Yes—with an important distinction. In August 2026, paid adoption of AI providers continued to rise among businesses observed by Ramp, while Ramp also reported lower spending by the top 1% of AI spenders and a lower effective price per token. Those figures measure different things: cheaper tokens do not prove that every company’s total AI bill fell.
What Ramp found in August 2026
Ramp Economics Lab’s September 9, 2026 update reports that paid AI adoption kept growing in August, though its growth rate was slowing. Within Ramp’s platform data, 43.8% of U.S. businesses paid for Anthropic subscriptions or tokens, and 39.8% paid for OpenAI subscriptions or tokens. Ramp lead economist Ara Kharazian wrote, “In August, Anthropic extended its lead in business AI adoption.” Ramp’s September 2026 AI Index update describes these as platform-based adoption measures, not a census of all U.S. businesses.
The update also reports two declines, but neither should be read as a direct measure of all companies’ total AI bills:
- Top-end spending: Ramp estimates that monthly AI spend per employee among the top 1% of spenders fell 9.7% from July to August, from $7,976 to $7,205.
- Effective token price: Ramp’s index put the effective price per million tokens at $0.68, 41% below its March 2026 peak of $1.15.
The top-1% figure comes from a small, volatile cohort. Ramp says late transactions led it to revise July’s estimate, so the month-to-month comparison may change as more transactions arrive.
Recommended Free Tools
Why lower token prices do not automatically mean lower AI bills
A business’s total spend depends on more than the price of a token. It also depends on how much it uses AI, which models it chooses, and what subscriptions or services it pays for. A falling effective price per million tokens can coincide with a rising bill if usage grows enough; it can also accompany lower spend if volume does not make up the difference.
Ramp offers several possible contributors to the lower effective price: provider-announced price cuts, greater use of lower-cost standard models, and summer seasonality. These are explanations Ramp raises, not a causal accounting that establishes how much each factor contributed. Ramp also says volume so far has not necessarily been sufficient to offset falling effective prices.
What the model mix suggests
Ramp reports that frontier models accounted for 45% of token share, down from an August peak of 53%. It says standard models, including GPT-5.6 Terra and Claude Sonnet, drove volume increases. Ramp’s interpretation is that businesses are using standard models that they consider performant and more cost-effective; the figures do not establish that every business changed models or why it did so.
The update says information, finance, and professional services lead on adoption. It also reports limited open-source adoption under its routing-platform proxy. That proxy is not a complete measure of all open-source AI use: the index is based on spending visible through Ramp and the routing platform it discusses.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →How to read Ramp’s adoption and cost figures
These measures answer different questions. Adoption is the share of businesses with a paid transaction; spend per employee measures dollars in a defined high-spending cohort; effective price per million tokens is a unit-price index; and token share describes usage mix. None alone establishes total market revenue, the average AI bill across all businesses, or how much AI any one company uses.
- Paid adoption: Useful as an indicator of paid AI activity among businesses in Ramp’s data, not a market-wide revenue share.
- Spend per employee: The reported decline applies to Ramp’s top 1% cohort, not to a typical company or employee.
- Effective token price: A lower unit-price index does not establish that total usage costs fell for each business.
- Model mix: Token share indicates the distribution of token usage in Ramp’s reporting, not the share of business spending captured by each provider.
What the index covers—and what it misses
Ramp describes its AI Index as spend-based research tracking how American businesses use AI. It reflects transactions visible through Ramp, so free tools and use through personal accounts can be missed; the results may also reflect the kinds of companies that use Ramp’s financial platform. The index is updated monthly, and late transactions can revise earlier estimates.
Ramp’s methodology page describes a sample of more than 70,000 American businesses, but that figure appears in the page’s search-result description rather than a page text that could be checked. Treat the sample size as reported methodology-page context, not as a fully verified account of the underlying sample. Ramp AI Index methodology page
The practical takeaway is narrow but useful: Ramp’s August 2026 data indicates that paid AI adoption continued to spread among its observed businesses even as its top-spender and effective-price measures declined. It does not prove that businesses overall—or each business—spent less on AI.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteQuick Recap
Best Value
- Family farms not data design for people against AI server farms, data center expansion, rural land buyouts, corporate agriculture, and industrial tech development replacing farmland and open space. Rural conservation and anti data center message.
- AI protest design for farmers, land conservation supporters, anti AI activists, sustainability groups, environmental advocates, rural communities, and people opposing server farm construction, power grid strain, and farmland destruction.
- Lightweight, Classic fit, Double-needle sleeve and bottom hem
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
