How to use the AWS Cost Estimator
- Enter the instance hourly rate and how many hours a month it runs.
- Enter storage volume and the per-GB-month rate.
- Enter estimated outbound data transfer and its per-GB rate.
- Add a margin for requests, support plans and cross-zone traffic.
How the calculation works
Compute is billed per instance-hour, and a full month is 730 hours on average — the figure providers use rather than 720. Storage is billed per GB-month on provisioned or stored volume, and outbound transfer is billed per GB with inbound almost always free. Those three lines account for most small-to-medium bills.
What they miss is the long tail: per-request charges on object storage and serverless functions, cross-availability-zone traffic between your own services, NAT gateway data processing, snapshot storage, load balancer hours and support plan percentages. Regional price differences add another 10 to 30% spread for identical resources, and commitment discounts — reserved or savings plans — can cut compute substantially in exchange for a one to three year term.
Monthly ≈ (instance rate × hours) + (GB stored × storage rate) + (GB egress × transfer rate) + request and support chargesSource: Standard cloud pricing structure; 730-hour average month convention. VERIFY current per-region rates against the provider's published price list.
Worked example
One instance at 0.096/hour running continuously, 320 GB of block storage at 0.08/GB-month, 900 GB egress at 0.09/GB.
- Compute = 0.096 × 730 = 70.08.
- Storage = 320 × 0.08 = 25.60.
- Egress = 900 × 0.09 = 81.00.
- Subtotal = 176.68, plus roughly 15% for requests, snapshots and load balancing.
About 203 a month, with egress the single largest line — a CDN in front of the service would likely cut it substantially.
Frequently asked questions
Why is data transfer such a large part of the bill?+
Outbound transfer is charged per gigabyte and scales with traffic rather than infrastructure, so a popular site pays more without adding servers. Caching at a CDN edge is usually the cheapest fix.
How much can reserved instances or savings plans save?+
Commitments of one to three years typically cut compute rates substantially, in exchange for locking in usage. They only pay off on genuinely steady workloads, not on ones you might re-architect.
Do prices differ between regions?+
Yes, often by 10 to 30% for the same resource. Latency and data residency usually decide the region, but for batch or background work a cheaper region can be a real saving.
What is commonly missing from an estimate like this?+
Per-request charges, cross-zone traffic between your own services, NAT gateway processing, snapshots, load balancer hours and any paid support plan. Treat the estimate as a floor.
Last reviewed August 31, 2026. We review this page whenever the underlying formula, tax year, published rate or standard changes.