How AWS Billing Works: Invoices, Pricing, Budgets and Cost Explorer

Simple explanation

Think of AWS billing like a running tab at a restaurant. Every time you order something (launch an instance, store a file, send data), the kitchen adds a line to your receipt. During the meal you can glance at the running total, but tips, taxes, and any promotional discounts are only applied when you ask for the final check. That final check is your AWS invoice.

The key difference from a fixed subscription is that your bill changes every month based on actual usage. Run more servers, store more data, or transfer more traffic and the bill goes up. Shut things down and it goes down. There is no minimum spend and no contract required for standard on-demand usage.

How AWS billing works step by step

1. You create or use AWS resources. Launching an EC2 instance, uploading files to S3, invoking a Lambda function, or sending data through a load balancer all generate usage.
2. AWS meters usage by service. Each service tracks consumption in its own units: seconds of compute, gigabytes of storage, number of requests, gigabytes of data transferred.
3. Estimated charges appear in the Billing console. Throughout the month, you can check the Bills page to see estimated costs broken down by service. These numbers update regularly but are not final.
4. The billing period closes. At the end of the calendar month, AWS stops accumulating usage for that period. Tiered pricing, volume discounts, credits, and taxes are applied.
5. AWS issues the invoice. Within the first few days of the next month, the finalized invoice appears in your account. AWS charges your default payment method.
6. One-time purchases are handled separately. Upfront Reserved Instance payments, AWS Marketplace purchases, and domain registrations can be billed immediately rather than waiting for the monthly cycle.

What actually creates charges in AWS?

A single AWS service can generate multiple line items on your bill. An RDS database, for example, shows up as instance hours, storage, I/O, and backup charges all on separate lines. Knowing which categories of usage cost money helps you predict and control spend. For a deeper dive into how each pricing model works, see the AWS Pricing Models guide.

Compute

Running EC2 instances is billed per second (minimum 60 seconds). Lambda is billed per request plus per GB-second of memory used. Fargate charges for vCPU and memory per second. The instance type, region, and operating system all affect the rate.

Storage

S3 charges per GB stored per month. EBS volumes charge per GB provisioned per month, whether or not the data fills the volume. RDS storage is billed the same way. Snapshots and backups stored in S3 add to the total.

Requests and API calls

S3 charges per PUT, GET, LIST, and other request types. API Gateway charges per million API calls. DynamoDB charges per read and write request unit. These per-request costs are small individually but add up at scale.

Data transfer and egress

Inbound data transfer to AWS is free. Outbound data transfer to the internet (egress) is charged per GB, with tiered pricing that decreases at higher volumes. Cross-region transfers also incur charges. This is one of the most common sources of surprise costs. See Network Egress Costs Explained for the full breakdown.

Managed service compound pricing

Services like RDS, ElastiCache, and Redshift combine multiple billing dimensions: instance hours, storage, I/O operations, and backup storage. A NAT Gateway charges both an hourly rate and a per-GB data processing fee, which catches many beginners off guard.

Bills vs Cost Explorer vs Budgets vs Pricing Calculator

AWS gives you four main tools for understanding and controlling costs. Think of them like this: Bills is your receipt, Cost Explorer is your spending diary, Budgets is your alarm clock, and Pricing Calculator is your planning spreadsheet. They serve different purposes, and confusing them is one of the most common beginner mistakes.

For a hands-on walkthrough of Cost Explorer filtering and report creation, see Cost Explorer Reports. To set up budget alerts step by step, see Setting Budgets and Billing Alerts.

When to use this knowledge

• Setting up your first AWS account. Create a budget alert immediately so you are never caught off guard. Even a $5 alert is valuable when you are learning.
• Before launching new services. Use the AWS cost estimation guide to predict monthly cost before you deploy.
• Debugging an unexpected bill. Check the Bills page for the service-level breakdown, then use Cost Explorer to find when the spike happened and which resources caused it.
• Setting up team cost ownership. Activate cost allocation tags and filter Cost Explorer by team or project.
• Comparing estimated vs final charges. If your mid-month estimates do not match the final invoice, check for credits, tax adjustments, or tiered pricing recalculations.

Savings options

On-demand pricing is the default, but AWS offers significant discounts if you can commit to usage or tolerate interruptions. Think of it like phone plans: on-demand is pay-as-you-go with no contract, Reserved Instances are a fixed plan with a discount, Savings Plans are a flexible contract, and Spot Instances are a standby ticket that is cheap but can be cancelled. For a full comparison and decision framework, see Cost Optimisation Strategies.

Reserved Instances

Commit to a specific EC2 instance type in a region for one or three years. Discounts reach up to 72% compared to on-demand. Payment options are all upfront, partial upfront, or no upfront. Best for steady-state workloads where you know the instance type and region will not change.

Savings Plans

Commit to a minimum dollar-per-hour spend on compute for one or three years. AWS applies the discount automatically across EC2, Lambda, and Fargate. More flexible than Reserved Instances because they work across instance families, sizes, regions, and operating systems. Best for teams that want commitment savings without locking into a single instance type.

Spot Instances

Use spare EC2 capacity at up to 90% off on-demand prices. AWS can reclaim Spot Instances with a two-minute warning when capacity is needed. Best for batch jobs, data pipelines, CI/CD builds, and any workload that can handle interruptions gracefully.

Consolidated billing with AWS Organizations

When you use AWS Organizations, all member accounts’ charges roll up into a single invoice on the management account. This simplifies payment and unlocks volume-based discounts calculated across all accounts combined.

What changes with consolidated billing

• One payment method covers all member accounts.
• Volume pricing tiers and Reserved Instance benefits are shared across accounts.
• The management account receives and pays the combined invoice.

What stays the same

• Each member account still generates its own usage data.
• You can filter Cost Explorer by individual account to see per-account spend.
• Cost allocation tags work across all accounts in the organization.
• Member account owners can still view their own costs if given permission.

Real-world example: why a simple app costs more than expected

Suppose you deploy a small web application using:

• One t3.small EC2 instance running 24/7
• An S3 bucket storing 10 GB of images
• A NAT Gateway so your instance in a private subnet can reach the internet
• CloudWatch basic monitoring (included free)

You might estimate the bill at roughly the EC2 cost alone. But the actual invoice includes:

• EC2 instance hours (the cost you expected)
• EBS volume storage for the instance’s root disk
• S3 storage plus GET/PUT request charges
• NAT Gateway hourly charge plus per-GB data processing fees
• Data transfer out to the internet for every response served through the NAT Gateway

Quick CLI reference

These commands give you a fast terminal-based view of your costs. Replace the date range with the period you want to check.

# View existing budgets
aws budgets describe-budgets \
  --account-id $(aws sts get-caller-identity --query Account --output text)

# Get last month's cost breakdown by service
# Replace START and END with the first and last day of the month you want
aws ce get-cost-and-usage \
  --time-period Start=$(date -d "$(date +%Y-%m-01) -1 month" +%Y-%m-%d),End=$(date +%Y-%m-01) \
  --granularity MONTHLY \
  --metrics "UnblendedCost" \
  --group-by Type=DIMENSION,Key=SERVICE \
  --output table