The AWS Cost and Usage Report, or CUR, is AWS's most granular billing dataset: every charge, for every resource, broken down by hour or day, with the pricing, usage and discount detail behind each line. It is delivered as files to an S3 bucket on a schedule you set, where you query it with Amazon Athena, load it into a warehouse, or feed it to a FinOps platform. If Cost Explorer answers "what did we spend," the CUR answers "exactly which resource, in which account, under which rate, at which hour," which is what you need to allocate cost accurately and find savings.
This explainer sits under our complete guide to AWS cost optimization, the pillar for this cluster, and it underpins the See step of our See, Cut, Lock, Run method: the CUR is the raw material that makes every dollar attributable to an owner. Once you can read it, you can act on it through tools like AWS Budgets and Cost Anomaly Detection.
Cost Explorer is a fast, visual summary with limited granularity and history. The CUR is the full record: every line item, hourly resolution, and the columns that let you build your own allocation logic. Use Cost Explorer to spot a trend, the CUR to explain and fix it.
What the CUR contains
Each row in the CUR is a usage or charge record with dozens of columns. The ones that matter most for cost work fall into a few groups. Identity columns tell you the linked account, the region, the availability zone, and the specific resource ID when resource-level data is enabled. Product columns describe the service, the usage type, and the operation, so you can tell an EC2 running hour from an EBS provisioned-storage charge. Cost columns separate unblended cost, blended cost, and amortized cost, which is how you spread the up-front portion of a Reserved Instance or Savings Plan across the hours it covers. Discount and reservation columns show how Savings Plans and RIs applied to each line. Finally, your cost allocation tags appear as their own columns once activated, which is what makes showback and chargeback possible.
How it is delivered and queried
You configure the CUR through AWS Data Exports in the Billing console, choosing the S3 bucket, the granularity, and the format. AWS offers a standard CUR schema and a FOCUS-aligned export, the open FinOps Open Cost and Usage Specification that normalizes billing across clouds. Files land on a schedule and are typically partitioned by month and refreshed as AWS finalizes charges. The common pattern is to point Amazon Athena at the bucket so you can run SQL directly, or to load the data into a warehouse such as Redshift or a third-party platform. Enable resource IDs and hourly granularity if you want to do real rightsizing and allocation work; without them you lose the detail that makes the CUR worth using.
What you actually do with it
Three jobs depend on the CUR. The first is allocation: joining tag columns to cost columns to answer who spent what, by team, product, or environment, which is the foundation of cost allocation tags that actually stick. The second is commitment analysis: using amortized cost and the reservation columns to measure coverage and utilization, so you know whether your Savings Plans are working. The third is waste and anomaly detection: grouping unblended cost by usage type and resource to surface idle resources, sudden spikes, and the data transfer charges that hide inside aggregate views, as in AWS data transfer and egress charges.
Want your CUR turned into a savings plan?
Our AWS cost audit ingests your Cost and Usage Report, allocates every dollar to an owner, and ranks every opportunity by size. On the performance model you pay only from realized savings. No savings, no fee.
Book an AWS cost audit →Common traps when reading the CUR
Three mistakes recur. Mixing cost types is the first: summing unblended and amortized cost together double counts commitments, so pick the right column for the question. Ignoring credits and refunds is the second: line items with negative cost and specific charge types can mask the true run rate if you filter them out by accident. Missing untagged spend is the third: a large pool of records with no allocation tag is not noise, it is unallocated cost that needs an owner before any showback is credible.
| Question | CUR column group | Use |
|---|---|---|
| Who spent this? | Account, region, tags | Allocation and showback |
| On what? | Product, usage type, operation | Service-level breakdown |
| At what rate? | Unblended, amortized, discount | Effective rate and commitment value |
| Covered by a commitment? | Reservation, Savings Plan | Coverage and utilization |
CUR schema details and the Data Exports interface reflect AWS as of May 2026. Verify column names and the FOCUS export options in the Billing console before building pipelines, as AWS revises the schema over time.
The AWS Cost Optimization Field Guide ships with the Athena queries we run against the CUR for allocation, commitment coverage, and waste detection. It is the downloadable companion to this explainer.
The short version
The AWS Cost and Usage Report is the full, hourly, line-item record of your AWS spend, delivered to S3 and queried with Athena or a warehouse. Enable resource IDs and hourly granularity, use the right cost column for each question, and join your tags to make spend attributable. It is the data layer beneath allocation, commitment analysis, and waste detection, and the starting point of any real AWS cost optimization engagement, exactly as in our SaaS on AWS case study.