * 0.02777778 months = 0.04166667 USD
- Total cost:
- Total cost: 0.01388889 USD (phase 1) + 0.04166667 USD (phase 2) = 0.05555556 USD
- Amount on the bill without VAT: 0.05 USD (not rounded)
3. VAT (for Singapore-based Entities)
- VAT applicability: A 9% Goods and Services Tax (GST) is levied on customers with legal entities registered in Singapore who use Greennode (LG Entity, Singapore) services.
- Other customers: No GST is applicable.
VAT Calculation Steps:
Determine the customer's jurisdiction:
- Refer to the contract or invoice to identify the country where the customer's business is registered.
- If the business is registered in Singapore, proceed to step 2. Otherwise, no GST is applicable.
Calculate the GST amount:
- Formula: GST amount (USD) = Amount on the bill without VAT (USD) * 9%
Calculate the total amount due:
- Formula: Total amount due (USD) = Total service value (USD) + GST amount (USD)
Example: Assuming a customer with a Singaporean legal entity uses Greennode services with a total service value of 7,000 SGD.
- Amount on the bill without VAT (USD) = 7,000 USD
- GST amount (USD) = 7,000 USD * 9% = 630 USD
- Total bill amount included VAT (USD) = 7,000 USD + 630 USD = 7,630 USD
D. Billing Methodology
1. Pay-as-you-go Model
Flexible: Only pay for actual usage
Description: With the pay-as-you-go billing method, you will be charged based on your actual usage of AI resources. This flexible billing model allows you to scale resources up or down as needed and pay only for what you use.
Best For: Variable workloads, testing. Ideal for users with fluctuating workloads or those who want to test the platform without committing to a fixed plan.
2. Billing Cycle
GreenNode operates on a continuous, usage-based billing cycle. This means you are charged for the resources you consume in near real-time. Here's a breakdown of how it works:
- Continuous Resource Monitoring: The GreenNode platform constantly monitors your active resources, including compute instances (GPUs, CPUs), storage, and other chargeable services. This monitoring occurs continuously, tracking your resource consumption as you use the platform.
- Minute Charge Increments: Every 5 minutes, the platform calculates the cost of your resource usage for that specific 5-minute interval. This calculation is based on the pricing structure for each service, which you can review on the GreenNode AI Platform pricing page.
- Automated Credit Deductions: The calculated usage cost for each 5-minute period is automatically deducted from your GreenNode credit balance. This process is fully automated and requires no manual intervention.
- Real-time Balance Updates: Your credit balance is updated immediately after each 5-minute deduction. You can view your current balance at any time on the Credits page.