Is ChatGPT Now 20% of Google Search? What It Means for Backlinks

A large search-bar shape made of linked chain segments beside a smaller chat-bubble search bar, showing ChatGPT as a growing share of search

OpenAI now says ChatGPT handles a volume of search-like queries equal to about one-fifth of Google's. The claim came from ChatGPT consumer product lead Adam Fry, and it reframes a question link builders have been circling for a year: if a meaningful slice of "search" is moving into a chat box, where does a backlink have to live to still be seen? The number is OpenAI's own, and it carries real caveats — but the direction is hard to argue with, and it changes how off-Google authority pays off.

Summary

  • OpenAI says ChatGPT's search-like query volume is "roughly one-fifth" of Google's — about 20% — per consumer product lead Adam Fry, in an interview with Alex Kantrowitz, reported in early November 2026.
  • The Google baseline was not disclosed, so this is OpenAI's characterization of its own query volume, not an audited market-share figure. Google says its own search volume is at an all-time high.
  • A "search" here means a prompt that triggers a web lookup — not every ChatGPT message — so the figure counts search-like activity, not total prompts.
  • Scale is still growing fast: Fry said ChatGPT went from one billion to 1.2 billion weekly users in about two months.
  • The backlink takeaway: as more "search" happens inside AI answers, being cited or mentioned in those answers becomes a visibility channel of its own — and that is earned with real, in-content links and brand mentions on sites the model trusts, not with another thin page.

Frequently asked questions

Did OpenAI say ChatGPT is 20% of Google Search?

OpenAI's Adam Fry said ChatGPT's volume of search-like queries is "roughly one-fifth the volume of queries of Google" — about 20%. He said it in an interview with Alex Kantrowitz, reported by Search Engine Roundtable and 9to5Google in early November 2026. It is a statement about query volume, and the Google number it is measured against was not made public.

Is "20% of Google" a verified market share?

No. It is OpenAI's own figure for its search-like query volume, and the Google baseline it is compared with was not revealed. Market share and query volume are different things: for context, Bing sits at roughly 5% of search market share, which is not the same kind of measurement. Treat the 20% as a directional signal that AI search is now large, not as an audited statistic.

What counts as a "search" inside ChatGPT?

By OpenAI's description, a search is a prompt that leads ChatGPT to look something up on the web — "every time you ask a question and go search the web, that's a search." So the figure does not count every message people send; it counts the subset where ChatGPT actually runs a web search and pulls in sources.

Why does ChatGPT's search volume matter for backlinks?

When ChatGPT runs a web search, it grounds its answer in sources it pulls from the open web and may cite them. If a growing share of "search" happens this way, then being one of the sources an answer draws on becomes a real path to visibility — and that eligibility is built the same way it always was: relevant, in-content links and brand mentions on sites with genuine authority.

How is this different from ranking in Google?

Classic Google ranking competes for a position on a results page. Getting surfaced in a ChatGPT answer competes to be a trusted source the model grounds its response in. The underlying credibility signals overlap heavily — who links to you and writes about you across the open web — but the "slot" you are competing for is a citation inside an answer, not a blue link.

Does getting cited in ChatGPT send real traffic?

It can. When ChatGPT shows clickable source links, those citations become referral paths, and brands that appear as cited sources can earn clicks from them. The exact volume depends on how often your pages are surfaced and linked, which is why being a credible, frequently-referenced source matters more than any single placement.

What is the difference between a backlink and an AI citation?

A backlink is a link another site places to yours; it signals authority to search engines and to the models that learn from the web. An AI citation is when an answer engine names or links your page as a source for its answer. They are connected: the authority you build with backlinks and mentions is a large part of what makes a model treat your page as citation-worthy in the first place.

Should I stop optimizing for Google because of this?

No. Google still represents the large majority of search by OpenAI's own framing, and it says its volume is at a record high. The right read is "and," not "instead": keep earning rankings while also building the off-Google authority that makes you eligible for AI answers. The same relevant editorial links tend to do both jobs at once.

How does Rixot fit into AI-era link building?

Rixot places relevant, in-content editorial links and brand mentions on real publisher sites your buyers actually read. That is precisely the kind of signal answer engines lean on when they decide which sources to trust. You can filter donor sites by niche, authority and traffic in the Rixot marketplace.

What OpenAI actually said

The headline figure is specific, and so are its limits. In an interview with Alex Kantrowitz, OpenAI's ChatGPT consumer product lead Adam Fry said ChatGPT's "volume of search-like queries is now roughly one-fifth the volume of queries of Google." Search Engine Roundtable reported it on 4 November 2026; 9to5Google covered the same remarks a day earlier. Both are reporting the same claim, and both flag the same gap in it.

The number, stated plainly

One-fifth of Google's query volume is a big claim for a product that pushed into web search only recently. Fry framed it as search-like queries — the prompts where ChatGPT goes and looks something up — rather than every message a user sends. That distinction matters: it means the 20% describes the search-shaped slice of ChatGPT usage, which is the slice that competes most directly with a Google query and the slice where sources get pulled in and cited.

The caveat nobody should skip

The Google baseline was never disclosed. As 9to5Google put it, there is "a lot of wiggle room" because the source of Google's search volume wasn't revealed, and OpenAI's own definition of a "search" leaves room for interpretation. This is OpenAI describing its own numbers against an unstated comparison, not an independent audit. Barry Schwartz of Search Engine Roundtable added the obvious counterpoint: Google says its search volume is at an all-time high, which makes a clean "one-fifth" ratio harder to pin down. So the honest reading is that AI search is now large and growing — not that a verified 20% of the world's searches have moved.

The growth behind it

Context explains why OpenAI can even make the claim. Fry said ChatGPT went "from one billion to 1.2 in just the last two months" — weekly users, not search users — describing roughly 100 million new users in 30 days and another 100 million in the next 30. Whatever the exact search ratio, a product adding users at that pace is becoming a place people go to ask questions, and asking questions is what search is.

Why this is a channel story, not just a stat

It is easy to file "ChatGPT is 20% of Google" under AI hype and move on. But for anyone who buys or earns links, the useful part is not the percentage — it is what it implies about where attention now lives. A growing share of the moments when a person has a question and wants an answer is happening inside a chat interface that reads the web, summarizes it, and points to a handful of sources. That is a new surface where your brand is either present or absent.

For a decade, "being visible for a query" meant ranking a page. The page was the unit, and the backlink was the lever that moved it. In an answer-engine world, the unit shifts: you are competing to be one of the sources the answer is built from. The backlink is still the lever — models and search engines alike infer authority from who links to and mentions you — but the prize it buys is a citation in the answer rather than only a rank on a page.

Where the "search" actually goes

When ChatGPT answers a question with a web search, it retrieves pages, grounds its response in them, and can show them as sources. Brands that are frequently referenced across relevant, trustworthy sites are more likely to be in that retrieved set. Brands that exist only on their own domain, with little independent coverage, are easy for an answer engine to overlook — there is nothing on the open web vouching for them.

The open web is still the moat

You cannot out-Google Google, and you cannot out-OpenAI OpenAI. What you can do is build presence on the independent publishers, niche blogs and trade sites that cover your category, so that whichever engine is reading the web for an answer finds credible outside sources that mention you. That presence is exactly what a strong backlink profile has always been — it just has a second payoff now.

How brands get surfaced inside ChatGPT answers

Answer engines are grounded in sources they treat as credible, and credibility on the open web is still a function of who links to and writes about you. A topically relevant, in-content link on a real publisher does something a standalone page cannot: it tells every engine — Google's AI Mode, ChatGPT, and others — that an independent, trusted site found you worth referencing. When the search surface compresses to a few cited sources, that independent signal is what decides whether you are one of them.

Mentions and links, together

AI answers increasingly surface named brands, not just URLs. That makes the pairing of a brand mention and an in-content link in the same relevant article unusually efficient: the mention feeds the model's sense of who the notable players are, and the link reinforces the authority behind the name. One placement does both jobs, which is why relevance beats raw volume here.

Relevance over raw authority

A topically matched placement on a site your buyers read is worth more to an answer engine than a generic high-authority link with no subject connection. Engines are trying to assemble a credible, on-topic answer; a link that sits inside genuinely relevant content is a stronger signal of fit than a powerful but unrelated domain. Our per-link pricing is built around that trade-off — relevance first, not DR for its own sake.

What this changes for link building

The practical shift is from a single-engine mindset to a multi-surface one. If roughly a fifth of search-shaped questions are now being answered in a chat interface, then a link-building program aimed only at Google rankings is leaving the AI surface uncovered. The fix is not a new tactic so much as a wider target: earn the same relevant editorial coverage, but value it for two outcomes — rankings and citations — instead of one.

Diversify the sources that mention you

Different engines read the web differently, and they do not all cite the same domains. A link or mention that one engine ignores may be exactly what another one surfaces. Building authority across a spread of relevant, real-traffic publishers hedges against any single engine's quirks and widens the set of places you can be found as a source.

Measure citations, not only positions

If visibility now includes "appears as a source in an AI answer," then your measurement has to include it too. Track whether your brand shows up as a cited source for your key questions in ChatGPT and other engines, and treat that as a metric alongside rankings and referring domains. You cannot improve what you do not watch, and position-only reporting now misses part of the picture.

What to do about it

For brands that want to stay visible as search splits across surfaces, the response is concrete:

  • Earn relevant, in-content editorial links on real-traffic publishers in your niche, so answer engines have credible outside sources that reference you. Browse and filter donors by niche, authority and traffic in the Rixot marketplace.
  • Pair every link with a brand mention in the same article, since AI answers surface named brands, not just links.
  • Prioritize relevance over raw DR — a topically matched placement is a stronger citation signal than an unrelated high-authority one.
  • Diversify across publishers and engines, so you are eligible to be cited wherever a given answer is assembled.
  • Add citation tracking to your reporting, next to rankings and traffic, so AI visibility is a number you manage rather than a guess.

What this does not mean

It does not mean Google is finished or that search has already moved. OpenAI's own framing leaves Google as the large majority of query volume, and Google says its searches are at a record high. It also does not mean the 20% is a hard law — it is one company's description of its own numbers against an undisclosed baseline, and it will move as both products change. The durable lesson is structural, not numerical: more of the moments when people ask questions are being answered by systems that read the open web and cite a few sources. The brands with real, relevant authority off their own domain are the ones those systems keep finding. Build that authority with relevant links and mentions, and you stay visible whichever box the next question gets typed into.

Sources