Few People Pay For AI, But Those Who Do Spend Big
AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

Andreessen Horowitz’s latest AI product ranking adds observed U.S. consumer card spending, using YipitData panels. It reports that 4.5% of U.S. consumers had an active paid personal subscription to ChatGPT, Gemini or Claude in August, while the top 1% of AI spenders accounted for nearly a fifth of observed spending and averaged about $900 a month.

Venture capital firm Andreessen Horowitz has added observed U.S. consumer spending to its latest ranking of popular AI products, finding that 4.5% of consumers had an active paid personal subscription to ChatGPT, Gemini or Claude in August. Within the spending captured by its data provider’s panels, the top 1% of spenders accounted for nearly a fifth of spending and averaged about $900 a month, a sharp contrast with the roughly $25 monthly outlay of a typical paying user.

The spending figures come from YipitData panels tracking U.S. consumer cards, according to Andreessen Horowitz. The firm cautions that the sample does not represent total company revenue; it offers an observed view of consumer payments rather than a full accounting of what AI companies earn. OpenAI and Anthropic lead the new spending ranking. Gemini is not included because its revenue cannot be separated from Google’s other subscriptions.

The report says the share of consumers with paid personal subscriptions to the three named assistants roughly doubled over the preceding year, from about half the August level. ChatGPT accounts for most of that paying group, according to the report. Among observed spenders, the top 1% spent nearly five times as much in aggregate as the bottom half combined. Their average monthly spending was about $900, while the typical paying user’s spending remained around $25 a month.

Spending growth was concentrated at the top: the report says the top 1% of payers were spending 79% more than in early 2025. Those customers were more likely than other users to buy building and automation products such as n8n, Manus and fal, as well as creative software including Higgsfield, Figma and HeyGen. Andreessen Horowitz characterizes this segment as prosumers using AI for professional work; that is the firm’s interpretation of the product choices, not a measure of every buyer’s motivation.

At a glance
reportWhen: Published in the seventh edition of And…
The developmentAndreessen Horowitz has added panel-based U.S. consumer spending data to its seventh ranking of popular AI products, showing a small paid audience and highly concentrated spending.

Heavy Users Drive AI Spending

The findings point to a gap between AI use and paid adoption: broad experimentation has not yet become routine subscription spending for most consumers. The contrast matters for companies trying to turn free users into revenue. A relatively small group of customers currently accounts for a large share of the consumer spending visible in the panels, and its purchases often extend beyond chatbot subscriptions to specialized tools for work and creative tasks.

That pattern may also shape how products are built and priced. Providers have an incentive to serve users willing to pay for advanced features and higher-tier plans, but a business dependent on a narrow group could be exposed if those customers change tools or reduce spending. The report does not establish that this concentration will persist, nor does its panel data show companies’ total revenues.

There is also a potential access gap. Andreessen Horowitz says paid accounts offer more features and can be harder to use, while many free users engage with only basic capabilities. If advanced use remains concentrated among people with the money, skills or professional need to pay, the divide could reinforce existing inequalities. The report frames that as a possibility; it does not measure resulting effects on income, education or employment.

Amazon

AI subscription software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

A New Spending Measure

The latest publication is the seventh edition of Andreessen Horowitz’s ranking of widely used consumer AI products. Its lists combine measures such as web traffic and mobile usage, with this edition adding observed card spending for U.S. consumers. The firm says the rankings show a relatively settled top tier, with only 11 products appearing for the first time, the fewest newcomers in its lists so far.

ChatGPT leads the web ranking, ahead of Gemini and Claude, and also ranks first across the report’s three top-50 lists. Canva and Notion make the web top 10. Only seven products appear in all three lists, according to the report. These measures describe different things: traffic and usage indicate reach, while the card panels capture a portion of consumer purchases. They should not be treated as interchangeable measures of market share or company revenue.

The report also describes shifts among the leading assistants. Claude’s paid U.S. subscriber count passed Gemini’s in March 2026, according to Andreessen Horowitz; after Google changed its pricing in June, the two were described as close. The firm says ChatGPT regained the lead in net new subscriptions by August after Claude’s spring momentum. Those comparisons are attributed to the report and do not, by themselves, explain the causes of changes in subscriptions.

“The top 1% of spenders accounts for nearly a fifth of all observed AI spending.”

— Andreessen Horowitz

Amazon

AI creative tools for professionals

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Limits of the Spending Data

The card-spending figures are based on YipitData panels, not a complete census of AI purchases. Andreessen Horowitz says they do not represent total revenue, and the available material does not specify the panel’s size, its full methodology or how closely it reflects all U.S. consumers. The reported 4.5% subscription share is a snapshot for August; the source material does not state the year for that particular figure.

Gemini’s absence from the spending ranking is a measurement limitation, because its revenue cannot be separated from Google’s other subscriptions. The data therefore cannot support a direct, like-for-like comparison of total consumer revenue across the three assistants. It is also unclear how much of the high spend reflects recurring subscriptions versus other purchases, or how many people use paid products for personal rather than work purposes.

The report suggests the skills gap between advanced and casual AI users could contribute to social inequality, while also raising the possibility that easier and cheaper tools may reduce that divide. Neither outcome is established by the spending data. The extent to which paid use translates into measurable benefits for users is not given.

Amazon

AI automation products for work

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Whether Paid Use Broadens

The next useful signal will be whether paid subscriptions spread beyond heavy users or whether spending continues to depend on a small group of high-value customers. Later editions or updated company data may clarify whether the August subscription share has changed and whether the reported spending concentration holds over time.

Readers should also watch for clearer, comparable disclosures on consumer revenue. Google’s bundled subscriptions make Gemini difficult to isolate in these panels, so a change in reporting or product packaging could affect future comparisons. For now, Andreessen Horowitz’s figures offer a panel-based snapshot of U.S. consumer spending, not a complete tally of the AI market.

Amazon

AI productivity software

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

How many U.S. consumers paid for a personal AI subscription?

Andreessen Horowitz reports that 4.5% of U.S. consumers had an active paid personal subscription to ChatGPT, Gemini or Claude in August. The source material does not specify the year for that August snapshot.

How much did the highest-spending users pay?

The report says the top 1% of observed spenders averaged about $900 per month. This is an average for the panel-based group, not a typical payment for all AI users.

Do these figures show total AI company revenue?

No. The spending estimates use YipitData consumer-card panels and, according to Andreessen Horowitz, do not represent total revenue. Gemini is also absent from the spending ranking because Google’s AI revenue cannot be separated from its other subscriptions.

Which companies lead the spending ranking?

OpenAI and Anthropic lead the new spending ranking, according to Andreessen Horowitz. The report does not provide a fully comparable spending position for Google’s Gemini because its subscription revenue is bundled with other Google services.

Source: rss

COLUMBUS DAY / I

Columbus Day / Indigenous Peoples' Day Picks

As an affiliate, we earn on qualifying purchases.

You May Also Like

Introducing Grok 4.7

Grok 4.7 has been officially introduced, marking a new version in the series. The development is gaining attention, though details remain unconfirmed.

Livenerf: Has Opus 5.5 Been Nerfed Yet?

A GitHub benchmark has collected six of 30 daily Opus 5.5 runs. Its first comparison with launch-week performance is not expected until late October.

DeepSeek V4.1 Flash

DeepSeek v4.1 Flash, the latest update to the AI search platform, has been launched, promising faster performance and improved accuracy. Details are confirmed, but some features remain unconfirmed.

X47.c Botnet Targets Windows And Drains AI APIs Using xAI Grok

A SecurityWeek headline links the x47.c Windows botnet to xAI’s Grok API, but available material gives no technical evidence, scale or confirmed impact.