AWS Clean Rooms cost: per compute-unit of collaboration queries

Quick answer

AWS Clean Rooms bills for the compute (Clean Rooms Processing Units, CRPU) that runs privacy-preserving analytics across multiple parties' data without sharing raw data. Cost scales with query volume and complexity, so running efficient, scoped queries and limiting how much data each query processes are the levers. The underlying data storage (S3) and any Clean Rooms ML add their own cost.

AWS Clean Rooms lets multiple parties analyze their combined data (for advertising measurement, research, or partnerships) without any party sharing its raw data with the others. It is priced by the compute that runs the collaboration queries, so cost tracks how much querying the collaboration does.

Per compute-unit

Component	Bills for
Query compute (CRPU)	Compute running collaboration queries
Underlying storage	S3 for the data (separate)
Clean Rooms ML	Priced separately if used

Queries run on Clean Rooms Processing Units (CRPU), and you pay for the compute a query consumes, so heavier queries over more data cost more. The parties' data stays in their own S3 (billed as S3), and the optional Clean Rooms ML for privacy-preserving modeling is priced separately. So the main Clean Rooms charge is the query compute.

What drives the bill

Query volume and complexity drive cost. Frequent, complex collaboration queries scanning large datasets consume more CRPU. So a collaboration running many analyses over large combined datasets costs more than occasional, scoped queries. The privacy controls (aggregation thresholds, allowed columns) shape queries but the compute is the cost.

Controlling Clean Rooms cost

Run scoped queries that process only the data the analysis needs rather than full-table scans, batch or schedule collaboration analyses rather than running many ad-hoc queries, and be mindful of the underlying S3 storage for the datasets. Because it is compute-per-query, the scan-less-and-query-efficiently discipline of Athena applies: process less data per query to spend less compute.

FAQ

How is AWS Clean Rooms priced?

By the compute (Clean Rooms Processing Units, CRPU) that runs privacy-preserving collaboration queries across multiple parties' data without sharing raw data. Cost scales with query volume and complexity. The underlying data storage (S3) and optional Clean Rooms ML are priced separately.

How do I reduce Clean Rooms cost?

Run scoped queries that process only the data the analysis needs rather than full-table scans, batch or schedule collaboration analyses rather than running many ad-hoc queries, and account for the underlying S3 storage. Since it is compute-per-query, processing less data per query is the main lever, like Athena.

What is a CRPU in AWS Clean Rooms?

A Clean Rooms Processing Unit, the compute unit that runs collaboration queries. You pay for the CRPU a query consumes, so heavier queries scanning more data cost more. Query volume and complexity drive the CRPU consumption, which is the main Clean Rooms charge on top of the underlying S3 storage.

What drives Clean Rooms cost?

Query volume and complexity. Frequent, complex collaboration queries scanning large combined datasets consume more CRPU compute. A collaboration running many analyses over large datasets costs more than occasional, scoped queries. The privacy controls shape what queries can run, but the compute they consume is the cost.

Does the data in Clean Rooms cost extra?

The parties' data stays in their own S3 and is billed as S3 storage, separate from the Clean Rooms query compute. So account for both the CRPU compute that runs the collaboration queries and the S3 storage for the underlying datasets, plus Clean Rooms ML if used for privacy-preserving modeling.

Does C3X estimate Clean Rooms cost?

Clean Rooms cost is driven by query compute (CRPU), a usage input tied to query volume and complexity. C3X prices the surrounding infrastructure, including the S3 storage for the datasets, and you model query patterns to estimate the collaboration compute charges.

What to do next

Estimate the infrastructure around your data collaboration before you build it. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.