Google Business Model: How Search Built a $400B Advertising and AI Machine

The Origin

In 1996, two Stanford PhD students — Larry Page and Sergey Brin — started "BackRub," analyzing the web's link structure. Their insight: the number and quality of links pointing to a page was the best signal of its importance. That became PageRank, and the Google business model still rests on the data flywheel it kicked off.

Google launched in September 1998 from a friend's garage in Menlo Park. While AltaVista and Yahoo cluttered their pages with portals, news and stock tickers, Google's homepage was famously empty — just a logo and a search box. That restraint became the brand.

The AdWords Revolution

In 2000, AdWords let any business buy text ads next to search results. The auction model aligned Google's revenue with user value: better, more relevant ads earned more clicks and more money. This is the core of how Google makes money — it monetizes intent at the exact moment someone is looking to buy. By 2004, Google generated $3.2B and went public at $85 a share.

Building the Ecosystem

Gmail (2004), Maps (2005), Android (2008) and Chrome (2008) each made the internet more useful while quietly creating new ad surfaces and richer data signals. The boldest bet was buying YouTube for $1.65B in 2006 — derided at the time as overpaying for a money-losing video site. Two decades on, YouTube alone pulls in $40B+ of advertising a year and is the most-watched streaming service on American TVs, a return that dwarfs the purchase price.

Twenty years later, that same playbook — useful product, attached to the ad engine — is how Gemini is being pushed into Search, Android and Workspace. The difference now is cost: AI answers are far more expensive to serve than a list of blue links, which is exactly why the next decade of the Google business model hinges on monetising AI without eroding the ad economics that pay for it.

The Search Problem

By the late 1990s the web was doubling in size every few months, but search engines were terrible at making sense of it. AltaVista, Lycos and Yahoo ranked pages mostly by how often a keyword appeared, which spammers gamed by stuffing invisible text. Type a query and you got a wall of junk. The web was becoming the world's library with no card catalogue.

The Advertising Problem

Online advertising was just as broken. Banner ads converted at under 1% because they ignored intent — they showed the same pitch to everyone. Small businesses couldn't afford the big media buys, there was no real way to measure return, and traditional TV and print wasted more than half of every dollar with no tracking at all. Advertisers knew half their spend was wasted; they just couldn't tell which half.

Google's opportunity was to fix both at once: organise the web well enough that people trusted it, then sell advertisers the one thing they had never been able to buy — a customer at the exact moment of intent.

This is how Google makes money: it gives ~4 billion people free tools, harvests the intent signal those tools produce, and auctions that intent to advertisers. Everything else is plumbing for that loop.

PageRank

Page and Brin treated the web like a democracy. Each inbound link was a "vote," and votes from important pages counted more. That recursive ranking buried the keyword-stuffers and gave users results they could trust — and trust is what made Google the default front door to the internet.

The Ad Auction

The genius wasn't the ad; it was the auction. Google runs a real-time second-price-style auction (GSP) on every query, ranking ads by bid times Quality Score so that a more relevant ad can win while paying less. This fixed all four of online advertising's old problems at once: ads appear only on matched intent, any SMB can self-serve in minutes, advertisers pay per click rather than per impression, and every click is measurable down to the conversion. Because relevance is rewarded, Google's incentives align with both the user and the advertiser — a rare three-way fit that still generates ~$340B a year.

The Ecosystem as Moat

Every free product feeds the auction. Chrome (~65% browser share) and Android (~70% of smartphones) guarantee Google is the pre-set default, so the queries flow to it by default rather than by choice. Gmail, Maps and YouTube each create new ad surfaces and richer intent signals — a Maps search for "dentist" is worth far more than a banner impression. The result is a flywheel: more users produce more data, better results and more advertisers, which funds the free products that attract the next billion users.

Layering AI On Top

The 2023–2026 move is to wrap this loop in Gemini — AI Overviews and AI Mode in Search, Gemini in Workspace and Android — while renting the same TPU infrastructure to enterprises through Cloud, now a $70B+ run-rate business. The unresolved tension: AI answers that satisfy a query without a click can cannibalise the very ad slots that pay for everything.

Google gives ~4 billion people free tools — Search, Gmail, Maps, Android, YouTube — and sells the resulting intent to advertisers. Advertising is roughly 75% of Alphabet's $402.83B FY2025 revenue.

The big cost is Traffic Acquisition — roughly $60B paid to partners like Apple (~$20B/yr) for default placement. Even after ~$52B of R&D and heavy AI capex, Alphabet clears a ~33% net margin ($132.2B net income).