generated: '2026-08-13'
method: searched
source: https://aws.amazon.com/clean-rooms/pricing/
docs: https://aws.amazon.com/clean-rooms/pricing/
currency: USD
model: usage-based
plan_count: 0
note: >-
  AWS Clean Rooms publishes no named subscription tiers — there is no Free /
  Pro / Enterprise ladder to record. It is priced purely on metered consumption
  (compute measured in CRPU-hours, plus per-record ML and Entity Resolution
  charges and per-GB intermediate table storage), billed through the AWS account
  with no separate signup. plan_count is therefore an honest 0; the metered
  dimensions the provider actually publishes are captured in dimensions[] below.
  Prices are the published US East (N. Virginia) rates; AWS varies rates by
  Region.
free_tier:
  available: false
  note: >-
    The pricing page states the AWS Free Tier is not available for PySpark in
    AWS Clean Rooms, AWS Clean Rooms ML, or AWS Entity Resolution.
dimensions:
  - id: spark-sql-compute
    name: Spark SQL compute
    unit: CRPU-hour
    price: 2.00
    currency: USD
    region: us-east-1
    billing_granularity: per-second
    minimum_billing_duration: 60 seconds
    default_allocation: 32 CRPUs per Spark SQL query
  - id: pyspark-compute
    name: PySpark job compute
    unit: CRPU-hour
    price: 2.00
    currency: USD
    region: us-east-1
    billing_granularity: per-second
    minimum_billing_duration: 10 minutes
    default_allocation: 32 CRPUs per PySpark job
  - id: differential-privacy
    name: Differential Privacy surcharge
    unit: CRPU-hour
    price: 2.00
    currency: USD
    region: us-east-1
    note: >-
      Additive to the base Spark SQL compute rate, for a published total of
      $4.00 per CRPU-hour when differential privacy is applied.
  - id: ml-custom-model-training
    name: Clean Rooms ML custom model training
    unit: 1,000 records
    price: 0.01
    currency: USD
  - id: ml-custom-model-inference
    name: Clean Rooms ML custom model inference
    unit: million records
    price: 10.00
    currency: USD
    tiering:
      - up_to: 1,000 million records
        price: 10.00
        unit: million records
      - above: 1,000 million records
        price: 0.20
        unit: million records
  - id: ml-lookalike-training
    name: Clean Rooms ML lookalike model training
    unit: 1,000 profiles
    price: 0.04
    currency: USD
  - id: ml-lookalike-segments
    name: Clean Rooms ML lookalike segments
    unit: 1,000 profiles
    price: 0.25
    currency: USD
  - id: synthetic-data-generation
    name: Synthetic data generation
    unit: SDGU (Synthetic Data Generation Unit)
    price: 2.00
    currency: USD
  - id: entity-resolution-data-prep
    name: AWS Entity Resolution data preparation
    unit: 1,000 records processed
    price: 0.10
    currency: USD
  - id: entity-resolution-rule-based
    name: AWS Entity Resolution rule-based matching
    unit: 1,000 records matched
    price: 0.50
    currency: USD
    base_fee: 100.00
    base_fee_unit: per collaboration
  - id: entity-resolution-provider-based
    name: AWS Entity Resolution provider-based matching
    unit: 1,000 records processed
    price: 0.10
    currency: USD
    note: Excludes third-party data provider subscription costs.
  - id: intermediate-table-storage
    name: Intermediate table storage
    unit: GB-month
    price_range:
      min: 0.023
      max: 0.025
    currency: USD
    note: Rate varies by Region.
payment_configuration:
  note: >-
    A collaboration creator can assign which member pays for query compute, job
    compute, ML training/inference and synthetic data generation via the payment
    configuration on the collaboration and membership — so the billed party is
    not always the caller.
  docs: https://docs.aws.amazon.com/clean-rooms/latest/userguide/what-is.html#billing