DEV Community

Cover image for Google Rebuilt Its Own Search Business — Where AI-Era Search Ads Actually Make Money
s3atoshi_leading_ai
s3atoshi_leading_ai

Posted on

Google Rebuilt Its Own Search Business — Where AI-Era Search Ads Actually Make Money

#ai #marketing #google #seo

TL;DR

  • The "Google is destroying the web that feeds it" chorus rests on two premises — AI shrinks search and links disappear. Google's own data breaks both: queries hit an all-time high, and Google explicitly commits to keeping links.
  • Google didn't make a foolish call. It read the irreversible shift in search behavior earlier than anyone and re-architected its monetization base — moving from selling clicks (an intermediate good) to monetizing attention and transactions (the final goods) on its own surface.
  • That's why click erosion (real) and double-digit ad-revenue growth coexist: Search & other revenue grew +19% YoY to $60.4B in Q1 2026, with Google's CBO stating AI Overviews/AI Mode monetize "at approximately the same rate" as traditional search.
  • The real fault line for creators and businesses: "content that gets clicked" vs. "structure that gets cited." The former loses value the moment AI summarizes it. The latter gains value by being summarized.

Prologue: The discomfort with the "search is dead" chorus

At Google I/O 2026 on May 19, Liz Reid, VP of Search, called the search box's redesign its "biggest upgrade in over 25 years."

The familiar search bar is being replaced by an intelligent search box that dynamically expands to fit long, conversational questions, and AI Mode is now built into the default search experience with Gemini 3.5 Flash as the default model (rolling out from May 20 across all countries and languages where AI Mode is available) [1].

In response, media and social feeds keep repeating the same lines:

  • "Google Search is over."
  • "If AI answers the question, no one clicks a link anymore."
  • "Google is destroying the very web that feeds it."

It sounds plausible. But one fact refuses to fit. The numbers Google itself published point in the opposite direction: search queries hit an all-time high [1][4].

If AI is killing search, why is search volume at its largest ever?

This piece dissects the now-popular "chicken-and-egg problem = Google's self-destruction" narrative — structurally, using primary sources and Google's already-published earnings data.

Let me state the conclusion up front. Google is not making a foolish management decision. It is the most rational player in the room — the first to grasp, at a structural level, the irreversible change in search behavior, and it has re-bet its entire monetization base on it. I want to frame this redesign not as "destruction" but as a re-architecting of both supply structure and revenue structure.

Chapter 1: Dissecting the narrative — what is the "chicken-and-egg" argument?

First, let me state the conventional "chicken-and-egg" argument in its strongest form.

The skeleton goes like this. The web is built by individuals who blog and outlets that do investigative journalism. They live on clicks — ad revenue, subscription funnels, product sales. All of it depends on users clicking a link and arriving at the site. But Google scrapes those same articles, packs a summary into the AI Overview, and stops sending users to the source. Revenue evaporates, the incentive to create content evaporates, and the creators vanish. Which leaves one final question:

When the creators are gone, where exactly does Google pull its answers from?

Google is strangling the very hen that lays the eggs it eats — that's the picture.

The argument carries one further implication:

"If Google is this irrational, a gap opens for a competitor. A contrarian player appears that returns raw organic links instead of AI answers, and the people who miss the old Google migrate there."

It's internally coherent. But internal coherence and factual truth are different things. Let me take it apart one premise at a time, each with a primary source.

Chapter 2: Both premises of the narrative collapse

The narrative rests on two premises: "AI shrinks search" and "links disappear."

Premise ① "Search shrinks" — refuted by Google's own data.

Liz Reid's official announcement (May 19) states that AI Mode crossed 1 billion monthly users within a year of launch, queries continue to double quarter over quarter, and people are "searching more than ever before" [1]. Per Sundar Pichai's keynote, the monthly tokens Google processes ballooned roughly 7x in one year [2].

People are asking AI longer, more complex questions.

Here I'll draw the line carefully. "Queries at an all-time high" is Google's own claim; there is currently no independent third-party audit of the absolute global query count. So I treat it as "a fact Google has published." Even so, the direction — AI expanding rather than shrinking search behavior — does not move.

Premise ② "Links disappear" — Google explicitly denies it.

Google officially states that users will "continue to get a range of results from Search, just like you do today," and that "links and supporting articles get even more relevant" as context deepens [1]. This isn't lip service. Eight days after I/O (May 27), Google posted "new ways to find your favorite sources and original content in AI Search," strengthening the visibility of original content with Preferred Sources and a "Highly Cited" badge [3].

Google posted "new ways to find your favorite sources and original content in AI Search,"

Why does Google deliberately keep — and reinforce — links and original content?

Because the quality of the AI's answer depends entirely on the quality of the source content. Google isn't strangling the hen; it's renovating the henhouse. The issue is not "links disappearing." It's "the meaning of the link changing."

Chapter 3: Redefining search — from a "device that answers" to a "device that completes"

Now the central issue: what is Google actually doing?

Draw a single line through the features unveiled at I/O 2026 and it reads: search is migrating from "a relay point to links" to "an OS that completes tasks."

First, Search agents

Background agents crawl blogs, news, and social 24/7 and notify you of changes in your areas of interest. This is the shift from "the user comes to search" to "an agent keeps watch on your behalf" [1]. Don't overlook what they crawl — blogs and news. Demand for content doesn't vanish; the route changes.

Second, Generative UI

Search writes code in response to your question and generates visualizations and mini-apps on the spot. The result page is no longer "a list of links" but "a tool built just for you."

Third, Universal Cart and UCP (Universal Commerce Protocol)

Spanning Search, Gemini, YouTube, and Gmail, agents handle everything from adding an item to checkout [7][8]. Search moves from "the entrance to shopping" to "the device that completes the purchase."

The structure running through all three layers is the true meaning of the 7x token figure. People are no longer typing words; they're handing entire complex tasks to AI. And this move toward a "device that completes" is, at the same time, Google's new revenue mechanism. The next chapter dissects that mechanism.

Chapter 4: Redesigning monetization — from "selling clicks" to "monetizing the answer and the transaction"

The narrative cannot answer one question: "If clicks fall, how does Google grow ad revenue?"

Google has already answered with a concrete mechanism — Google Marketing Live 2026, May 20 [5]. Monetization is redesigned across three layers.

Layer 1: Ads move from "next to the link" to "inside the answer"

This is the core. Even when clicks stop flowing to external sites, placing the ad unit inside the AI answer monetizes attention where it actually gathers. Google began testing two new formats for AI Mode [5]. Conversational Discovery ads have Gemini generate ad creative in real time per query, answering the user's specific question directly inside AI Mode. Highlighted Answers place ads directly inside the recommendation list AI Mode generates. Both carry a "Sponsored" label — the ad is no longer "a banner beside the link" but "the answer itself / the recommendation itself." Existing text and Shopping ads are already served above and below AI Overviews across 200+ markets [6]. This is what CBO Schindler (below) means by "monetizing at the same rate as traditional search."

Layer 2: The long tail that could never be monetized finally becomes inventory

The average AI Mode search is 3x longer than a traditional query [5]. Long, conversational queries are high-intent but were hard to monetize under old keyword-bidding logic. By having Gemini generate creative per query, this long-tail region becomes ad inventory for the first time. "7x tokens / all-time-high queries" means an expansion of the monetizable surface area itself. The more search grows, the more questions can be monetized.

Layer 3: From "selling clicks" to "completing the purchase"

The UCP / Universal Cart from Chapter 3 becomes the revenue mechanism here. Google used to send traffic to external sites and charge for ads (the referral model). In the new structure, users complete a purchase in a few taps on Google's surface, using Google Pay. The merchant of record stays on the retail side, but the payment happens on Google's surface [7][8].

One logic runs through all three layers: Google stopped selling the intermediate good — the click — and moved to the side that monetizes the final goods, attention and transactions, on its own surface. This is why click erosion (the destruction is real) and double-digit ad-revenue growth (see Epilogue) coexist.

Chapter 5: So what actually breaks — the layer that is destroyed vs. the layer that is redesigned

Now let's face the narrative head-on. If you ask, "So nothing is breaking?" — the answer is No. Destruction is real. There is hard data.

The narrative is half right: traffic destruction is genuinely happening. But here is the structural fork. What's breaking is a specific layer — not "the whole web," and not "Google's self-destruction."

The decisive point: this "destruction of a specific layer" and the chicken-and-egg argument's "Google self-destruction" are entirely different stories. The former is the culling of a specific business model — and it's a fact. The latter is the claim that Google is starving its own food supply — and the earnings data in the Epilogue refutes it.

In Japan, another dynamic is at work. On April 20, 2026, the Japan Newspaper Publishers & Editors Association issued a statement that AI content usage could constitute "abuse of a superior bargaining position" under antitrust law [11], and Nikkei and Asahi sued Perplexity in the Tokyo District Court [12]. The "layer being destroyed" isn't disappearing quietly — it's taking a seat at the table of law and negotiation. That, too, is part of the "redesign of supply structure."

Chapter 6: What the AI era demands — from "content that gets clicked" to "structure that gets cited"

What separates the layer that's destroyed from the layer that's redesigned? In a phrase: do you build "content that gets clicked," or "structure that gets cited"? This is the value watershed of the AI-search era.

Traditional SEO optimized for "rank high in results and get the click." But in AI search, the user reads the AI's answer before clicking a link. The question becomes: can your content become the citation source for the AI's answer? The emerging fields called GEO (Generative Engine Optimization) and AEO are already being formalized academically.

Here is the enormous opportunity the narrative missed. A substantial share of the URLs cited as sources in AI Overviews are pulled up from pages that ranked outside the top tier in traditional organic search [10]. AI search resets the incumbents who dominated the top via "domain authority," and opens a path for upsets to content with "structure worth citing as an answer."

I'm proving this myself. The open-source books I publish are designed as "structure to be cited by AI." Among them, my open-source book dissecting Palantir's Ontology Strategy reached #1 on Google globally within weeks of publication — ranking above Palantir's own official Ontology page. Not domain authority, but depth of structure judged worth citing.

🔗 https://github.com/Leading-AI-IO/palantir-ontology-strategy

"Content that gets clicked" loses its value the instant AI summarizes it. "Structure that gets cited" gains value by being summarized. This asymmetry is the core strategy of the AI-search era.

Epilogue: Google is excellent — the rationality of betting on the irreversibility of search behavior

Finally, let's settle the biggest premise — the emotional core of the chicken-and-egg argument: "Google is a foolish self-destroyer that pivoted to AI search even at the cost of damaging ad revenue." The earnings refute it.

While rolling out AI search across the board, search ad revenue is not falling — it's growing double digits. Google CBO Philipp Schindler stated plainly that AI Overviews and AI Mode "monetize at approximately the same rate" as traditional search [4]. In Q1 2026, Search & other revenue grew +19% YoY to $60.4B [4].

The three monetization layers from Chapter 4 sit behind this number. Ads move inside the answer, long-tail queries become inventory, and purchase completion becomes transaction revenue. Looking at the latest earnings, there is no objective data supporting the narrative that Google "pivoted at the cost of damaging ad revenue." Google is expanding the monetizable surface area itself.

History gives us déjà vu. The music industry was said to "end" with Napster. It didn't. Spotify redesigned the supply structure and built a new economy called streaming. What got destroyed was "the business model of selling CDs," not "music." The same thing is happening now.

Google is not destroying the web that feeds it. It understood, earlier than anyone, that search behavior changed irreversibly — and it is re-architecting its entire monetization base from a device that makes you click links into a device that completes answers and tasks. On top of that, it is redesigning the rules of a new supply structure: who gets cited, and who gets culled. This is not self-destruction. It's a rational bet on an irreversible change.

So, within that new structure — can your content, can your business, stand on the side that "gets cited" and "monetizes"? Or do you end up on the side that "gets summarized and disappears"?

Asking that question, and designing for it, is the heart of strategy in the AI era.

📚 Related open-source books (CC BY 4.0, free)

I publish a 17+ title open-source library on AI industry structure, all free under CC BY 4.0:

Full library 👉 https://github.com/Leading-AI-IO

References (primary sources)

  1. Google, "Google Search's I/O 2026 updates: AI agents and more" (Liz Reid, 2026/5/19) — https://blog.google/products-and-platforms/products/search/search-io-2026/
  2. Google, "Sundar Pichai's I/O 2026 opening keynote" (monthly tokens 9.7T → 480T → 3,200T, ~7x) — https://blog.google/innovation-and-ai/sundar-pichai-io-2026/
  3. Google, "New ways to find your favorite sources and original content in AI Search" (2026/5/27) — https://blog.google/products-and-platforms/products/search/original-high-quality-content-search/
  4. Alphabet Inc. Q1 2026 earnings (SEC Form 8-K, Exhibit 99.1 — Search & other +19% / $60.4B / Schindler remarks) — https://www.sec.gov/Archives/edgar/data/0001652044/000165204426000043/googexhibit991q12026.htm
  5. Google Marketing Live 2026 new ad formats (Conversational Discovery ads / Highlighted Answers / 3x query length — Search Engine Land) — https://searchengineland.com/google-tests-new-conversational-ad-formats-in-ai-mode-and-search-478115
  6. Google Ads Help, "About ads and AI Overviews" (ads in AI Overviews, 200+ markets) — https://support.google.com/google-ads/answer/16297775
  7. Google Developers Blog, "Under the Hood: Universal Commerce Protocol (UCP)" — https://developers.googleblog.com/under-the-hood-universal-commerce-protocol-ucp/
  8. Google, "Universal Commerce Protocol features and AI tools" (Universal Cart / Google Pay) — https://blog.google/products-and-platforms/products/shopping/shopping-updates-google-marketing-live/
  9. Campaign, "How much should publishers fret about Google AI Overviews?" (DMG Media / Daily Mail: CTR 25.2% → 2.8%) — https://www.campaignlive.com/article/the-weather-changing-publishers-fret-google-ai-overviews/1933229
  10. Search Engine Journal, "Google AI Overviews Impact On Publishers" (Pew / Chegg / zero-click / citation-source analysis) — https://www.searchenginejournal.com/impact-of-ai-overviews-how-publishers-need-to-adapt/556843/
  11. Japan Newspaper Publishers & Editors Association, statement on AI search services (2026/4/20) — https://www.pressnet.or.jp/statement/copyright/260420_16248.html
  12. Impress Watch, "Asahi and Nikkei sue Perplexity over AI search" (Tokyo District Court) — https://www.watch.impress.co.jp/docs/news/2041960.html

Originally published in Japanese on note. I'm Satoshi Yamauchi — AI Strategist at Sun* / Founder & CEO of Leading.AI. I publish an open-source library on AI industry structure. Find me on X @s3atoshi.

Top comments (0)