Revenue is still growing strongly, users are searching more than ever, yet stocks have plummeted, losing about $500 billion in capitalization in just one session. This article is about what's going on with the way Search makes money, why Google is willing to take a cut of its own business, and how it plans to make money back.

One quarter of growth but still heavily penalized

At first glance, Alphabet's quarter wasn't bad at all. Revenue was about $127.6 billion, up about 25% over the same period. The Cloud segment also more than doubled, thanks to companies pouring money into renting AI infrastructure. But according to CNBC, both revenue and profit were lower than analysts' expectations.

Investors worry about two things. First, search advertising, the largest source of revenue, grew only 12% and missed forecasts for the second consecutive quarter. Second, negative free cash flow is about $9 billion because Google still plans to spend nearly $200 billion on infrastructure this year, and says it will spend much more next year. As a result, shares fell about 13% on October 29, the worst session since Google went public in 2004.

The market values ​​Google by upcoming growth, not by current revenue, and Search's growth is slowing.

More searchers, but less money per search

Sundar Pichai says searches are growing the fastest in the past five years. Questions in AI Mode are three times longer than traditional searches, and many people also ask by voice, image or document. Yet Search revenue growth gradually decreased, from 19% in the first quarter to 17% in the second quarter, then to 12% in the third quarter.

When asked why, Philipp Schindler, Google's business director, said the biggest reason was foreign exchange rates, plus some retail and media groups spending less. But he also admitted that the sharp increase in AI Mode has dragged down revenue, because AI Mode is making less money. Google doesn't say how much less.

The problem is that a product that is better for the user is not necessarily better for the wallet. The old results page had dozens of links, each click was an opportunity to sell advertising. A neat AI answer will satisfy users immediately, they will click less, and there will be less space to place ads.

Advertisement

Change yourself before someone else does

The story of a company's new product taking away the revenue of its old product is not uncommon in the technology industry. Apple is an often mentioned example: the iPhone took away parts of the iPod, but Apple still did it, because if they didn't do it, someone else would.

Google is right there. CNBC writes that Search is under pressure from OpenAI's ChatGPT or Anthropic's Claude, users may be looking elsewhere for answers. If Google stands still and keeps the ten-link page to protect revenue, they risk losing users. So they choose to put AI into Search themselves, accepting that each search will temporarily earn less money, to maintain their position.

Google also made their thinking clear. Pichai compared the transition to mobile more than ten years ago: create the best experience first, make money later. They don't want to separate AI into a separate product, but combine AI Overviews, AI Mode and traditional search into one seamless experience. That means Google still wants to be the first place people go to when they need to know something, it's just that that place now knows how to answer instead of just giving links.

Advertisement

Conversation advertising: How Google plans to make money again

If AI Mode has fewer ads, Google's fix is ​​very straightforward: add more ads. Schindler said ad tests in AI Mode have shown positive results, and Google will expand ads in both AI Mode and AI Overviews to more people and more types of questions. Ads in AI Mode will be available for the first time in the international market, and in the US they are launching multi-pass ads, meaning ads can appear throughout a conversation, not just on the first question.

Google also offers some reassuring arguments. According to Schindler, most of the questions in AI Mode are new questions, do not replace old searches, and many questions are at the top of the funnel, when users are still researching and have not yet finalized a purchase. With products in AI Mode, click rates on sponsored and organic tiles were roughly equal. But Search Engine Land notes that this comparison only applies to product listings, and Google does not release specific numbers.

The risk here is trust. An AI answer that is suspected to have been paid for has its value immediately reduced. Google must consider how to advertise enough to shareholders, but be subtle enough so that users still believe the answer.

Money pouring into AI and the question of profit

The other half of the story is cost. If you want good AI, you have to build data centers, buy electricity, and make your own chips. Google is spending nearly $200 billion on infrastructure this year alone, and NBC News said the company has committed more than $1,000 billion to multi-year contracts on infrastructure, energy and space rental for AI.

The good news is that this investment is making money in the Cloud segment, where businesses rent AI infrastructure, including TPU chips made by Google. From Q2, Google began selling TPUs to external customers, competing with Nvidia's GPUs. The bad news is that investors may not be patient enough. They see the money spent very clearly, but the money earned from AI in Search is unclear.

What can I learn

Google's quarter shows that products change first, monetization follows, and the in-betweens are always difficult. Google is not the first time going through this. When moving to mobile, they were also worried that advertising on small screens would make less money.

Those who do advertising or online sales in Vietnam should pay attention to this. When advertising in AI Mode and AI Overviews expand internationally, the way users find products will change, and the questions will become longer and more consultative. Content that accurately answers customers' real needs will be more profitable than pages that are just stuffed with keywords.

As for Google, the big question is still whether they can turn the AI ​​Q&A experience into a money printing machine like the old ten-link page. What do you think, does advertising inserted into AI answers make you trust Google less?

Sources: Alphabet, CNBC, Search Engine Land, Adweek, NBC News, The Motley Fool