🔍 Read the full analysis: OpenAI Says ChatGPT Gets One-Fifth Of Google’s Searches—What The Data Shows on ThorstenMeyerAI.com
Get movie nights delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
OpenAI consumer product head Adam Fry said ChatGPT receives roughly one-fifth as many “search-like” queries as Google, but OpenAI has not disclosed the comparison method or underlying figures. Independent estimates range from about 7.5% to 12% for modeled search-like activity, while traffic panels show ChatGPT sends far fewer visitors to websites than Google.
OpenAI consumer product head Adam Fry said November 2 that ChatGPT handles “roughly one-fifth” as many search-like queries as Google, a comparison that the company has not backed with a public methodology or underlying data. Independent estimates put ChatGPT’s search-like activity lower, while website traffic panels show a much larger gap between the companies in visits sent to sites.
Fry made the statement during a fireside chat with journalist Alex Kantrowitz at an OpenAI-hosted dinner in New York. Big Technology published the conversation on November 3, and The Verge described the gathering as a press event. Asked how OpenAI counts a search, Fry said: “Every time you ask a question and go search the web, that’s a search.” The Verge reported that simple requests, such as defining a word or asking for a basic how-to explanation, do not count under that description.
Fry did not identify which Google query figure OpenAI used as its comparison, the period covered, the countries included, or the data behind the calculation. The Verge reported that he did not specify the source of Google’s search-volume number. Google had not commented on the claim in the reports reviewed by ThorstenMeyerAI.com. The statement is confirmed as something Fry said; the one-fifth comparison itself cannot be independently verified from the details released.
Two outside estimates based on models put the ratio below Fry’s figure. Ahrefs estimated ChatGPT’s search-like activity at about 12% of Google’s in February 2026. seoClarity estimated about 7.5% on October 28, while describing its result as a modeled comparison rather than measured market share. These figures use different assumptions and are not direct measurements of the same activity.
Infographic 01 · ChatGPT as a share of Google
Questions or clicks? ChatGPT is a fifth of Google, or less than 2 percent
OpenAI says one fifth. Independent models say less. Traffic panels say far less, because a chat answer often ends without a visit.
Counting questionsHow many search-like questions?
Counting clicksHow many visits reach websites?
All values: ChatGPT as a percentage of Google on the same yardstick. OpenAI’s figure is a company claim (striped bar); the two models are estimates; the panels see only their own sites. Bars scaled to 20 %.
Sources: Big Technology, 3 Nov 2026 · Ahrefs, 11 Feb 2026 · seoClarity, 28 Oct 2026 · Ahrefs traffic panel, Oct 2026 · StatCounter, Oct 2025–Oct 2026.
Powered by Thorsten Meyer AI
Search Volume Is Not Website Traffic
The comparison matters to businesses that depend on people finding their websites through search. A user asking ChatGPT to look something up is not necessarily equivalent to a Google search that produces a list of links, and neither measure tells a publisher how many visitors arrive. Search-like queries and referral clicks are different metrics; treating them as interchangeable can make ChatGPT’s reach appear comparable to Google’s in ways that do not reflect traffic to websites.
In Ahrefs’ panel of 118,671 websites in October 2026, Google accounted for 24.74% of visits and ChatGPT for 0.36%. That is a traffic-panel observation, not a count of all searches or all web use. StatCounter’s comparison of clicks from search engines and chatbots to sites in its network found ChatGPT’s monthly share ranging from 0.08% to 0.41% between October 2025 and October 2026. Those figures describe referrals, not the number of questions people ask ChatGPT.
For companies, the data supports tracking both search rankings and citations or referrals from AI tools, rather than assuming one channel has replaced the other. For readers, an AI-generated answer that includes sources still depends on the quality of those sources. The available figures do not establish how often users click through, but they show why a query-volume claim alone cannot answer how much traffic AI tools deliver.
Popular right now
Sony WH-1000XM4C (M4 2nd Gen) Wireless Noise-Cancelling Headphones, Black
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
As an affiliate, we earn on qualifying purchases.
How the Estimates Were Built
The independent estimates cited by ThorstenMeyerAI.com are models with different definitions. Ahrefs’ Patrick Stox started with 2.5 billion ChatGPT prompts per day from OpenAI’s 2025 usage study, then judged that 65% represented activities people might use Google for. Stox said that 65% was his own judgment and could be lower depending on how “search” is defined. It is an estimate of potentially search-like prompts, not a direct count of ChatGPT web searches.
seoClarity’s October estimate extrapolated ChatGPT message volume using OpenAI user-growth figures. It counted the “seeking information” category and one-quarter of “practical guidance” as search-like, then compared that result with an assumed Google total of five trillion yearly searches plus 20% growth. The company called its result “a modeled comparison, not a measured market-share figure.” Because the assumptions differ from Ahrefs’ and OpenAI has not provided its own calculation, the percentages should not be read as a single consistent time series.
OpenAI has also reported that “seeking information” represented 20.2% of consumer ChatGPT messages in September 2026. That category is not the same as Fry’s definition of a search that triggers a web lookup. ThorstenMeyerAI.com notes that, by its rough arithmetic, a one-fifth comparison would require about six in ten ChatGPT messages to trigger a web search. That calculation depends on assumptions about Google’s query total and ChatGPT’s message volume; the published usage category alone does not confirm or disprove Fry’s claim.
Google’s own reported performance provides another measure of scale, though not a direct test of the ratio. Alphabet said Google query growth in the third quarter was its fastest in five years, while search revenue rose 12% to $63.35 billion. These company-reported figures indicate continued growth in Google Search, but they do not reveal the specific query count OpenAI used.
““A modeled comparison, not a measured market-share figure.””
— seoClarity
The Missing Comparison Method
The central unknown is what Google query total OpenAI used and whether it covered the same dates, countries, and types of activity as ChatGPT’s figure. OpenAI has not published the underlying counts or a methodology in the reporting cited here, and Google has not responded to the claim in those reports. Without those details, outsiders cannot reproduce the one-fifth calculation.
It is also unclear how OpenAI distinguishes a web search from other ChatGPT activity in the figure, beyond Fry’s description that a question must trigger a web lookup. The independent estimates use different models, categories, and assumptions, while the traffic panels measure referrals rather than searches. They can help frame the claim, but none directly verifies or disproves OpenAI’s number. The reported figures also cover different time windows, so they should not be compared as if they were simultaneous measurements.
More Detail Needed From OpenAI
The next step for checking the claim is for OpenAI to publish the query counts, time period, geographic scope, and definition behind its comparison, including the Google benchmark it used. Until then, the one-fifth figure remains a statement by an OpenAI executive rather than a publicly reproducible statistic. Further reporting could also establish whether Google responds or whether OpenAI provides more detail in a formal release.
For readers and site operators, the practical next step is to keep the measures separate: monitor AI citations and referrals alongside conventional search traffic, and treat modeled estimates as estimates. New traffic-panel data may show whether ChatGPT’s share of visits changes, but it would not by itself settle how many search-like questions people ask inside ChatGPT.
Key Questions
What did OpenAI say about ChatGPT and Google searches?
OpenAI consumer product head Adam Fry said on November 2, 2026, that ChatGPT’s volume of “search-like queries” was roughly one-fifth of Google’s query volume. OpenAI has not released the calculation or underlying data in the reporting cited here.
Can the one-fifth figure be independently verified?
Not from the information released so far. Fry described a search as a question that triggers a web lookup, but did not specify the Google benchmark, time period, countries, or supporting counts.
What do independent estimates say?
Ahrefs estimated about 12% in February 2026, while seoClarity estimated about 7.5% on October 28. Both results are models based on assumptions, not direct measurements, and their methods differ.
Does ChatGPT send as much website traffic as Google?
The cited traffic panels show a much smaller ChatGPT share of visits. Ahrefs recorded 0.36% of visits from ChatGPT and 24.74% from Google across its 118,671-site panel in October 2026. These are panel results and measure referrals, not search-query volume.
Why do the query and traffic numbers differ?
A person can ask ChatGPT to search without clicking through to a website, while a traffic panel counts visits that reach sites. Query volume measures activity inside a service; referral traffic measures visits sent out, so the figures answer different questions.
Source: ThorstenMeyerAI.com
