Alphabet’s search business is still growing while artificial intelligence rewrites how people find information. The real investment question is not whether Google survives. It is what survival costs—and whether the new version of Search can remain as profitable as the old one.
For most of my adult life, “Google it” has been less a suggestion than a reflex.
I do not announce that I am about to use a search engine. I simply open a browser, type half a thought into a box, and expect the accumulated knowledge of civilization to arrange itself helpfully before I lose interest. Google became so embedded in daily behavior that its brand stopped functioning like a company name and started behaving like a verb, a utility, and occasionally a substitute for consulting a qualified physician.
Then generative artificial intelligence arrived and asked an impolite question: What if people no longer want ten blue links?
What if they want one direct answer?
What if they ask ChatGPT, Perplexity, Claude, or another AI assistant to research a topic, compare products, plan a trip, explain a diagnosis, summarize a document, or complete a task without ever visiting a traditional search results page?
For Alphabet investors, this is not a side issue. Google Search is one of the most successful businesses ever created. In the second quarter of 2026, Google Search and other advertising revenue reached $63.3 billion, up 17% from the prior year. That is not the financial profile of a product currently being wheeled toward hospice.
But disruption rarely waits for the incumbent’s revenue to decline before becoming real. Sometimes the old business keeps growing while the ground beneath it changes. The numbers look healthy, management sounds confident, and somewhere in the distance a new user habit is quietly learning to walk.
So I want to examine both sides honestly.
The bull case says Google owns the distribution, data, infrastructure, advertiser relationships, and engineering talent needed to transform Search before anyone can replace it. The bear case says Google is being forced to rebuild its most profitable product into something more expensive, less predictable, and easier for competitors to imitate—all while regulators pry open the advantages that protected it.
Both arguments are credible.
That is what makes the stock interesting.
The Question Is Not Whether Search Changes
Search has already changed.
Google has woven AI Overviews and AI Mode into a more conversational experience. Users can ask longer questions, request comparisons, upload images and files, continue with follow-ups, and increasingly ask the system to perform tasks rather than merely locate pages. Google says AI Overviews serves more than 2.5 billion monthly users, while AI Mode has surpassed 1 billion monthly active users.
Those figures matter because they challenge the simplistic bear argument that Google somehow slept through the AI revolution and woke up to discover a chatbot parked on its lawn.
Google did stumble publicly. Its early chatbot rollout produced the kind of avoidable errors that transform a product launch into a global comedy workshop. The company appeared defensive and culturally slow compared with younger AI labs that had no enormous advertising machine to protect.
But Google has since moved with considerable force. Gemini now touches Search, advertising, Workspace, Cloud, YouTube, Android, and the company’s developer ecosystem. The Gemini app reportedly reached 950 million monthly active users by the second quarter of 2026. Google’s model APIs were processing about 22 billion tokens per minute. This is not a laboratory experiment taped to the side of the company. It is a full-scale reconstruction project.
The investment question is whether Google can change the user experience without damaging the economic engine underneath it.
That is much harder than adding an answer box.
The Bull Case Begins With the Revenue Nobody Has Managed to Kill
The strongest argument for Alphabet is painfully simple: Search is still growing.
Despite several years of predictions that generative AI would devour Google’s core business, Search and other advertising revenue grew 17% year over year in the second quarter of 2026. Google Services revenue rose 15% to $94.5 billion, and the segment produced $39.5 billion in operating income at a 41.8% margin.
I can construct impressive theories about disruption, but eventually the theory has to visit the income statement.
So far, users have not abandoned Google at scale. According to management, AI features are increasing overall query volume. Search usage reached an all-time high during the 2026 World Cup, and AI Mode queries have more than doubled every quarter since launch. People appear to be using AI search for longer, more complicated requests they might not have attempted with keywords.
That expansion is the heart of the bull thesis.
Traditional search handled a limited set of intentions well. Find a website. Check a score. Locate a restaurant. Compare a product. AI can broaden the range of searchable activity into planning, reasoning, brainstorming, analysis, and multimodal questions. If Google turns previously unsearchable thoughts into commercial queries, AI may enlarge the market rather than simply cannibalize it.
The company does not need every AI query to replace an old search. It wants AI to create more total opportunities to connect intent with an advertiser.
That is a familiar Google talent: take a vague human desire, convert it into structured commercial intent, and place a paid result nearby with such impeccable timing that the advertisement feels less like advertising and more like a helpful coincidence.
Google Owns the Most Valuable Starting Point: Habit
Investors often talk about distribution as though it were a row in a spreadsheet. In consumer technology, distribution is frequently psychology.
Google is the default behavior for billions of people. It is integrated into Chrome, Android, browsers, phones, address bars, operating systems, and routines built over decades. Users do not evaluate every search provider before asking who played a supporting role in a film from 1997. They use whatever is already there.
That habit gives Google time.
A competitor may have a better model for a particular task, but “better” must overcome convenience, familiarity, trust, speed, and default placement. The history of technology is filled with superior products that discovered users are remarkably reluctant to relocate a reflex.
Google can also introduce AI to people who never deliberately choose an AI product. A user who would not subscribe to a standalone assistant can encounter an AI Overview inside the search box already used daily. That lowers adoption friction to almost nothing.
The bull case does not require Gemini to win every benchmark. It requires Google’s integrated experience to remain good enough that most users see no compelling reason to leave.
“Good enough” sounds insulting until one notices how many trillion-dollar platforms were built on it.
The Advertiser Machine Is an Unfairly Large Advantage
Google’s real moat is not merely that people search. It is that millions of advertisers know how to pay for those searches.
The company has spent more than two decades building auctions, measurement systems, campaign tools, conversion tracking, merchant feeds, local listings, and relationships with businesses ranging from global retailers to the dentist who appears whenever I search “tooth pain should I panic.”
An AI competitor can build a compelling answer interface. Building an advertising marketplace with scale, demand, trust, measurement, and enough commercial inventory is another undertaking entirely.
Google is already adapting its monetization systems to longer AI queries. Management says Gemini improves query understanding and helps surface relevant ads for searches that were previously difficult to monetize. In one example, the company reported a 20% improvement in the relevance of Shopping ads. Advertisers using AI-powered campaign products such as AI Max or Performance Max reportedly see an average of 15% more conversions or value on Search at similar returns on ad spending.
The company is also testing sponsored links inside AI responses, conversational ad experiences, direct offers, brand agents, and forms of commerce that can move a user from research to purchase without leaving Google’s ecosystem.
This matters because the bear case often assumes AI answers eliminate the advertising opportunity. They may instead relocate it.
A query such as “best running shoe” produces obvious ads. A conversational request—“I am training for my first half marathon, have flat feet, run mostly on pavement, and need something under $150”—contains richer intent. If Google can answer helpfully while presenting a relevant commercial option, that query may be more valuable than its keyword-based ancestor.
The uncomfortable truth for anyone hoping AI creates a pristine advertising-free internet is that conversational systems may understand our purchasing intentions better than search engines ever did.
Capitalism has noticed.
Google Has the Full Stack—and the Bill to Prove It
Google’s technical position strengthens the bull case. The company designs custom tensor processing units, operates global data centers, develops frontier models through DeepMind, owns massive consumer products, runs one of the largest cloud platforms, and controls software distribution across Android and Chrome.
This full-stack structure can reduce costs and optimize AI from the chip through the user interface. Alphabet says it lowered the cost of AI Mode responses to the lowest level since launch even while adding more advanced capabilities.
That sentence deserves attention.
The central economic challenge of AI search is inference cost. A traditional query can be served cheaply through an enormously optimized system. A generative response requires more computation. If every search becomes a miniature reasoning session, the cost of serving the world’s curiosity rises dramatically.
Google has more tools than most competitors to attack that problem. It can improve model efficiency, build specialized hardware, route simple queries to cheaper systems, reserve expensive models for difficult tasks, and spread infrastructure investment across Search, Cloud, YouTube, Workspace, Gemini, and external customers.
The infrastructure serves both defense and offense. The same capital expenditure that protects Search can fuel Google Cloud, which grew revenue 82% to $24.8 billion in the second quarter of 2026 and ended the quarter with a $514 billion backlog. Cloud operating income reached $8.8 billion.
In the bull version of events, Alphabet is not spending desperately to preserve an aging search box. It is building the computing layer for an AI economy and using the cash generated by Search to finance the transition.
That is a far more attractive story.
It is also an extraordinarily expensive one.
The Bear Case Starts With $200 Billion of Nervous Energy
Alphabet spent $44.9 billion on capital expenditures in the second quarter alone. Management raised its full-year 2026 capital spending guidance to between $195 billion and $205 billion and expects spending to increase significantly again in 2027.
For perspective, this is the kind of number that makes “we are investing for the future” sound less like a strategy and more like the construction of a small artificial moon.
The company produced negative free cash flow of $5.9 billion in the quarter, although trailing-12-month free cash flow remained positive at $53.3 billion. Alphabet has the balance sheet to spend aggressively. It ended the quarter with $242.5 billion in cash and marketable securities, according to management.
But affordability does not guarantee attractive returns.
The bear worries that AI changes Search from an extraordinarily capital-efficient toll road into a computationally intensive service that must spend more merely to preserve existing behavior. Revenue may grow while returns on incremental investment decline. Depreciation, energy, data center operations, and AI talent costs can pressure margins long after the initial construction headlines fade.
This is the difference between having money and using money well.
If Alphabet spends $200 billion to create entirely new profit pools, investors may celebrate. If it spends $200 billion to keep users from defecting while competitors force prices down, the same expenditure looks like a defensive tax.
The numbers can be identical while the economics are opposites.
The Innovator’s Dilemma Has an Advertising Auction Inside It
Google must improve Search in ways that could weaken the business model that made Search valuable.
Traditional search sends users through a page of links and advertisements. Generative search aims to resolve the question directly. The more complete the answer, the less reason the user may have to click anything—including an ad.
Google says monetization on queries that display AI Overviews remains encouraging. That is good evidence for the bulls, but it is not the end of the argument. AI Mode is more conversational and agentic than an overview placed above traditional links. A system that plans, compares, negotiates, and eventually transacts on behalf of the user may concentrate commercial decisions into fewer interactions.
That could make each interaction more valuable. It could also reduce the number of opportunities available for advertisers to bid.
Imagine an AI agent planning an entire vacation. Traditional search might generate dozens of queries for flights, hotels, restaurants, attractions, transportation, insurance, and reviews. An agent may collapse the process into one conversation and a handful of recommendations.
Google must capture enough value from the compressed journey to replace all the economic activity it eliminated.
This is not impossible. It is simply unproven at the scale and profitability investors have come to expect.
The company is effectively renovating a casino while every machine remains in use. It must change the games, install new wiring, keep customers entertained, preserve revenue, satisfy regulators, and make sure the new machines do not cost more to operate than they collect.
No pressure.
AI Competitors Do Not Need to Replace Google Entirely
The bear case is often framed too dramatically. A competitor does not need to destroy Google Search to damage Alphabet’s economics.
It only needs to capture valuable categories of queries.
Research-heavy questions can move to answer engines. Coding questions can move to specialized assistants. Product discovery can move to Amazon, TikTok, Reddit, or AI shopping agents. Travel planning can begin inside a chatbot. Local discovery can happen through maps, social video, or vertical applications. Browser-level agents may eventually perform tasks without presenting a conventional results page at all.
Search fragmentation matters because not all queries are economically equal. Losing a low-value factual query may be irrelevant. Losing a commercial query moments before a purchase is more painful.
Younger users may also develop habits that differ from those of older generations. If they begin complex tasks inside AI assistants rather than Google, the behavioral moat weakens gradually. The decline might not appear as a dramatic collapse in total search volume. It may appear as slower growth in the most profitable query categories, higher traffic-acquisition spending, greater promotional costs, or pressure on pricing.
Investors should watch the mix, not merely the headline usage number.
A platform can retain billions of users while losing the part of their attention that advertisers value most.
Regulation Is Now Part of the Product Roadmap
Google’s defense of Search is not taking place on an empty field. Courts and regulators have already concluded that the company maintained monopoly power through unlawful practices.
The U.S. remedy imposed in 2025 prohibited exclusive distribution contracts involving Google Search, Chrome, Google Assistant, and Gemini. It also required Google to make certain search-index and user-interaction data available to rivals and offer search and search-ad syndication services under specified conditions. Implementation and appeals remained active into 2026.
For bears, these measures attack the mechanisms that helped Google preserve scale. Rivals may gain data, distribution opportunities, or infrastructure they previously lacked. Device makers and browser companies may have more freedom to promote alternatives. Google may need to compete more openly for defaults without using exclusivity as a wall.
For bulls, the outcome could have been much worse. Google was not ordered to divest Chrome in the final judgment, and the company retains ownership of its core products, technology, brand, and advertiser relationships. The remedies may weaken the moat without draining it.
Still, regulation creates a persistent uncertainty premium. The search case is not Alphabet’s only antitrust exposure. Advertising technology has faced separate litigation, and governments around the world continue examining platform power, privacy, AI training data, publisher economics, and digital competition.
The company now has to innovate while lawyers stand beside the engineers asking whether the new feature will become an exhibit.
That is rarely a productivity enhancement.
The Web Ecosystem Cannot Be Treated Like Free Raw Material Forever
Google’s AI depends on an open web filled with publishers, creators, merchants, forums, experts, and ordinary people contributing information. Generative answers can summarize that information so effectively that users have less reason to visit the original sources.
This creates a structural problem.
If AI search reduces publisher traffic, publishers may produce less, block crawlers, demand licensing payments, erect paywalls, or prioritize content designed for other platforms. The search engine becomes more useful in the short term while weakening the ecosystem that supplies its knowledge over the long term.
Google says its AI features send billions of clicks to websites every week and has introduced more visible links and publisher controls. But publishers care about whether their own traffic and revenue remain healthy, not whether an enormous aggregate number sounds impressive in a presentation.
The bear sees a dangerous loop: AI answers reduce clicks, weaker economics reduce original content, declining content quality makes AI answers less trustworthy, and users migrate toward closed databases, social communities, or specialized sources.
The bull sees Google as the company best positioned to solve the problem because it already understands crawling, ranking, attribution, spam, freshness, and publisher relationships better than any AI-native rival.
Both views can be true for a long time, which is deeply inconvenient for anyone demanding a clean forecast.
What I Would Watch as an Investor
I would not base the thesis on whether Gemini wins one benchmark or whether a competing chatbot briefly tops an application-store chart. Those events attract attention but reveal little about durable economics.
I would watch five things.
First, Search revenue growth. If AI usage rises while Search advertising continues growing at a healthy double-digit rate, the bull thesis remains intact. If query volume climbs but revenue growth slows sharply, Google may be generating engagement it cannot monetize efficiently.
Second, Google Services margins. Rising revenue is less impressive if inference, depreciation, energy, and traffic-acquisition costs consume the gains. The 41.8% operating margin reported in the second quarter of 2026 provides a strong starting point. The direction from here matters.
Third, capital intensity and free cash flow. Investors need evidence that the extraordinary infrastructure build produces returns across Search and Cloud. “Demand remains strong” cannot serve as a permanent substitute for cash generation.
Fourth, commercial behavior inside AI Mode. I want to know whether users click ads, accept offers, complete transactions, and return for purchase-related tasks. General monthly-active-user figures demonstrate distribution; commercial engagement demonstrates the business.
Fifth, competitive habit formation. Are users beginning their high-value research elsewhere? Google can defend its overall share while losing the starting point for complex decisions. Survey data, referral patterns, browser behavior, advertiser returns, and management commentary may reveal that shift before the income statement does.
The market will obsess over AI model rankings. I will obsess over whether advertisers continue receiving measurable returns.
One of those things pays Alphabet’s bills.
My Verdict: The Bull Case Leads, but It Is Paying for the Privilege
Today, I lean bullish on Google’s ability to defend Search.
The evidence is stronger than the obituary. Search revenue is growing. AI features are reaching enormous audiences. Query volume is expanding. Google has successfully placed Gemini inside products people already use, and its advertising infrastructure is beginning to monetize more detailed conversational intent. The company possesses distribution, models, chips, data centers, cash, engineers, and advertiser demand at a scale few competitors can match.
Most importantly, Google understands that the threat is existential. Complacent incumbents protect the old interface. Google is actively willing to replace it.
But my confidence is not unconditional.
The company is spending at a staggering rate. The free-cash-flow pressure is real. AI queries are more expensive, the long-term ad format remains unsettled, regulators have weakened distribution advantages, and competitors can erode the business without replacing it outright.
The bull case wins if AI expands Search faster than it compresses advertising opportunities, while infrastructure efficiencies preserve attractive margins. The bear case wins if Google must spend ever-larger amounts to serve queries that generate less incremental profit and face more competition.
That is the contest.
Google does not need to preserve the old search engine. In fact, trying to preserve it would be the surest way to lose. It needs to preserve the economic relationship underneath Search: users reveal intent, Google organizes information, advertisers pay for relevant access, and everyone receives enough value to return.
The interface can change from links to answers, from answers to conversations, and from conversations to agents. The investment survives if that relationship survives.
For now, the numbers suggest it is surviving remarkably well.
Yet Alphabet’s greatest advantage may also be its greatest burden. It has the cash to fund the AI transition because Search is so profitable. That same profitability means every new product is judged against one of the finest business models ever built.
Google is not merely racing its competitors.
It is racing its own margins.
And that may be the only opponent large enough to make this fight genuinely close.
Based on information available through August 14, 2026. Sources: Alphabet’s second-quarter 2026 earnings call, Google’s 2026 AI Search announcements, Google’s advertising updates for AI Search, and the U.S. Department of Justice’s summary of the Google Search remedies.
This article is commentary and analysis for informational purposes and is not personalized investment advice.
Comments
Post a Comment