The State of AI Search in 2026: 50 Statistics, Every One Sourced

AI search in 2026 is large, fragmenting, and sending fewer clicks per answer than classic search. ChatGPT passed one billion weekly users in August 2026, Google says AI Overviews reach more than 2.5 billion people a month, and ChatGPT's share of AI referrals to B2B sites fell from 89% to 63% in eight months. The problem is that most AI search statistics in circulation have no traceable source, so what follows is 50 figures where the producing organisation, the period, and the method are all identifiable, each linked to where it was published.

TL;DR

• Every figure below names the organisation that produced it and the period it covers, with a link to the source. Where the method is weak, the entry says so.
• Scale: ChatGPT reached 1 billion weekly active users (OpenAI, August 2026), the Gemini app 950 million monthly users and Google's AI Mode more than 1 billion monthly users (Alphabet, July 2026).
• Concentration is collapsing: ChatGPT fell from 89.1% to 62.6% of B2B AI referral share in Goodie's GA4 panel, and from about 76% to 53% of generative AI web visits in Similarweb's panel.
• Clicks shrink when an AI answer appears: Pew measured 8% of Google visits producing a result click with an AI summary, against 15% without one.
• The most-quoted number in the field, the Princeton study's 40% visibility gain, is a maximum rather than an average and applies only once a source is already in context.
• Crawler policy moved again: Cloudflare's new defaults, which block Training and Agent crawlers on ad-displaying pages of new domains, took effect on September 15, 2026.
• A closing section lists widely circulated statistics that were excluded because no primary source exists for them.

How we compiled this

Last verified: September 2026. For the September 2026 refresh, every figure on this page was checked against the page where the producing organisation published it. Sources checked: company earnings releases and transcripts (Alphabet, Meta), vendor research posts (Semrush, Ahrefs, Similarweb, Goodie, XFunnel, Cloudflare), Pew Research Center, and the original arXiv paper. Figures that had been superseded by a newer disclosure were replaced, and the date on each entry was corrected to when the data was actually published.

Three rules were applied. A statistic is included only if the organisation that produced it can be named. Where a figure comes from a company measuring its own property, that is stated, because self-reported single-site data is directionally useful and not generalisable. Where a figure is an estimate by a third party rather than a disclosure by the operator, it is labelled as such. Two earlier entries, user-count estimates for Perplexity and Claude with no named producer, failed the first rule and were replaced with Similarweb's measured traffic shares. A BrightEdge coverage-growth figure was removed because it could not be traced to a BrightEdge publication.

Nothing here is first-party data from me. This page is an index of other people's work, which is exactly what it should be. For the tooling side of the market, the vendors, engines, and measurement guides live on GEO Compass.

Audience and adoption

#	Statistic	Source and period
1	ChatGPT reached 1 billion weekly active users, up from 900 million in February 2026	OpenAI, announced August 6, 2026
2	Google AI Overviews reached more than 2.5 billion monthly users	Google, Sundar Pichai at Google I/O, May 2026
3	Google's AI Mode passed 1 billion monthly active users, less than a year after its global expansion	Alphabet, Q2 2026 earnings remarks, July 22, 2026
4	The Gemini app reached 950 million monthly active users, with daily actives tripling year over year	Alphabet, Q2 2026 earnings remarks, July 22, 2026
5	Meta AI passed roughly 1 billion monthly active users across Meta's apps. Meta has not given an updated assistant user count since; its Q2 2026 earnings call reports engagement growth, not users	Meta, May 2025
6	ChatGPT's share of worldwide generative AI website visits fell from about 76% to roughly 53%	Similarweb, June 2025 to May 2026
7	Gemini's share of generative AI website visits rose from under 9% to roughly 27 to 28%	Similarweb, June 2025 to May 2026
8	Claude's share of generative AI website visits rose from about 2% to roughly 9%	Similarweb, June 2025 to May 2026
9	On iOS, ChatGPT ranked first, Claude second, and Gemini third by free downloads	SensorTower, early May 2026
10	On iOS top-grossing, ChatGPT ranked first, Claude second, Grok third	SensorTower, early May 2026

Entries 6 to 8 measure web visits only, so they understate engines whose usage is concentrated in mobile apps or inside other products. That is why Gemini's web share and its 950 million app users tell different stories, and why neither is a measure of "search market share".

Referral share and traffic

Statistics 11 through 20 come from Goodie's 2026 AI Search Traffic Report (published May 21, 2026), built on a GA4 panel of anonymised B2B brand sites measured between August 2025 and May 2026. Wave 1 covered 41 brand sites and 2,802,519 AI sessions. It is a B2B panel, so it does not describe consumer categories.

#	Engine	May-Aug 2025	Mar-Apr 2026
11	ChatGPT share of AI referrals	89.1%	62.6%
12	Claude share of AI referrals	1.4%	18.5%
13	Gemini share of AI referrals	2.4%	10.6%
14	Perplexity share of AI referrals	3.1%	7.3%
15	Copilot share of AI referrals	3.2%	4.0%

16. Of sixteen brands with sufficient Claude volume, fourteen moved upward between waves. The report gives a one-tailed binomial p-value of 0.0021. (Goodie, 2026)

17. Claude held 1.29% of measured platform visits and produced 18.0% of B2B referrals, roughly fourteen times its visit share. (Goodie, Jan to Apr 2026)

18. Gemini held 29.0% of platform visits and produced 10.3% of referrals, underperforming its size. (Goodie, Jan to Apr 2026)

19. Perplexity held 1.85% of visits and produced 7.1% of referrals. (Goodie, Jan to Apr 2026)

20. Grok held 3.5% of platform visits and produced effectively no measurable B2B referral output. (Goodie, Jan to Apr 2026)

All referral figures undercount AI's real reach. Users copy URLs out of answers or type the domain later, which registers as direct traffic, and native mobile apps frequently strip the referrer. Google bundles its AI surfaces into organic attribution, so an AI Overviews visit is indistinguishable from a blue-link visit in Search Console.

Engagement and conversion

21. AI-sourced sessions averaged 58.5 seconds of engagement with a 67.8% engagement rate. (Goodie, Feb to May 2026)

22. Google organic averaged 44.2 seconds and a 61.8% engagement rate over the same window, making AI traffic roughly 30% more engaged. (Goodie, Feb to May 2026)

23. Bing organic averaged 47.9 seconds and a 58.4% engagement rate. (Goodie, Feb to May 2026)

24. The average AI search visitor was 4.4 times as valuable as the average traditional organic visitor, based on conversion rate. (Semrush AI search study, 2025)

25. On Ahrefs' own website, AI search visitors converted at 23 times the rate of organic: 0.5% of traffic over 30 days produced 12.1% of signups. (Ahrefs, self-reported single-site data, June 2025)

Entries 24 and 25 differ by a factor of five because they measure different things. Semrush aggregates across many properties; Ahrefs reports one site with an unusually technical audience. Both support the direction, neither supports a precise multiplier for your business.

Zero-click and AI Overviews

26. Google AI Overviews were triggered for 15.69% of queries in November 2025, down from a peak of 24.61% in July 2025 and up from 6.49% in January 2025. (Semrush AI Overviews study)

27. Google users clicked a traditional search result on 8% of visits when an AI summary appeared, against 15% when none did. (Pew Research Center, 900 U.S. adults, March 2025 browsing data)

28. Users clicked a source link cited inside the AI summary on just 1% of visits. (Pew Research Center, March 2025)

29. 18% of the 68,879 Google searches in the panel produced an AI summary, and 26% of visits with a summary ended the browsing session, against 16% without one. (Pew Research Center, March 2025)

30. Google's AI Mode shows a zero-click rate of approximately 93%, based on a clickstream analysis of about 69 million U.S. desktop search sessions. (Semrush, May to July 2025)

31. Pages cited by ChatGPT ranked at position 21 or worse in conventional Google results almost 90% of the time. (Semrush, July 2025)

32. The share of U.S. ChatGPT prompts whose answers included a source citation rose from about 1.6% to roughly 6.8%. (Similarweb, June 2025 to May 2026)

Statistic 31 is the most strategically important entry on this page and the least quoted. It says AI citation is not a reward for ranking. A retrieval system reaching past the first two pages of Google means conventional rank is a weak predictor of citation, which is precisely why passage structure matters more than domain position. Statistic 32 is the counterweight: most ChatGPT answers still cite nothing, so being citable only pays off on the prompts where the engine decides to ground its answer.

Crawler traffic and blocking

33. AI crawlers accounted for 20.3% of verified bot traffic, with AI-search bots adding a further 6.5%, putting AI-related activity at roughly 26.7% of verified bot traffic. (Cloudflare Radar, May 2026 snapshot)

34. GPTBot accounted for 11.48% of AI bot HTTP requests. (Cloudflare Radar, May 2026 snapshot)

35. ClaudeBot accounted for 9.73% of AI bot HTTP requests. (Cloudflare Radar, May 2026 snapshot)

36. Training-purpose crawling accounted for nearly 80% of AI bot traffic, while user-action and undeclared purposes together stayed under 5%. (Cloudflare, August 2025)

37. In news and publications, GPTBot represented 17.4% of AI crawler activity and ChatGPT-User 14.9%. (Cloudflare, August 2025)

#	Operator	Crawl-to-refer, all sites	Computer and electronics
38	Anthropic	50,000:1	8,800:1
39	OpenAI	887:1	401.7:1
40	Perplexity	118:1	88:1

(Cloudflare, first week of August 2025.)

41. GPTBot appeared in 5.52% of robots.txt DISALLOW rules, the most-blocked AI crawler, ahead of CCBot at 5.08%, ClaudeBot at 4.88%, and Google-Extended at 4.44%. (Q1 2026 analysis of robots.txt across Cloudflare's network, Cloudflare Radar)

42. 35.7% of the top 1,000 websites blocked GPTBot. (Originality.ai, 2024)

43. The GPTBot block rate across a sample of 140 million websites was 5.89%. (Ahrefs, 2025)

Statistics 42 and 43 are not in conflict. Large publishers block at far higher rates than the long tail, which is why a single "how many sites block AI crawlers" number is close to meaningless without knowing which sites.

44. Cloudflare changed its default to block AI crawlers for new domains, announced on July 1, 2025. (Cloudflare)

45. Since September 15, 2026, new domains onboarding to Cloudflare get defaults across three AI traffic categories: Training and Agent blocked on ad-displaying pages, Search allowed. Mixed-purpose crawlers follow the most restrictive rule that applies. (Cloudflare, announced July 1, 2026)

46. 17% of Cloudflare sites enable some mechanism to block AI training, while fewer than 1% block search bots. (Cloudflare, September 15, 2026)

The practical reading, set out in the AI crawler reference table: blocking training is a defensible business decision, blocking retrieval removes you from live answers, and almost nobody makes the second choice deliberately.

Content signals and citation behaviour

47. Adding quotations, statistics, or citations each raised visibility in generative engine answers by roughly 25 to 40 percent, with the three strongest methods achieving 30 to 40 percent relative improvement on a position-adjusted word count metric. Tested across approximately 10,000 queries and nine datasets. (Aggarwal et al., "GEO: Generative Engine Optimization", KDD 2024)

The important caveat, which almost every guide quoting "40%" omits: the study measured how well a source competes once it is already in a fixed retrieval context. It establishes nothing about whether you get retrieved in the first place, and nothing about durable traffic.

48. URLs cited by AI assistants averaged 1,064 days old against 1,432 days for URLs in organic results, making cited content 25.7% fresher, across 16.975 million citations. Google's AI Overviews were the exception and leaned toward older pages. (Ahrefs, July 2025)

49. Product content accounted for 46 to 70% of AI citations, reaching up to 70% in B2B, across a study of 768,000 citations from ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews. (XFunnel, April 2025)

50. In the same study, blogs earned 3 to 6% of AI citations and PR material under 2%. (XFunnel, April 2025)

Statistics 49 and 50 come from a vendor measuring citation behaviour across its own query set, so treat the ranges as directional. The direction is still the most actionable finding in this list: most B2B teams spend the bulk of their content budget on the format AI engines cite least. The operator-side argument is in GEO's product-content advantage.

llms.txt adoption

Three supplementary figures, included because llms.txt is the most-recommended and least-evidenced tactic in the field.

• llms.txt files grew from 4,088 in June 2025 to 36,120 in May 2026 across a monitored set of more than 3 million websites, an 8.8x rise. (Originality.ai)
• Of roughly 38,000 domains with a valid llms.txt file, 97% received zero requests for it in May 2026, in a server-log analysis of 137,000 domains. (Ahrefs, June 2026)
• Among files that were fetched, SEO audit tools accounted for 21.7% of requests, GPTBot 4.51%, ClaudeBot 0.80%, and AI retrieval bots in aggregate 1.1%. (Ahrefs, May 2026)

The longer version of what this means is in the llms.txt guide. The short version: publishing a file is cheap, but the logs say AI retrieval bots barely read it yet.

What the numbers say has settled, and what has not

Taken together, the data supports a small number of planning conclusions for 2026.

• "AI search" is several channels, not one. Referral share (entries 11 to 15) and web traffic share (entries 6 to 8) have both fragmented across ChatGPT, Gemini, Claude, Perplexity, Copilot, and Google's own AI Overviews and AI Mode. Each grounds and cites differently; the engine profiles on GEO Compass document the differences. A program that measures AI search as one channel is mis-segmenting its data.
• Citation share is the metric that survives. Clicks fall when answers appear (entries 27 to 30), so traffic alone undercounts influence. Citation share on a fixed query set, tracked consistently, is the measure that ties content to AI visibility. The method is in the measurement guide, and the board-level case is in citation share for cybersecurity CMOs.
• Product content beats blog content in B2B. Documentation, integration, pricing, security, and methodology pages are primary citation assets (entries 49 and 50), not supporting material.
• Gated content is invisible. A crawler cannot cite what sits behind a form. The argument is made in full in why gated whitepapers are killing your AI visibility.

Four things have not settled.

• Measurement methodology. Two visibility vendors will report different citation-share numbers for the same domain. Pick one method, run it consistently, and trust the trend over the absolute number. The GEO Compass vendor matrix compares the options.
• AI shopping assistants. Shopping surfaces in ChatGPT, Perplexity, Google, and Amazon change month to month. Clean Product schema and merchant feeds pay off whichever one wins.
• llms.txt. Adoption is rising fast while retrieval-bot usage stays near zero (see above).
• The publisher and AI economic settlement. Licensing deals (Axel Springer with OpenAI, Reuters with Meta), litigation (The New York Times against OpenAI and Microsoft), and network-level controls like Cloudflare's crawler defaults (entries 44 to 46) are all running at once. None is dominant.

What was excluded, and why

The following circulate widely and were left out. Each is quoted constantly, and none has a published method behind it.

• "Passages of 40 to 75 words are cited 3.1x more often." Attributed to a 2025 analysis of 10,000 AI citations that has never been published.
• "Pages with tables are cited 4.2x more often than prose." Same origin, same problem. The structural argument for tables is sound without it.
• "FAQ sections get cited 200% more often" and "structured data increases citation probability by 195%." Traced to Medium posts citing other Medium posts.
• "Content over 1,500 words has 180% higher citation probability." Same chain.
• "Only 12% of websites are optimised for AI search engines." No definition of optimised, no sample, no measurement.
• A fixed share of B2B buyers (40%, 60 to 80%) "now research vendors primarily in AI assistants". These figures appear without a named survey, sample, or question wording. Quote a buyer survey only when you can cite all three.
• Any figure describing "AI search market share" as a share of total search. The denominators are not comparable across sources and the resulting numbers vary by an order of magnitude.

If you are building a slide from AI search statistics, the exclusion list is more valuable than the inclusion list. Quoting an unsourced multiplier to a CFO who then asks for the study is a bad afternoon.

For what to do with these numbers, see what actually changes between GEO and SEO and the extraction patterns.

Frequently Asked Questions

How many people use AI search in 2026?

Hundreds of millions to billions, depending on the surface. ChatGPT reached 1 billion weekly active users in August 2026 (OpenAI). Google reports more than 2.5 billion monthly users for AI Overviews, more than 1 billion for AI Mode, and 950 million for the Gemini app. These counts overlap heavily and are not additive.

What percentage of AI referral traffic comes from ChatGPT?

62.6% of B2B AI referrals in Goodie's GA4 brand panel for March to April 2026, down from 89.1% in the May to August 2025 wave. Claude was second at 18.5%. The figure covers B2B sites only and excludes Google AI Overviews, where the answer usually ends the session without a referral.

Do AI search visitors convert better than organic?

The available evidence says yes, with wide variance in the size of the effect. Semrush measured AI search visitors as 4.4 times as valuable as organic, based on conversion rate across many properties. Ahrefs reported 23x on its own site, where 0.5% of traffic drove 12.1% of signups. Use the direction, not the multiplier.

Do AI Overviews reduce clicks to websites?

Yes, in the best independent measurement available. Pew found Google users clicked a traditional result on 8% of visits with an AI summary against 15% without one, and clicked a link inside the summary on only 1% of visits. Semrush measured about 93% of Google AI Mode sessions ending without a click.

Does ranking on Google predict AI citation?

Weakly. Semrush found that pages cited by ChatGPT ranked at position 21 or worse in conventional results almost 90% of the time. Retrieval systems reach well past the first page, which is why passage-level structure matters more than domain position.

How many websites block AI crawlers?

It depends entirely on which websites and which crawler purpose. Originality.ai found 35.7% of the top 1,000 sites blocking GPTBot in 2024, while Ahrefs measured a 5.89% block rate across 140 million sites. On Cloudflare, 17% of sites block AI training but fewer than 1% block search bots (September 2026). Any single headline number is misleading without the sample.