Google earnings raise AI cannibalization worries for search profits

Could Google's own AI be eating into its search business? And can Gemini generate enough revenue to make up the difference?

Growing AI usage does not automatically mean higher profits. Alphabet's most recent quarterly results sharpen a dilemma for shareholders: can ad revenue and subscription income replace the value of the searches that AI is absorbing?

Analysis | October 29, 2026 | Alphabet: GOOGL and GOOG

Google shares are down almost 11% in pre-market trading. For today, I am watching the $280-$284 range as a potential dip-buying zone. Whether this is temporary remains unclear, but my working thesis is that Google's AI tools may be displacing some search advertising opportunities faster than new revenue streams can offset them. Alphabet's earnings give that concern tangible evidence, though they do not prove an overall loss: search revenue continues to grow, and the company does not disclose enough detail about Gemini's economics to settle the debate.

Like many users, I have noticed myself clicking fewer links in Google search results over the past year. That personal shift makes me wonder: if more people get answers directly from AI, what happens to the business model that relies on searches and subsequent clicks?

The danger I see is that Google successfully changes how people find information while undermining the economics that made its search franchise so profitable.

Recent earnings provide concrete support for my hypothesis.

Alphabet's results for the quarter ending September 30, released October 28, showed Google Search and other revenue up 12% to $63.35 billion. Total company revenue rose 25% to $127.6 billion. Those are growth numbers, not signs of an imminent collapse.

Markets were less enthusiastic than the top-line growth might suggest. According to investingLive, Alphabet shares dropped more than 9% after hours as a Search miss and $9 billion cash burn outweighed Cloud growth. That market reaction reflects multiple concerns and should not be taken as definitive proof of cannibalization.

What draws my attention, however, is management's own explanation for the slowdown.

On the earnings call, chief business officer Philipp Schindler noted that rapid growth in AI Mode queries had impacted performance because AI Mode monetizes at a lower rate. He also pointed to foreign exchange as the biggest factor in Search's deceleration from the prior quarter, along with changes in retail advertising and World Cup-related spending.

That distinction is important. AI was one factor, not the only explanation.

Meanwhile, CEO Sundar Pichai said search query growth was its strongest in five years and the Gemini app had reached one billion users.

To me, those disclosures highlight a key possibility: Google can drive more activity without generating the same revenue from each interaction.

They do not clarify how much revenue would have existed without AI. Some AI queries may represent entirely new demand. Others may replace traditional searches. Public disclosures do not cleanly differentiate between the two.

Zero clicks for publishers does not equal zero revenue for Google

My personal browsing habits are just one data point. Broader evidence is worth examining.

Pew Research Center studied browsing behavior from 900 U.S. adults in March 2025. Users clicked a traditional result on 8% of visits to Google pages with an AI summary, compared with 15% without one. Links within the summaries attracted clicks on just 1% of visits.

This was an observational study, not a randomized experiment or a direct measure of Google's ad revenue. Search results were reconstructed in April, and queries that generate summaries differ from those that do not. Still, it provides evidence that reading an answer can replace visiting a website.

Three products also need to be distinguished: AI Overviews summarize information within search results; AI Mode offers conversational search; Gemini is a separate assistant. Their revenue models may not be identical.

Google does not make money every time someone clicks an ordinary, unpaid search result. Losing a publisher referral therefore does not automatically cost Google a paid click. Ads can appear within or alongside AI Overviews and AI Mode.

The risk I find more concerning comes when a single AI conversation replaces a series of commercially valuable searches.

A person researching a vacation might previously have searched multiple times for destinations, hotels and comparisons. An assistant could compress that process into a shortlist. If fewer valuable ad opportunities remain, and new placements do not compensate, Google loses potential revenue even while offering a better user experience.

Conversely, an assistant that helps someone reach a purchase faster could create a more valuable ad opportunity. The outcome depends on what users do next, not simply whether they leave Google.

Gemini subscriptions are not an automatic substitute

I would not assume that paid AI plans can simply replace whatever advertising revenue disappears.

The math is more complicated.

Google lists AI Pro at $19.99 per month in the U.S., bundled with benefits beyond the Gemini assistant. That price is not equivalent to incremental Gemini profit. Existing subscribers may upgrade from another paid plan, promotions affect receipts, and serving their requests costs money.

Alphabet's financial reporting does not provide a standalone Gemini consumer subscription revenue and profit figure. Its broader subscriptions, platforms and devices category mixes several businesses. Nor does one billion Gemini users mean one billion paying customers.

Let's put some numbers around that question.

Suppose the annualized value of Search revenue lost relative to an otherwise achievable outcome equaled just 1% of its latest quarterly revenue run rate. Using $63.35 billion per quarter, that would be roughly $2.53 billion a year.

Replacing that gross revenue at $19.99 a month would require about 10.6 million additional subscribers paying for a full year.

This is an illustration, not an estimate of actual cannibalization. It ignores discounts, churn, existing subscription spending and the cost of delivering AI. Replacing lost profit would require comparing the contribution margins of both businesses.

It also shows why subscriptions should not be dismissed. A sufficiently large, durable paying audience could make a meaningful difference.

The harder problem is converting enough users while keeping the free service attractive. A fixed monthly fee brings predictable revenue, but intensive usage can add costs without adding another subscription payment. Usage limits and higher-priced tiers can help, though they may also drive price-sensitive users elsewhere.

Switching AI assistants may be easy. Rebuilding a working relationship may not be.

For someone asking occasional questions, moving to another AI platform can be as simple as opening a different app.

That creates a challenge for premium subscriptions. If several assistants answer everyday questions well enough, users can cancel, rotate subscriptions or use multiple free products.

Switching becomes harder when an assistant holds project history, understands preferences and connects to everyday work. An enterprise may also need to rebuild integrations, retest reliability and repeat security reviews.

My view is that the more durable advantage could come from how deeply an assistant fits into someone's daily work. A temporary lead on a model benchmark may be easier for competitors to challenge.

But even that advantage is contestable.

Google already offers tools to import memories and chat histories from other AI services. Its documentation says project files, attachments and generated images do not transfer through the chat-history import.

This shows that switching friction can be reduced, not eliminated. A transferred conversation is not necessarily a functioning replacement for an established workflow.

There is an irony here: Google wants portability when it helps Gemini win customers. If comparable portability becomes widespread, it could also make Gemini customers easier to lose.

Cheaper, open AI could squeeze the subscription cushion

Open-weight models, whose trained parameters developers can obtain and deploy subject to their licenses, create another source of competitive pressure.

They can support lower-priced services, specialized assistants and some applications running on users' own devices. Google participates in this market through Gemma, which it makes available for applications, hardware and hosted services.

The competitive threat does not require every open model to beat every premium model. For many routine tasks, being sufficiently useful at a lower price may be enough.

Stanford's 2026 AI Index described convergence among leading models on one widely used evaluation, while also finding that the gap between the best closed and open models had widened again after previously narrowing. That is evidence of active competition, not proof that all models are interchangeable.

Nor does an open model guarantee a cheaper finished service. Hosting, security, maintenance, reliable information retrieval and customer support still cost money.

This is the potential squeeze that concerns me: AI could reduce some advertising opportunities while competition limits the subscription prices available to replace them.

Lower computing costs also help Google. They could make AI answers cheaper to deliver and improve margins even if customer prices fall.

Google could turn the threat into a stronger business

I also see a credible case for Google emerging from this stronger.

Google is developing commercial formats for AI Search, including conversational ads and offers linked to checkout. These are real product initiatives, though their long-term profitability remains unproven.

If an assistant understands what someone wants and helps complete the transaction, Google could capture value closer to the purchase. Fewer browsing steps would not necessarily mean less revenue.

Alphabet also has revenue sources beyond consumer subscriptions. Google Cloud revenue rose 103% to $30.7 billion in the latest quarter. Google Services operating income increased 17%, and its operating margin reached 40.7%.

Those figures challenge any claim that AI has already broken Alphabet's operating model. Services includes more than Search, however, and shared AI research expenses also sit at the Alphabet level. The segment margin does not reveal Gemini's standalone profitability.

Cash flow adds another complication. Alphabet reported negative quarterly free cash flow of $9.1 billion. Management attributed the pressure to capital spending, inventory for its chip systems business and a $5.2 billion European Commission fine payment. That cash outflow cannot simply be labeled a loss from AI search.

A related investor concern appears in investingLive's report on Samsung's share-price slide amid disappointment over the timing of shareholder returns. Samsung shares fell as much as 6.6% as investors awaited details of its broader return plan, due in late January, despite strong AI-related earnings.

Samsung's situation does not establish a problem with Google's search model. It illustrates a broader test for AI investments: how much cash reaches shareholders, and when? Strong demand and rising revenue do not settle that question on their own.

I also have to consider what happens if Google does not change. Preserving familiar Search economics temporarily could come at the cost of driving users toward competing assistants.

Internal cannibalization could be the cost of defending a business that would otherwise face external disruption.

Three paths I would consider as an investor

A profitable transition. AI generates new demand, advertising catches up, subscriptions retain customers and serving costs fall. Google protects its customer relationship and expands the ways it earns money from it.

A larger business with weaker economics. Usage keeps rising, but commercially valuable searches are partly replaced by less profitable interactions. Subscriptions and Cloud help, yet more investment is needed to generate each additional dollar of cash flow.

A shift in who controls the customer. Rival assistants become the starting point for research and purchases. Google may still supply models, computing or information behind the scenes, but another platform controls the recommendation and commercial relationship.

These are scenarios, not forecasts. Different parts of Alphabet could experience them simultaneously.

When assessing the next results, these are the questions I would focus on:

• Does Search advertising growth strengthen after accounting for currency and comparison effects?
• Does management provide evidence that AI advertising earns more per commercial task, alongside useful advertiser conversion data?
• Do paid AI plans produce durable incremental revenue, rather than mainly trials, discounts or upgrades from existing subscriptions?
• Do operating cash generation and investment spending move toward a more sustainable balance?
• Are users becoming more committed to one assistant, or spreading their activity across interchangeable services?

Query counts alone cannot settle those questions. A long conversation may contain many prompts without representing many separate purchase opportunities.

My hypothesis remains: Google could remain a leader in AI while surrendering some of the exceptional economics of traditional search.

Gemini subscriptions may help close that gap. What I want to understand is whether they can do enough, alongside AI advertising and other businesses, to protect Google's profitability. User growth alone does not answer that question. In terms of simplicity, IMHO, stock must, in the near future (till the end of Q1 2027) should reclaim back $310, in order to reignite the bullish case. That would be my line in the sand for Alphabet stock (ticker GOOG).