The AWS Cost and Usage Report, usually called the CUR, is the granular billing export AWS delivers to an S3 bucket, listing every usage line item by hour or day with the account, service, resource identifier, usage type, region, tags, and several different cost columns. It is the source of truth for cost work because it is far more detailed than the console dashboards, which summarize and round. The single most important thing to understand is the difference between its cost measures: unblended cost is what AWS charged on the day, so an upfront commitment payment lands as one large charge, while amortized cost spreads that upfront portion of Savings Plans and Reserved Instances evenly across the term so each day shows its true economic cost. Use amortized cost for trends and unit economics, and unblended cost to reconcile to the invoice. Get that one distinction right and the rest of cost analysis follows.
Here is what the report contains, how to read its cost columns, and how to query it.
What does the Cost and Usage Report contain?
The CUR is a wide table where each row is a slice of usage for a period as short as an hour. Every row carries the linked account that incurred it, the service and the product code, the usage type that names exactly what was billed, the resource identifier where available, the region and availability zone, any cost allocation tags you have activated, and the cost columns. That resolution is the point: the console can tell you that compute spend rose, but only the CUR can tell you which account, which usage type, and which resource drove it, which is what you need to act. For a gentler tour of how a bill is built before diving into the CUR, see the anatomy of a cloud bill, and for reading the summarized view, reading your AWS bill line by line.
Which cost column should you read?
The CUR carries several cost measures, and choosing the wrong one is the most common analysis error. The table sets them straight.
| Cost column | What it shows | When to use it |
|---|---|---|
| Unblended cost | What AWS charged on the day, with upfront commitment fees as one time charges | Reconciling to the invoice, cash accounting |
| Amortized cost | Upfront Savings Plans and Reserved Instance fees spread evenly across the term | Trend analysis, unit economics, showback and chargeback |
| Net amortized cost | Amortized cost after private rate or discount adjustments | True effective cost under an enterprise agreement |
| Public on demand cost | What the usage would cost at list price with no discounts | Measuring how much your commitments actually save |
Reporting unblended cost for trends makes a month with an upfront purchase look like a spike that did not really happen, and reporting amortized cost to finance for invoice reconciliation will never tie out. Comparing net amortized cost against public on demand cost is how you measure your effective savings rate, covered in measuring effective savings rate on AWS.
How do you query the report?
The CUR is delivered to S3 in a columnar format suited to analytical queries. The common patterns are to query it in place with Athena, to load it into a data warehouse for joins against business metadata, or to enable the FOCUS export so the same data lands in the cross cloud standard schema. Querying directly is what unlocks the granularity: you can group cost by tag to attribute spend to teams, by usage type to find the exact thing that grew, or by resource to chase a single runaway. Because Athena bills by data scanned, partition the CUR by date and select only the columns you need, the same discipline described in Athena and query cost discipline, so the report that finds your savings does not itself become a cost.
Ask your team which cost column their dashboards use. If they cannot say, or if trend charts use unblended cost, the numbers leadership sees are distorted by every upfront purchase, and the first fix is to standardize on amortized cost.
What does the report reveal in practice?
The CUR turns vague cost questions into precise answers, which is where savings start.
A Fortune 500 retailer saw compute spend climbing in the console but could not explain it. We queried the CUR grouped by usage type and linked account, and the increase resolved to a single non production account running a fleet of oversized instances around the clock, plus a data transfer usage type quietly growing from a chatty service. Neither was visible in the summarized console view. Rightsizing the fleet, scheduling it off out of hours, and fixing the transfer pattern removed the increase, and standardizing reporting on amortized cost gave finance a trend it could finally trust. The work was part of a program that left the AWS estate materially lighter. Figures are verified against billing data and anonymized.
Frequently asked questions
What is the AWS Cost and Usage Report?
What is the difference between amortized and unblended cost?
How do you query the Cost and Usage Report?
Turn your CUR into a saving plan
We query your Cost and Usage Report to the resource level, surface the spend the console hides, and turn it into a safe reduction plan, independent of any provider and taking zero provider commissions. Start with our playbook, then bring us your report.
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.