Compute Thresholds

{
  "id": "compute-thresholds",
  "numericId": null,
  "path": "/knowledge-base/responses/compute-thresholds/",
  "filePath": "knowledge-base/responses/compute-thresholds.mdx",
  "title": "Compute Thresholds",
  "quality": null,
  "importance": null,
  "contentFormat": "article",
  "tractability": null,
  "neglectedness": null,
  "uncertainty": null,
  "causalLevel": null,
  "lastUpdated": "2026-02-09",
  "llmSummary": null,
  "structuredSummary": null,
  "description": "Regulatory thresholds based on training compute to trigger safety requirements",
  "ratings": null,
  "category": "responses",
  "subcategory": null,
  "clusters": [
    "ai-safety"
  ],
  "metrics": {
    "wordCount": 7,
    "tableCount": 0,
    "diagramCount": 0,
    "internalLinks": 0,
    "externalLinks": 0,
    "footnoteCount": 0,
    "bulletRatio": 0,
    "sectionCount": 0,
    "hasOverview": false,
    "structuralScore": 2
  },
  "suggestedQuality": 13,
  "updateFrequency": null,
  "evergreen": true,
  "wordCount": 7,
  "unconvertedLinks": [],
  "unconvertedLinkCount": 0,
  "convertedLinkCount": 0,
  "backlinkCount": 0,
  "redundancy": {
    "maxSimilarity": 0,
    "similarPages": []
  }
}

{
  "id": "compute-thresholds",
  "type": "policy",
  "title": "Compute Thresholds",
  "description": "Compute thresholds define capability boundaries using training compute (measured in FLOP) as a proxy. The EU AI Act uses 10^25 FLOP for GPAI obligations; the US Executive Order uses 10^26 FLOP for reporting requirements. These thresholds aim to capture frontier models while minimizing regulatory burden on smaller systems.",
  "tags": [
    "compute-governance",
    "regulation",
    "flop-thresholds"
  ],
  "relatedEntries": [
    {
      "id": "compute-governance",
      "type": "policy"
    },
    {
      "id": "eu-ai-act",
      "type": "policy"
    },
    {
      "id": "ai-executive-order",
      "type": "policy"
    }
  ],
  "sources": [
    {
      "title": "EU AI Act GPAI Thresholds",
      "url": "https://artificialintelligenceact.eu/"
    },
    {
      "title": "US Executive Order Compute Thresholds",
      "url": "https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/"
    }
  ],
  "lastUpdated": "2025-12",
  "customFields": [
    {
      "label": "Approach",
      "value": "Define capability boundaries via compute"
    }
  ]
}

No facts for this entity

No external links

No backlinks

{
  "title": "Compute Thresholds",
  "description": "Regulatory thresholds based on training compute to trigger safety requirements",
  "sidebar": {
    "order": 50
  },
  "quality": 0,
  "importance": 0,
  "lastEdited": "2026-02-09",
  "entityType": "approach"
}

---
title: "Compute Thresholds"
description: "Regulatory thresholds based on training compute to trigger safety requirements"
sidebar:
  order: 50
quality: 0
importance: 0
lastEdited: "2026-02-09"
entityType: approach
---

This page is a stub. Content needed.