Business & Industry

Business & Industry is a data page on the AI desk of Glassbox Auditing. How many U.S. businesses use AI (Census, every two weeks), by industry beside each industry's business conditions, by state, and what near-frontier AI costs. Adoption is where model progress turns into business spending, productivity and changes in hiring.

Last updated: 7 Nov 2026.

What is the latest reading?

U.S. businesses using AI in any business function: 24.2% as of 18 Oct 2026, +6.9 pts since Nov 2025. Source: U.S. Census Bureau, Business Trends and Outlook Survey.

What is on this page today?

The cheapest model within 5 points of the frontier costs $5.00 per million output tokens (GPT-6.1 Terra), against $50.00 for the frontier model (GPT-6 Astra) — $10.00 when the terminal began logging it on 2026-09-25 · OpenRouter's public price list against Epoch's capabilities index · a million tokens is roughly 750,000 words (OpenRouter and Epoch AI, 7 Nov 2026)
24.1% of U.S. businesses use AI — 27.8% plan to within six months (Census Bureau (Business Trends and Outlook Survey), 7 Nov 2026)
Information leads industries on AI use: 48% of businesses, +7.9 points over 6 months — lowest: Agriculture, forestry, fishing and hunting at 9% · four-survey average of firms using AI in any business function · the page sets each industry's use beside its business conditions (Census Bureau, 17 Oct 2026)
District of Columbia leads states on AI use: 37% of businesses, +13.2 points above the nation — the nation 24.1% · change over 6 months · four-survey average; small states have wide survey margins and Census suppresses the least reliable cells (Census Bureau, 17 Oct 2026)

Where does the data come from?

Items on this page come from OpenRouter and Epoch AI; Census Bureau (Business Trends and Outlook Survey); Census Bureau. Each item links to its own source. All sources, how items are checked.

What does this page lead to?

Labor Market (Macro): Firms that adopt AI change which tasks their staff do and, in some cases, how many they hire; Census asks about both, and the effect on jobs shows in BLS data with a lag. (either way, depending on the cause; months to years; economy scale) Source: U.S. Census Bureau, Business Trends and Outlook Survey.
US Map (Domestic Intelligence): AI use is highest in states with large information, finance and professional-services sectors; those industries carry much of those states' output and wage growth (BEA). (either way, depending on the cause; years; economy scale) Source: BEA GDP by state and industry (SAGDP).
Cloud & Hubs (Companies): Most businesses reach AI through cloud platforms (Azure OpenAI, Amazon Bedrock, Google Vertex), so wider use adds to the clouds' customers and usage. (moves the same way; quarters; company scale) Source: AWS and Google Cloud customer stories.
Business Activity (Macro): The survey that measures AI use also reports each industry's revenue, demand and outlook; AI-heavy industries were stronger to begin with, so the link is an association, not cause. (either way, depending on the cause; months; economy scale) Source: U.S. Census Bureau, Business Trends and Outlook Survey.
Job Postings (Feed): Firms that adopt AI post roles to build and run it (data, machine-learning and software jobs), which shows in postings across the wider job market. (moves the same way; months; sector scale) Source: freehire public API (open job postings crawled from company career sites and job boards).

What feeds this page?

Progress Level (AI): As capability rises and near-frontier prices fall, more tasks become worth handing to AI; Census's business survey shows how many firms actually do. (moves the same way; months to years; economy scale) Source: U.S. Census Bureau, Business Trends and Outlook Survey.
Labs & Countries (AI): A small gap between open-weights and closed models lets firms run capable models on their own hardware or cheap hosts, which tends to lower what near-frontier AI costs a business. (either way, depending on the cause; months; sector scale) Source: OpenRouter model prices.
AI Models (AI): Cheaper, more capable models lower the cost of putting AI into a business process; Census's survey shows how many firms do. (moves the same way; months; economy scale) Source: OpenRouter model prices.