specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Amazon SageMaker
providerId: amazon-sagemaker
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: true
tags:
  - FinOps
  - FOCUS
  - AWS
  - Machine Learning
  - SageMaker
  - Cost Management
description: 'FOCUS-aligned FinOps for Amazon SageMaker AI: per-ML-instance-hour billing across notebook,
  training, batch transform, real-time, and asynchronous inference, plus serverless inference (GB-second
  + GB-data) and SageMaker Savings Plans for 1- or 3-year commitments.'
sources:
  - https://aws.amazon.com/sagemaker-ai/pricing/
alignedWith:
  framework: FinOps Foundation Framework
  frameworkUrl: https://www.finops.org/framework/
  dataSpec: FOCUS
  dataSpecVersion: '1.3'
  dataSpecUrl: https://focus.finops.org/focus-specification/v1-3/
publisherName: Amazon Web Services, Inc.
serviceCategory: AI / Machine Learning
billingModel:
  pricingCategory: Pay-As-You-Go + Committed Use
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
    - Usage
    - Purchase
    - Tax
    - Credit
    - Adjustment
focusColumns:
  ServiceName: Amazon SageMaker
  ServiceCategory: AI and Machine Learning
  ServiceSubcategory: ML Platform
  ProviderName: AWS
  PublisherName: Amazon Web Services, Inc.
  InvoiceIssuerName: Amazon Web Services, Inc.
  BillingCurrency: USD
  ChargeCategory: Usage
meters:
  - name: notebook_instance_hours
    unit: instance-hour
    aggregation: sum
    dimensions:
      - region
      - instance_type
      - user
  - name: training_instance_hours
    unit: instance-hour
    aggregation: sum
    dimensions:
      - region
      - instance_type
      - job_name
  - name: batch_transform_instance_hours
    unit: instance-hour
    aggregation: sum
    dimensions:
      - region
      - instance_type
      - job_name
  - name: realtime_endpoint_instance_hours
    unit: instance-hour
    aggregation: sum
    dimensions:
      - region
      - instance_type
      - endpoint_name
      - variant_name
  - name: serverless_inference_gb_seconds
    unit: GB-second
    aggregation: sum
    dimensions:
      - region
      - endpoint_name
  - name: serverless_inference_data_processed_gb
    unit: GB
    aggregation: sum
    dimensions:
      - region
      - endpoint_name
  - name: storage_gb_month
    unit: GB-month
    aggregation: sum
    dimensions:
      - region
      - workload
principles:
  - name: Visibility
    description: Use SageMaker Studio cost dashboards, CloudWatch metrics (Invocations, ModelLatency, CPUUtilization,
      GPUUtilization), and CUR / FOCUS export for per-job and per-endpoint spend.
  - name: Allocation
    description: Tag training jobs, endpoints, notebooks, and Studio profiles with Project, Team, Owner;
      enable cost allocation tags so charges flow into FOCUS columns.
  - name: Optimization
    description: Use Spot training for fault-tolerant jobs, right-size endpoint instance types, scale endpoints
      to zero (Serverless or async) when traffic is sparse, commit to Savings Plans for steady inference,
      use Inferentia/Trainium families for higher price-performance, and stop idle notebooks.
  - name: Accountability
    description: Set AWS Budgets per project tag, alert on training job duration anomalies, and review
      monthly endpoint utilization against InvocationsPerInstance.
maintainers:
  - name: Kin Lane
    email: kin@apievangelist.com
    url: https://apievangelist.com
  - name: Amazon Web Services
    email: support@aws.amazon.com
    url: https://aws.amazon.com