01Global AI Ecosystem: Key Statistics
- ChatGPT lost 16.9 points of share among the top eight AI assistants between October 2025 and October 2026; Gemini and Claude gained most
- Muse is the fastest-growing AI product with 10M+ visits (+228.3% MoM)
- Education & Research is the fastest-growing category this month (+9.8% like-for-like)
| Metric | Latest figure / period | What it tells us |
|---|---|---|
| Tracked web visits | 17.9BOctober 2026 | 17,643 products with traffic data; measures visit volume within our tracked universe. |
| Like-for-like web growth | +2.9%October 2026 | 37.3% of comparable products grew. Read aggregate growth alongside its breadth. |
| Assistant sample leader | ChatGPT · 56.1%October 2026 | Share of web visits among eight fixed assistants; a different denominator from all tracked websites. |
| App download leader | ChatGPT · 80.0MOctober 2026 | Estimated trailing 30-day downloads: a signal of distribution and acquisition, not active users. |
| App IAP revenue leader | ChatGPT · $306.4MOctober 2026 | Estimated trailing 30-day in-app purchase revenue, including subscriptions; a partial view of mobile monetization. |
This report follows three questions: where demand lands, who drives growth, and how attention relates to spending. Start with scale and breadth, then compare assistants, categories and countries before turning to app downloads and in-app purchase revenue.
Web visits track browsing, downloads track installation activity, and IAP revenue tracks part of mobile spending. Each has its own time window and coverage; read them as complementary signals.
02Global AI Traffic: Volume and Growth
In October 2026, combined visits to 16,121 comparable products changed +2.9%. 6,010 grew, 10,111 declined and 0 were flat; the median product’s change was -9.2%.
Total visits rose, but a majority of products did not grow. In a market weighted toward large products, a few gains can lift the aggregate; category demand and distribution deserve a closer look.
Removing the current traffic leader, ChatGPT, from the same comparable sample leaves +3.3% growth for the remaining products. This helps assess the leader’s influence; other large products remain included.
The comparable sample covers 99.7% of current tracked visits. Newly tracked products, missing-month data and suspected breaks are excluded from growth calculations.
03Leading AI Apps: Market Share and Landscape
Between October 2025 and October 2026, ChatGPT’s share fell 16.9 points; Gemini gained the most (+10.5 points), followed by Claude (+9.3).
| Assistant | October 2025 | October 2026 | YoY change (pts) | Visits |
|---|---|---|---|---|
| 73.1% | 56.1% | -16.9 | 6.0B |
| 14.0% | 24.5% | +10.5 | 2.6B |
| 2.3% | 11.6% | +9.3 | 1.3B |
| 4.2% | 3.5% | -0.7 | 372.8M |
| 2.4% | 1.9% | -0.5 | 203.8M |
| 1.0% | 1.3% | +0.2 | 135.8M |
| 2.6% | 0.8% | -1.8 | 88.3M |
| 0.3% | 0.3% | -0.0 | 34.2M |
Monthly share breakdown
| Month | ChatGPT | Gemini | Claude | DeepSeek | Grok | 豆包 | PerplexityAI | Kimi |
|---|---|---|---|---|---|---|---|---|
| 2025-10 | 73.1% | 14.0% | 2.3% | 4.2% | 2.4% | 1.0% | 2.6% | 0.3% |
| 2025-11 | 70.6% | 16.3% | 2.2% | 4.2% | 2.8% | 1.1% | 2.3% | 0.5% |
| 2025-12 | 66.1% | 20.8% | 2.1% | 3.9% | 3.2% | 1.3% | 2.2% | 0.5% |
| 2026-01 | 64.1% | 23.2% | 2.3% | 3.3% | 3.5% | 1.3% | 1.9% | 0.4% |
| 2026-02 | 62.1% | 24.5% | 3.4% | 3.2% | 3.5% | 1.1% | 1.8% | 0.5% |
| 2026-03 | 57.4% | 26.0% | 6.1% | 3.5% | 3.3% | 1.6% | 1.7% | 0.5% |
| 2026-04 | 54.3% | 27.2% | 8.1% | 4.1% | 2.8% | 1.6% | 1.5% | 0.4% |
| 2026-05 | 53.3% | 27.8% | 9.1% | 4.1% | 2.3% | 1.5% | 1.3% | 0.4% |
| 2026-06 | 53.4% | 28.4% | 9.4% | 3.7% | 2.1% | 1.5% | 1.1% | 0.4% |
| 2026-07 | 54.6% | 26.9% | 9.7% | 3.5% | 2.1% | 1.5% | 1.0% | 0.6% |
| 2026-08 | 55.9% | 26.0% | 9.4% | 3.5% | 2.1% | 1.6% | 0.9% | 0.6% |
| 2026-09 | 56.8% | 25.3% | 9.7% | 3.4% | 2.0% | 1.5% | 0.9% | 0.4% |
| 2026-10 | 56.1% | 24.5% | 11.6% | 3.5% | 1.9% | 1.3% | 0.8% | 0.3% |
Share of web visits among these eight products; the sample is fixed so months are comparable. Not total market share, and not share of users.
From one dominant player to a multi-player race. ChatGPT still leads the sample — about 2.3x Gemini’s share — but it gave up 16.9 points between October 2025 and October 2026. This shows a redistribution of visit share; user-level data would be needed to establish switching or multi-product use.
The lost share was not spread evenly: Gemini and Claude captured 99% of all share gains. A plausible explanation: Gemini rides on distribution through Google Search, Android and Workspace; Claude has built a strong reputation in coding, long documents and research. For newcomers, the general-assistant entry point is being re-divided among a few giants; differentiation is more likely to come from vertical use cases.
04Breakout Products and Traffic Slowdown Alerts
Among AI products with 10M+ monthly visits, Muse grew fastest (+228.3% month over month) to 45.8M visits.
Products with 10M+ monthly visits only; suspected data breaks excluded (see methodology). Full fastest-growing chart →
5 products with 10M+ visits grew more than 40% this month. Large monthly jumps can reflect a low prior base, campaigns or traffic volatility; one month is a signal worth following. Check each product’s 12-month trend before drawing conclusions.
A decline is not necessarily decay: seasonality, redesigns and domain changes can all dent a single month. Two or three consecutive months of decline are the real warning sign.
05Concentration at the Top, the Long Tail
ChatGPT alone accounts for 33.8% of all tracked visits; the top 10 AI websites take 68.3% and the top 100 take 84.1% — the remaining 17,543 sites share the other 15.9%.
Top 20 AI websites by traffic →
A “typical” AI website (the median) gets about 13.2K visits a month — 5 orders of magnitude below ChatGPT. For most AI products the real competition is not ChatGPT but other products in the same niche.
Concentration makes distribution worth examining, but visits alone cannot establish acquisition cost, retention or payment quality. For long-tail products, peers in the same category and size range are often the more useful benchmark.
06Vertical Categories: Where Growth Lands and Who Leads
General AI takes 65.7% of all visits. Among categories with 50M+ visits, Education & Research grew fastest (+9.8% like-for-like) and Marketing & Sales declined the most (-2.2%).
| Category | Share of visits | Visits | MoM | Sites | Leader |
|---|---|---|---|---|---|
| General AI | 65.7% | 11.8B | +3.3% | 2,421 | |
| Image & Design | 9.8% | 1.8B | +1.9% | 1,480 |  |
| Office & Collaboration | 6.1% | 1.1B | +0.3% | 1,243 |  |
| Characters & Fun | 3.9% | 694.9M | +2.8% | 350 |  |
| Video | 2.3% | 413.3M | +5.4% | 957 |  |
| Programming | 2.2% | 397.0M | +1.2% | 1,732 |  |
| Education & Research | 2.0% | 357.6M | +9.8% | 1,015 |  |
| Industry Services | 2.0% | 349.6M | -1.5% | 3,109 |  |
| Audio | 1.8% | 322.2M | +3.9% | 784 |  |
| Marketing & Sales | 1.8% | 315.5M | -2.2% | 2,161 |  |
| Writing | 1.2% | 216.5M | +2.4% | 636 |  |
| News & Analytics | 0.7% | 117.1M | +7.3% | 537 |  |
| Other Industries | 0.6% | 115.5M | +3.8% | 1,218 |  |
Education & Research is a category to watch, with Wayground AI accounting for 13.2% of its current visits. Category growth is a starting point; leader concentration and individual product trends help show how widely the opportunity is shared.
Industry Services has the most products (3,109 sites) yet only 2.0% of visits; its average site gets 11% of the overall average — crowded and fragmented, a classic long-tail category: individual product scale can vary widely; check the category leader’s share too.
07AI Preferences in China, the US and India
In China, DeepSeek leads on the web and 豆包 leads in apps; in the US, ChatGPT leads on the web and Muse from Meta leads in apps; in India, ChatGPT leads on both web and apps.
China, the US is split: different products win on the web and in apps, using two distinct measures — while in India the same product leads both rankings. Evaluate web visits and mobile downloads separately; these metrics do not establish individual cross-device choices.
08AI Mobile Apps: Downloads and Spending
ChatGPT leads both downloads (80.0M over 30 days, October 2026) and in-app purchase revenue (about $306.4M, October 2026); Claude ranks second by revenue.
| App | Category | Top 10 by in-app purchase revenue |
|---|---|---|
| General AI | ||
| General AI | ||
| Video | ||
| General AI | ||
| Image & Design | ||
| Image & Design | ||
| Image & Design | ||
| Characters & Fun | ||
| Audio | ||
| Image & Design |
| App | Category | Top 10 by downloads |
|---|---|---|
| General AI | ||
| General AI | ||
| Video | ||
| General AI | ||
| General AI | ||
| Image & Design | ||
| General AI | ||
| Image & Design | ||
| Video | ||
| General AI |
In-app purchase revenue covers App Store and Google Play only (including subscriptions); it excludes Android app stores in mainland China and other revenue such as web subscriptions or API sales, so it is not total company revenue. AI app downloads chart →
Within the in-app purchase revenue sample: ChatGPT earns about 4.3x the runner-up Claude, and 53% of the total for the 10 apps shown.
Downloads are not revenue: Google Gemini, Meta AI, Dola are top-10 by downloads but not by in-app revenue — the difference is worth following, but rankings alone do not establish conversion or a business model.
5 of the 10 earners shown are image or video creation apps (CapCut, Canva, Meitu, …): on mobile, creative tools also have a place beyond AI assistants; persistence needs to be checked across multiple periods.
09What the Data Means
01Market: separate scale from breadth
The top ten products account for 68.3% of tracked visits. Benchmark new products against peers in the same category and size range, then look for sustained share gains. Web traffic alone does not establish retention, profit or valuation.
Compare global web rankings →02Categories: put growth in context
Read category growth alongside leader share and product counts to distinguish narrow gains from broader demand. Follow through to each product’s multi-month trend.
Explore categories →03Monetization: look beyond downloads
Downloads indicate distribution momentum; IAP revenue offers a partial view of mobile spending. Check the dates before comparing. Web subscriptions, enterprise contracts, API sales and mainland China’s Android stores need separate research; dividing the two rankings cannot yield a conversion rate.
Follow app downloads →FAQ
How much traffic do AI websites get?
According to AlphaMoat, the 17,643 AI websites with traffic data received 17.9B visits in October 2026 (+2.9% like-for-like month over month).
What is ChatGPT’s market share?
Among eight leading AI assistants (ChatGPT, Gemini, Claude, DeepSeek, Grok, Doubao, Kimi, Perplexity), ChatGPT held 56.1% of web visits in October 2026, compared with 73.1% in October 2025. This is share of web visits among these eight products, not share of users.
What is the most visited AI website?
ChatGPT was the most visited AI website in October 2026 (6.0B visits), 33.8% of all tracked visits.
Which AI tools are growing fastest?
Among AI products with 10M+ monthly visits, Muse grew fastest (+228.3% month over month) to 45.8M visits.
Which AI app makes the most money?
By App Store and Google Play in-app purchase revenue (including subscriptions), ChatGPT earned the most over the last 30 days as of October 2026 — about $306.4M. This excludes Android app stores in mainland China and other channels such as web subscriptions, so it is not total company revenue.
Where does this data come from, and how often is it updated?
Web figures are third-party estimates of monthly visits, with a product’s multiple domains consolidated. App figures are trailing 30-day download and in-app purchase revenue estimates. The page updates monthly; web and app data are each labelled with their own data month.
Do AI website visits represent user counts?
No. A person can visit repeatedly and use several AI products. Visits are not unique users, monthly active users or time spent. AI embedded in other products and API usage are not fully captured by standalone website traffic either.
Why can total AI traffic grow while most products do not?
Total visits are weighted by product size, while growth breadth counts products. This report pairs like-for-like aggregate growth with the percentage of products growing and the median product change to distinguish large-product gains from broader growth.
Can app revenue divided by downloads measure paid conversion?
No. In-app purchase revenue can come from existing users, downloads are not unique new users, and the figures may cover different months. Conversion requires activity, subscription and payment data for the same user cohort.
Methodology
- Visits are third-party estimates of monthly web visits (desktop and mobile web), best read for scale and trend; a product’s multiple domains are consolidated.
- Like-for-like month-over-month only compares products with data in both months; newly tracked products are excluded so that catalogue growth is not mistaken for traffic growth.
- Products whose visits fell below 30% of, or rose above 4x, the previous month are treated as suspected data breaks and excluded from growth figures and mover lists (1,353 this month, notably ArtCraft (115.0K → 26.6M), AlphaEngine (551.9K → 9.8M), Google Lens (6.6M → 1.8M)). Their raw figures remain on their own product pages.
- AI assistant share is each product’s share of web visits among eight fixed products (ChatGPT, Gemini, Claude, DeepSeek, Grok, Doubao, Kimi, Perplexity) — comparable month to month, but not total market share and not share of users.
- App downloads are trailing 30-day estimates. In-app purchase revenue covers App Store and Google Play only (including subscriptions); it excludes Android app stores in mainland China and other revenue such as web subscriptions or API sales, so it is not total company revenue.
- Standalone website visits do not fully capture API usage, embedded AI features or internal enterprise use; unmeasured activity does not mean zero demand.
- Web, app-download and app-revenue data may cover different months; each is labelled separately and they are never added together.