specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Shadeform
providerId: shadeform
created: '2026-06-21'
modified: '2026-06-21'
reconciled: false
tags:
- GPU
- Cloud
- Marketplace
- Compute
- Infrastructure
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Shadeform spend. Shadeform is a GPU cloud marketplace that bills
  usage-based, per-GPU-hour rates that pass through from the underlying cloud
  provider, plus capacity-over-time charges for persistent volumes. There are no
  per-token or per-request fees. The dominant cost driver is GPU hours, so cost
  control centers on right-sizing GPU type, routing to the cheapest available
  provider/region, and aggressively reclaiming idle instances with auto-delete and
  spend alerts. Per-GPU rates are dynamic marketplace passthrough and are not
  reconciled here.
notes: >-
  Per-GPU-hour rates change continuously with marketplace supply and demand;
  verify live rates via GET /instances/types (hourly_price, in cents) and the
  Shadeform GPU directory during reconciliation.
sources:
- https://www.shadeform.ai
- https://docs.shadeform.ai/api-reference/instances/instances-types
- 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: Shadeform
serviceCategory: Compute
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Shadeform GPU Marketplace
  ServiceCategory: Compute
  ProviderName: Shadeform
  PublisherName: Shadeform
  InvoiceIssuerName: Shadeform
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: gpu_instance_hours
  description: GPU instance runtime billed per hour at the underlying provider's passthrough rate.
  unit: hours
  aggregation: sum
  dimensions:
  - account
  - cloud
  - region
  - shade_instance_type
  - gpu_type
- name: volume_gb_hours
  description: Persistent volume capacity billed over time (GB-hours).
  unit: gb_hours
  aggregation: sum
  dimensions:
  - account
  - cloud
  - region
- name: instance_count
  description: Number of active instances; tracked for concurrency and allocation, not directly billed.
  unit: instances
  aggregation: max
  dimensions:
  - account
  - cloud
  - tags
principles:
- name: Visibility
  description: Pull per-instance cost_estimate and hourly_price from the API; tag instances per workload/team for chargeback.
- name: Allocation
  description: Use instance tags and per-key usage to map GPU hours and volume storage to internal cost centers.
- name: Optimization
  description: Route launches to the cheapest available cloud/region/GPU type via GET /instances/types sorted by price; right-size GPU type; set auto_delete and spend alerts to reclaim idle instances.
- name: Accountability
  description: Assign owners per project; review GPU-hour and volume spend against budgets and enforce auto-delete thresholds.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com