specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Anyscale
providerId: anyscale
created: '2026-05-08'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-08, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-08'
reconciled: true
tags:
- AI
- Distributed Computing
- Ray
- ML Platform
- Inference
- GPU
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Anyscale spend. Anyscale meters compute by instance-hour per
  node type (CPU and NVIDIA T4 / L4 / A10G / A100 / H100 / H200). Hosted
  customers are invoiced directly by Anyscale; BYOC customers are charged
  the underlying cloud spend by AWS / GCP plus an Anyscale platform fee per
  instance-hour, often via a cloud marketplace.
notes: >-
  In BYOC mode the FOCUS dataset will combine an Anyscale ProviderName line
  item with separate AWS / GCP ProviderName line items - reconcile both for
  total cost-of-Ray.
sources:
- https://www.anyscale.com/pricing
- https://docs.anyscale.com/
- https://focus.finops.org/focus-specification/v1-3/
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: Anyscale
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Anyscale Platform
  ServiceCategory: AI and Machine Learning
  ProviderName: Anyscale
  PublisherName: Anyscale
  InvoiceIssuerName: Anyscale
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: cpu_instance_hours
  description: CPU-only node hours billed per instance-hour ($0.0135/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
  - workload_type
- name: gpu_instance_hours_t4
  description: NVIDIA T4 GPU instance hours ($0.5682/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: gpu_instance_hours_l4
  description: NVIDIA L4 GPU instance hours ($0.9542/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: gpu_instance_hours_a10g
  description: NVIDIA A10G GPU instance hours ($1.3635/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: gpu_instance_hours_a100
  description: NVIDIA A100 GPU instance hours ($4.9591/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: gpu_instance_hours_h100
  description: NVIDIA H100 GPU instance hours ($9.2880/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: gpu_instance_hours_h200
  description: NVIDIA H200 GPU instance hours ($10.6812/hr hosted).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
- name: byoc_platform_hours
  description: BYOC platform-fee instance-hours invoiced via marketplace (rate contract-dependent).
  unit: hours
  aggregation: sum
  dimensions:
  - organization
  - cluster
principles:
- name: Visibility
  description: Pull Anyscale usage exports per cluster/project; in BYOC, also pull AWS/GCP cost exports tagged with cluster IDs.
- name: Allocation
  description: Tag clusters and services to projects/teams; map to internal cost centers.
- name: Optimization
  description: Right-size compute configs and autoscaler bounds; idle workspaces auto-suspend; use committed-spend discounts and BYOC GPU reservations.
- name: Accountability
  description: Assign owners per project; review monthly GPU-hour burn and idle utilization.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com