Alphabet just reported the fastest growth in Google search queries in five years; in the same quarter, its search revenue grew at the slowest rate in 2026.
Roger covered the earnings call earlier this week on Search Engine Journal, and it’s a great overview of the high points. I’m diving deeper to make a specific point here.
Search & other revenue grew 12% to $63.35 billion in the third quarter, according to Alphabet’s earnings release, down from 19% in the first quarter and 17% in the second. That was roughly $1.6 billion below analyst expectations, and CNBC reported it was the second straight quarter search came in under estimates. On the earnings call, Google’s chief business officer, Philipp Schindler, named one of the causes directly: strong growth in AI Mode queries “had an impact because of the lower monetization we were seeing in AI Mode.”
For two years, the argument over whether AI answers cut into search has run on inferred numbers: click-through studies, panel estimates, and referral samples, many from vendors with something to sell. This quarter, the effect shows up in a counted number, Google’s own revenue line, with a Google executive naming AI Mode as a cause. That changes the quality of the evidence, even if it doesn’t finalize the argument.
A quick disclosure before going further: I run a company that measures visibility inside AI answers, so a story about money moving into those answers isn’t neutral ground for me.
The market reaction was severe. Alphabet shares fell about 13% on Thursday, the steepest one-day drop on record, erasing roughly half a trillion dollars in market value. Search wasn’t the only reason. Free cash flow came in at negative $9.07 billion against an expected shortfall of about $2.7 billion, and capital spending is still climbing. So I won’t claim investors sold because they rejected Google’s explanation for search. What I will claim is that the explanation needs a closer read than it’s getting, because the part that matters most is the part Google spent the least time on.
Every Reason Google Gave Is Temporary Except One
Schindler listed four drivers of the slowdown from the second quarter to the third. Foreign exchange came first, and he called it the single largest driver of the search revenue deceleration. Second, some retail advertisers in Asia-Pacific pulled back after the EU eliminated its de minimis exemption for low-value imports. Third, media and entertainment spending normalized after a second quarter inflated by the FIFA World Cup. Fourth, AI Mode queries grew quickly and earn less than traditional search. CFO Anat Ashkenazi added that Google has started lapping the search acceleration that began in the third quarter of 2025, which makes year-over-year comparisons harder for a few quarters.
Put those drivers on a timeline, and they sort themselves. A currency swing reverses or gets absorbed. The World Cup is over. A trade-rule change gets priced into advertiser budgets within a few quarters. Lapping a strong year is arithmetic that fades on its own. AI Mode is the only factor on the list that gets larger each quarter, because Google is actively growing it, and the lower monetization comes along with every query it adds.
Google’s Q3 explanation for slower search revenue named three temporary factors and one structural one: AI Mode queries are growing faster than they earn, and that gap widens with every quarter of adoption until ad yield inside the answer catches up.
‘Largest’ Ranks The Causes, But It Doesn’t Size Them
I understand the instinct to dismiss the currency explanation as convenient. I’d rather check it, because Alphabet publishes enough to do that. Total revenue grew 25% in the third quarter whether or not you adjust for currency; the constant-currency reconciliation tables in the release show it, and Ashkenazi described FX as a slight headwind on year-over-year growth. In the second quarter, revenue grew 24%, or 23% in constant currency, which means currency added about a point. Across the whole company, then, currency moved from a slight tailwind to a slight headwind, a swing of a point or two at most.
Ashkenazi also said FX had a more significant impact on a sequential basis, and the release bears that out: Revenue grew 6% from the second quarter as reported and 7% in constant currency. That’s a real effect. It’s also a one-point effect, and it applies to the company as a whole, not to search specifically.
Search growth fell five points over the same period. Google doesn’t publish a constant-currency figure for search alone, so this can’t be determined precisely. But unless search is far more exposed to currency than the rest of Alphabet, my estimate is that FX accounts for something like one to two of those five points. That makes it plausible as the single largest driver, and it still leaves most of the slowdown to the other factors combined.
That distinction matters. “Single largest driver” is a ranking. It tells you FX beat each of the other causes individually. It doesn’t tell you FX explains most of the decline, and it says nothing about which causes will still be there next year. The currency explanation is probably true. It just isn’t the answer to the question investors and practitioners are actually asking, which is whether the new way people search with Google earns as much as the old one.
What Counts As A Query Now
The other headline number deserves the same scrutiny. Sundar Pichai said Google’s year-over-year query growth rate is the highest it has been in five years. Google didn’t give the rate, and it didn’t say how a query is counted inside an AI Mode session. According to the call transcript, AI Mode queries average three times the length of traditional searches, and Google is now running multi-turn conversations in AI Mode, with ads designed for them. If each turn in a conversation counts as a query, query growth rises mechanically as conversations get longer, whether or not more people are searching for more things.
I’m not claiming that’s how Google counts. I’m saying the record query growth is a counted number with a definition we can’t see, and that kind of number should be held loosely until the definition is published. Revenue is much harder to redefine, which is why the revenue line is where the AI Mode story finally became visible.
Google Hasn’t Reached Parity, So The Ads Move Into The Answer
In my forecast earlier in October that the SERP stops being Google’s default between 2028 and 2030, I argued the trigger is ad yield. Google flips the default when an answer earns roughly what a results page earns per query, because before that point flipping costs money and after it keeping the old default does. One of the signals I said would push the timeline later was weak returns from AI Mode ad placements.
Q3 is Google confirming, in its own revenue, that parity hasn’t arrived. That fits the forecast’s explanation for why Google isn’t in a hurry to flip anything. It also tells you where Google’s effort goes next. Google isn’t going to slow AI Mode down to protect the results page; Pichai credits AI for the query growth, and the company is plainly committed to it. Instead, Google closes the yield gap by putting more advertising inside the answer.
The call laid that plan out. Schindler said Google is scaling ads within both AI Overviews and AI Mode to more users and more queries, taking AI Mode ads international for the first time, and introducing multi-turn ads in AI Mode in the U.S. He also said Google’s ad experiments have shown promising results, a phrase that will either become a reported metric in coming quarters or quietly disappear.
None of this surprised me. In August 2025, I described an ad system that builds the ad at query time inside the response, with advertisers supplying inputs rather than finished creative, and placement decided by how closely a brand aligns with the intent behind the prompt. Google shipped an early version at Google Marketing Live in May. Q3 explains why the rest of that design is likely to arrive faster: Google now has a quarterly revenue reason to finish building it.
Inside The Answer, Paid And Organic Now Compete On Equal Terms
One line from the call should get more attention from SEOs than it will. Schindler said click-through rates “are already similar between sponsored and organic product list items within AI Mode.” On a traditional results page, organic and paid listings occupied different territory with different click behavior, and practitioners built entire strategies around that separation. Inside an AI Mode answer, by Google’s own account, a sponsored product listing and an organic one draw clicks at about the same rate.
Put that next to the yield gap and my read is straightforward, though I want to be clear it’s my opinion and not something Google said. An answer has a small number of places where a brand can appear. As Google adds ads to close the gap, more of those places become available for purchase, and if sponsored and organic placements perform alike, advertisers have every reason to buy them. Organic visibility inside the answer doesn’t disappear, but it now competes for space with inventory that carries a direct revenue target.
Google also said the majority of AI Mode queries are incremental, and that many are upper-funnel queries carrying new commercial intent. That’s Google’s claim, and nobody outside Google can verify it. If it’s accurate, AI Mode is creating commercial moments that never existed on the results page, and the decision about which brands appear in them happens inside the answer, before anyone clicks anything. That’s a strong argument for measuring presence in the answer and not just clicks out of it.
The practical move is to establish a baseline now, before ad load inside AI Mode scales. For the queries that matter most in your category, record which answers cite you, which cite competitors, and which already carry sponsored units alongside the citations. A quarter from now, that record tells you whether paid placements are taking positions organic sources used to hold, and it gives you evidence to bring to budget conversations instead of an impression. Google will report its side of this story in aggregate. Nobody is going to report yours for you.
Reading The Next Report Like A Practitioner
This is where the difference between counted and inferred numbers stops being academic. Google’s revenue line is counted and audited. The query growth claim is counted but undefined. The incremental-query claim is asserted. Click-through studies of AI answers are inferred from samples. Each deserves a different level of trust, and the mistake I see most often is treating them as interchangeable because they all arrive as percentages. Measuring how brands actually show up inside AI answers, rather than inferring it from traffic, is the problem I built CitationIQ to address, and it becomes harder and more necessary as paid placements fill those answers.
Google’s fourth-quarter report arrives in early February, and Ashkenazi has already said FX will be a slightly larger headwind and lapping will cost a couple of points of growth. Q4 comes with the same cover built in. Here’s what I’ll be looking at instead:
- Whether Google quantifies its query growth or explains how turns in an AI Mode conversation are counted.
- Whether search revenue growth recovers once FX and lapping are set aside, which would suggest AI Mode monetization is catching up.
- Whether Google starts reporting revenue from ads in AI Overviews and AI Mode as its own figure, which I expect only once that figure looks good.
- Whether “promising results” from ad experiments turns into a reported metric or stays a phrase.
- Whether ad load inside AI Mode answers rises in your own categories, which you can check yourself without waiting for an earnings call.
The third-quarter report is the first time the gap between how much people use Google’s AI search and how much it earns appeared in Google’s own revenue, and Google’s response to that gap will be more advertising inside AI answers, where paid and organic placements now compete for the same small set of positions.
More Resources:
- SERPs Stop Being The Default Between 2028 & 2030
- How To Measure AI Search Visibility
- Google’s Q2: Precise Revenue Figures, Click Claims You Can’t Check
This post was originally published on Duane Forrester Decodes.
Featured Image: Remo_Designer/Shutterstock
