This edition · October 25, 2026

Enterprise AI Goes Mainstream

Read the edition note

Adoption and spend

Capability and cost

  • Frontier AI capability

    168.8 points

    Claude Opus 5.5

    ±0.0 points since September 3, 2026As of September 22, 2026

    Monitor

  • Open-weight model gap

    11.0 points

    Claude Opus 5.5 vs Kimi K3

    +0.1 points since September 3, 2026As of September 22, 2026

    Evaluate

  • AI agent task length

    17.4 hours

    claude-mythos-preview-early

    +45.4% since February 5, 2026As of April 7, 2026

    Act now

  • Price of near-frontier AI

    $4.00 per million tokens

    Gemini 4 Argon

    As of October 25, 2026

    Evaluate

Talent and workforce

Risk and regulation

By industry

Latest sector figures from each survey, under the survey's own sector names on each indicator page.

IndustryUS businesses using AIUS businesses planning AI useEU enterprises using AI
Financial Services40.5%47.4%—
Manufacturing22.9%27.0%17.3%
Retail & Consumer18.5%22.5%15.5%
Technology & Telecom46.4%51.1%62.5%
Energy & Utilities——33.6%
Public Sector33.4%36.7%—
Professional Services43.8%48.8%40.4%

Regulatory dates ahead

  1. · European Union

    EU AI Act deadline to mark AI-generated content from existing systems

    Providers of generative AI systems placed on the EU market before 2 August 2026 must mark their outputs as AI-generated and make them detectable (Article 50(2)). Systems placed on the market later already had to comply. European Commission, AI Act Service Desk

  2. · European Union

    EU AI Act bans AI-generated non-consensual intimate imagery

    New prohibitions apply to AI systems that generate or manipulate non-consensual intimate material and child sexual abuse material. Providers and deployers of image and video generation need safeguards in place. European Commission, AI Act Service Desk

  3. · California, United States

    California rules for automated decisionmaking technology apply

    Businesses subject to the CCPA that use automated decisionmaking technology (ADMT) to make significant decisions about consumers must comply with the ADMT requirements of the California Privacy Protection Agency's regulations. California Privacy Protection Agency

  4. · European Union

    EU AI Act rules for high-risk AI use cases apply

    Obligations for high-risk AI systems in sensitive areas listed in Annex III, including biometrics, critical infrastructure, education, employment, migration and border control, apply after the AI Omnibus extended the deadline. Deployers in these areas need risk management, data governance and human oversight in place. European Commission, AI Act Service Desk

  5. · European Union

    EU AI Act rules for AI in regulated products apply

    Obligations for high-risk AI systems embedded in products covered by EU product legislation (Annex I), such as machinery, lifts, toys and medical devices, apply at the end of their extended transition period. European Commission, AI Act Service Desk

Download the data

Every indicator and reading as open data, with sources. Our compilation is free to reuse under CC BY 4.0 with attribution to Enterprise AI in Numbers; each figure stays subject to its original source's terms, listed on its indicator page.

CSV JSONMethodology and sources