specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Thunder Compute
providerId: thundercompute
created: '2026-06-21'
modified: '2026-06-21'
reconciled: true
tags:
- GPU
- Cloud
- Infrastructure
- AI
- Compute
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Thunder Compute spend. Thunder Compute bills on-demand
  virtual GPU instances per minute at a fixed per-GPU-hour rate that varies
  by GPU model and tier (development vs. production). Additional vCPUs are
  billed per vCPU-hour, additional persistent storage per 100GB-hour while
  running, and snapshots per GB-month. There are no data egress fees. The
  dominant cost driver is GPU run time, so the primary optimization levers
  are stopping/deleting idle instances, right-sizing GPU model and count, and
  using snapshots to persist state cheaply between sessions.
notes: >-
  Per-GPU-hour and per-resource rates reconciled against the public pricing
  page on 2026-06-21; verify against GET /v1/pricing and the pricing page,
  which are authoritative and change over time.
sources:
- https://www.thundercompute.com/pricing
- https://www.thundercompute.com/docs
- 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: Thunder Compute
serviceCategory: Compute
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Thunder Compute GPU Instances
  ServiceCategory: Compute
  ProviderName: Thunder Compute
  PublisherName: Thunder Compute
  InvoiceIssuerName: Thunder Compute
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: gpu_hours
  description: GPU run time billed per minute at the per-GPU-hour rate for the GPU model and tier.
  unit: hours
  aggregation: sum
  dimensions:
  - account
  - instance
  - gpu_type
  - mode
- name: vcpu_hours
  description: Additional vCPU-hours beyond the 4 included per instance, billed per vCPU/hour.
  unit: hours
  aggregation: sum
  dimensions:
  - account
  - instance
- name: storage_gb_hours
  description: Additional persistent storage beyond the included 100GB, billed per 100GB/hour while running.
  unit: gb_hours
  aggregation: sum
  dimensions:
  - account
  - instance
- name: snapshot_gb_months
  description: Long-term snapshot storage billed per GB/month.
  unit: gb_months
  aggregation: sum
  dimensions:
  - account
  - snapshot
principles:
- name: Visibility
  description: Track per-instance GPU run time and storage via GET /v1/instances/list and console billing; attribute spend by GPU type and mode.
- name: Allocation
  description: Tag instances and API tokens per workload/team and map to internal cost centers.
- name: Optimization
  description: Stop or delete idle instances (GPU time is the main driver), right-size GPU model/count, snapshot and tear down between sessions, and prefer development tier when production performance is unnecessary.
- name: Accountability
  description: Assign owners per instance/token; review GPU-hour and storage spend monthly against budget.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com