Much of Google’s business growth over the past two decades is based on a very simple process. You enter a query into your browser, and the search engine shows you a list of links in which some ads appear When someone clicks on one of them, the company pays. This dynamic has worked like clockwork for them over the years.
However, this process is twisted with the arrival of AI mode, the version of the search engine where Gemini solves queries through a text rather than a list of links. Users like it, but it leaves less room for ads and less reason to click. Google has admitted to its investors that AI infrastructure needs the most money.
Simple search engines pay the bills. To see Google’s real problems, you need to look at its third-quarter results. Alphabet’s turnover was $127,577 million, of which search contributed $63,350 million, almost half of the total. This number grew by 12% last quarter, but in the previous two quarters, it grew by 19% and 17%, respectively.
Meanwhile, investment in AI infrastructure and its development costs eat into cash flow at a frantic pace. The accounts reveal that Google spent 51,355 million on equipment and facilities in just three months, more than the 42,290 million generated by its activities. The item that supports all these investment efforts is advertising. The service division, under which these revenues operate, had an operating profit of 39,157 million, three times that of Cloud.

Schindler admitted that the AI mode had problems. In an entry on his corporate blog, Sundar Pichai boasted that searches are growing at the fastest rate in the last five years. An analyst then asked why revenue doesn’t do the same. Philipp Schindler, commercial director, then mentioned exchange rates, retail trade and media spending. But in the call where the results were presented to shareholders, the manager keyed in on the problem: “It should be noted that the strong increase in queries in AI mode was influenced by the low monetization that we were seeing in this mode.”
as indicated Search Engine JournalSchindler doesn’t mention how poorly the AI mode performs or how he compares it, but his defense is that many of these questions are new. as collected Register“Most AI mod questions are incremental.”
No links, no clicks. In classic search engines, advertisers pay for each click on their link. That is, Google charges when someone clicks on a sponsored link in its results list. However, AI Mode changes that charging system, as it answers questions directly instead of providing a list of links to charge. People who already have the answer on the screen have less reason to continue searching for it, and the ad loses its meaning without clicking checkout.
There is a piece of information that, although it doesn’t come from official Google data, helps measure this impact of the arrival of AI mode. As measured by a study Pew Research CenterWith one of those AI-generated summaries, only 8% of visits resulted in a click on the referral link. Without this summary, the percentage increases to 15%. Google has questioned the methodology of this study. Nevertheless, if the pattern repeats itself, Google runs the risk of killing its goose by laying its golden egg by stubbornly pushing for solutions that don’t lay eggs.
More ads in reply. The solution that Google seems to have found is to put more ads in the answers. On October 28, he brought AI mode ads to ten more countries. It will arrive in the UK, Brazil and Mexico in the fourth quarter. According to internal company data in the US, clicks on sponsored and organic products are already similar
Pichai defended to analysts that Google is “focused on creating the best user experience in search, followed by monetization”. He did the same on mobile phones a decade ago. While that promise is fulfilled, each AI mode query causes Google to lose a little money compared to classic search. And to pay for the investment in AI that drives that system continues to be provided by general search engines.
In Xataka | AI no longer pays for itself: even big businesses like Meta or Google aren’t big enough to finance it
Picture | Unsplash (Farmby)