For the first time since 2015, the company that answers 14 billion questions daily dropped below 90% market share. That might sound like a small dip, but it's actually the beginning of the most expensive business suicide in history.
Here's what makes this wild: Google has the world's best AI engineers and unlimited resources. But they can't use any of it. Because every time they make search better, they lose billions of dollars.
For the first time in a decade Google’s global share has slipped under 90 %—to 89.65 % in April 2025. On desktop it’s already 79.1 %, and in Europe just 77.78 %.
Translate that one-point drop: ±50 million users a month just walked out the door. They didn’t bounce to Yahoo; they bounced to AI tools that answer instead of link.

The Innovator’s Dilemma Hits Google in the Face
Google optimized for ad revenue.
Users optimized for answers.
Now they’re in conflict.
Clayton Christensen warned us about this. It’s the Innovator’s Dilemma playing out in public. Google has the AI talent. They’ve got the compute. But they’re locked into a $300B business model that depends on blue links and click-throughs.
Every AI summary pushed to the top kills a little more of their own ad real estate. And yet… they can’t not do it.
The Numbers That Should Terrify Everyone
March 2025: 89.71% market share. The first crack in Google's empire since Obama was president.
But here's the real story: Google processes 14 billion searches daily. ChatGPT handles 37.5 million search-like prompts. That's a 373-to-1 advantage for Google.
So why is Google panicking?
Because those numbers are like comparing Netflix to TikTok in 2018. Scale doesn't matter when user behavior fundamentally shifts. Perplexity AI (a company most people have never heard of) grew 204% last year. That's exactly how disruption begins: small, fast, and in places the giant can't afford to compete.
The scariest part? Google can't fight back without destroying themselves.
Where Did The Search Traffic Go?
Product | Monthly visits | YoY growth |
|---|---|---|
ChatGPT | 5.2 B | + — |
Perplexity | 159.7 M | +204 % |
Claude | 76.8 M | — |
Even Bing (yes, that Bing) has inched up to ~4 % worldwide, powered by OpenAI integrations.

Learn AI in 5 minutes a day
This is the easiest way for a busy person wanting to learn AI in as little time as possible:
Sign up for The Rundown AI newsletter
They send you 5-minute email updates on the latest AI news and how to use it
You learn how to become 2x more productive by leveraging AI
The $300 Billion Handcuffs
Here's the trap that's about to destroy the internet's most powerful company:
Google made $198 billion from search advertising in 2024 (57% of their total revenue). They can’t just shut this off.
When you search "best restaurants near me," Google makes money when you click links. The more you click, the more ads they show.
But AI search gives you the answer directly: "Based on reviews, here are the top 3 restaurants within 2 miles." No clicking. No ads. No revenue.
Every AI answer Google provides is literally a bullet through their own business model.
This is the classic "innovator's dilemma, when your current success prevents you from adapting. Google is living through the most expensive version of this trap in corporate history.
So Google tries to “thread the needle”, rolling out AI Overviews to 1.5 billion users while still serving ads. And when an AI overview appears, paid-ad CTR collapses to 9.8% (vs 21.3 % without it).
But every smart summary pushes paid links lower. Every direct answer reduces inventory. It’s like renovating your house while you’re living in it and still paying rent.

AI Search Is The New Front Line
The raw volume gap is still massive. Google fields 14 billion searches a day. ChatGPT? Around 37.5 million prompts that look like search. But that stat hides what actually matters:
User expectations have changed.
People don’t want ten blue links anymore. They want one smart answer, with context, citations, and the ability to ask follow-ups.
Gen Z doesn't search like everyone else. While older generations think in terms of "search engines," younger users default to conversational AI. They ask questions naturally and expect direct answers, not homework assignments disguised as research.

That’s what AI-native tools like Perplexity and ChatGPT are delivering.

Perplexity CEO Aravind Srinivas dropping a reply on my post about Perplexity’s new agentic browser, Comet
Perplexity traffic is up 200%+ year over year. ChatGPT handles over a billion messages a day. And a third of those are functionally search.
The Zero-Click Apocalypse
Here's the statistic that could kill the internet economy: 70.5% of Google searches now end without a single click.
Seven out of ten times you search, you get your answer directly from Google and never visit another website. When Google's AI overviews appear, that jumps to 74.8%.
That's billions of lost clicks for publishers, content creators, and anyone who depends on search traffic. The entire ecosystem built around Google search is being strangled by Google's own AI improvements.

The Google Patent That Reveals Everything
Mike King unearthed patent US 20240289407 A1, the brains behind Google’s “AI Mode.” Google’s “Search with Stateful Chat” patent is their AI search playbook.
The flow matters more than the jargon:
Intent normalisation
Query fan-out – dozens of synthetic queries proliferate.
Custom corpus build – Google assembles a private mini-index for your question.
Dense retrieval + LLM orchestration – passages, not pages, are selected.
Answer composition – the model stitches a single narrative; ranking is gone.
If your page satisfies the head term but misses the synthetic swarm (“best foods for LDL”, “exercise to drop cholesterol”, “risks of natural lowering”), you’re invisible.
The patent shows Google's plan to transform search using sophisticated AI reasoning. When you ask a question, their system generates dozens of related queries and provides personalized answers based on your entire Google history.
It's incredibly advanced technology. It's also incredibly dangerous to their business.
The system maintains "stateful memory" of your conversations, personalizing responses based on your Gmail, location, and search history. No two users get the same results, making traditional SEO completely changed.
But here's what the patent doesn't explain: how to make money from AI search.

What This Means for You (And How to Adapt)
If you're building anything on the internet (content, products, a brand, a business) here’s the deal:
Google isn’t dead, but it’s bleeding. AI is now the front door to the web. That changes the game for how people find you, trust you, and convert.
Success is no longer about getting people to visit your website, it's about getting AI systems to reference your content when answering questions.
Instead of optimizing for keywords and rankings, you now need to engineer for "relevance value". Basically, how likely your information is to be synthesized into AI answers.
Here’s how to adapt before it’s too late:
1. Stop chasing keywords. Start chasing quotability.
AI doesn’t care if you rank. It cares if you’re citable. Use bullet points, bold claims, and clear takeaways. Become the source AI wants to reference.
2. Design for extraction.
Structure your content so AI can pull from it easily. Think paragraph-as-API. If it reads like a rambling blog post from 2012, it’s invisible in 2025.
3. Be everywhere AI looks.
Your brand, product specs, and answers need to be structured, accurate, and consistent across the web. AI pulls from the commons. Own your data footprint.
4. Build your own pipeline.
Google is no longer your growth engine. Double down on direct channels like email, SMS, apps, and community. Push value directly. Own the relationship.
5. Think like your buyer already asked ChatGPT.
Because they did. When they land on your site, your job isn’t to convince, it's to confirm. Skip the fluff. Deliver signal. Anticipate objections.
6. Track different metrics.
Pageviews are vanity. What matters now is AI citations. Are you showing up in ChatGPT answers, Perplexity results, or Google's AI overviews? That’s the new homepage.
Here’s a cheat sheet:
Old SEO Playbook | New AI Playbook |
|---|---|
Rank for keywords | Be cited in LLM answers |
Optimize headlines | Optimize answers inside content |
Track CTR | Track zero-click visibility & AIV |
Blog-first | Structured content, punchy answers |
SERP position matters | AI summarization suitability matters |
The $300 Billion Question
Can a company voluntarily disrupt its own $300 billion revenue stream fast enough to survive the transition to AI search?
History suggests no. Kodak invented digital photography but couldn't abandon film. Blockbuster had streaming technology but couldn't cannibalize retail. Nokia had smartphones before the iPhone but couldn't risk their existing business.

The innovator's dilemma has a nearly perfect track record of destroying dominant companies.
But Google has advantages previous companies didn't: unlimited resources, the world's best AI talent, and data advantages that are nearly impossible to replicate.
The stakes extend far beyond Google. This will determine how billions access information, how the internet economy functions, and whether the advertising model that built the modern web can survive the AI revolution.
Google's 10-year streak is over. The question now is whether their 25-year reign will follow.
While everyone else scrambles to understand what's happening, smart operators are already positioning themselves to win in the post-Google world.
We're helping forward-thinking brands engineer their AI visibility, and to use AI to create scalable content growth.
Cheers,
Matt Berman
If this newsletter lit a fire under you, forward it to one person who would benefit.
🎁 Get rewarded for sharing! My team grew NEW accounts to over 50 million views in just a few months. We made an AI viral hook generator so you can follow the same hook frameworks that we do.
Invite your friends to join, and when they sign up, you’ll unlock our AI Viral Hook Generator—the ultimate tool for creating scroll-stopping content.

{{rp_personalized_text}}
Copy and paste this link: {{rp_refer_url}}



