Every AI search prediction eventually asks the same question: is Google Search dying? Alphabet’s latest earnings report suggests a different reality: search is still growing, but the economics around it are changing.
Search and referral traffic aren’t the same story
Google Search and other revenue increased $9.1 billion in the quarter. The 10-Q attributes this to growth in search queries, driven by increased user adoption and mobile usage. Paid clicks were up 13% year over year; cost-per-click was up only 3%.
Paid clicks are growing at four times the rate of cost-per-click. That means more search activity is still being monetized at roughly the same rate. If AI Overviews were already causing a meaningful collapse in Google’s core search behavior, one place we would expect to see pressure is query growth. So far, Alphabet’s numbers do not show that.
That said, “AI search is killing Google” and “AI search is killing referral traffic to my site” are different claims.
The former isn’t supported by Alphabet’s own numbers so far. The latter appears elsewhere in the business: Google Network revenue, the line that includes AdSense and AdMob, is essentially flat, down $51 million for the quarter, driven by a decline in AdSense partially offset by growth in AdMob.
AdSense is the open-web publisher line; AdMob is in-app advertising, which isn’t the open web in the sense most people mean when they talk about referral traffic.
The open web signal is weaker than the headline suggests. Google Network revenue looks flat, but the composition underneath tells a different story.
The sharper stat: Network impressions were down 12% while cost-per-impression rose 13%. Fewer ad views, monetized much harder per view.
If your business model depends on being a link in someone’s AI Overview, visibility is becoming harder to measure with a single metric. Search can grow while clicks leave the ecosystem.
The other question is whether the infrastructure economics can keep up with the pace of change.
What running AI at scale actually costs
Google Cloud revenue was $24.8 billion, up 82%, and Cloud operating income more than tripled to $8.8 billion. Cloud is no longer just an infrastructure cost center; it is becoming a meaningful profit engine.
But put it next to the other number in this filing: $44.9 billion in capex in the same quarter, company-wide, not Cloud-only. Cloud’s profit is still a fraction of what Alphabet is spending on infrastructure in a single quarter. It is a genuine second growth engine, but it is not yet large enough to fund the AI buildout on its own.
That gap is why the rest of the filing looks the way it does. Alphabet expanded its external financing activity this year across several instruments. External financing activity exceeded $100 billion during the first half while capex reached $80.6 billion, primarily to expand AI infrastructure.
The assumption that AI search can simply be layered on top of existing search infrastructure without changing the economics deserves more scrutiny.
The interesting signal is not that Alphabet cannot fund AI. It clearly can. The signal is that even Alphabet is choosing to finance infrastructure expansion at a scale where traditional operating cash flow is no longer the only variable investors are watching.
One number to ignore, and what to watch instead
Net income for the quarter was $112.2 billion, up 298% year over year, but that does not mean the business is suddenly performing three times better.
The headline number is heavily influenced by unrealized investment gains. Income from operations was $40.8 billion. That is the better measure of how the business actually ran.
Next quarter, look beyond rankings and headline net income.
Watch Network impressions and cost-per-impression for the real referral-traffic signal. Watch Cloud operating income against total capex to see whether Alphabet’s second growth engine is catching up with the cost of its AI buildout. Watch Cloud operating income against total capex to see whether Alphabet’s second growth engine is catching up with the cost of expanding AI infrastructure.
The search business is fine.
What it costs to defend that business against AI is the actual story in this filing.
Till next time 👋
Ilias