Google admits AI Mode cuts search revenue, shares fall 15%

On its earnings call, Alphabet confirmed what analysts have long feared: AI answers in search are taking money away from search itself. Google Chief Business Officer Philipp Schindler said strong growth in AI Mode queries affected results because this mode monetizes less. AI Mode is a search tab where a neural network writes a finished answer instead of the usual ten links. Amid weak reports from other tech companies, Google shares fell 15%. That is roughly $500 billion in market value and the sharpest drop since the IPO in 2004.

Why a ready-made answer sells poorly

Search advertising works simply: a person enters a query, sees ads next to the links, and clicks. Google gets paid per click. When a neural network immediately writes an exhaustive answer, there is nothing to click and no reason to. Swyx, editor of the Latent Space newsletter, puts it this way: any sufficiently good summary removes the reader's need to open the source. Until now, this conflict was only with publishers, whose data Google showed in snippets and knowledge panels. Now it affects Google itself, because its own AI overviews are reducing its revenue.

There is no exact figure for the losses. Google does not break out AI Mode in its reporting, so a rigorous calculation is impossible. Everyone is citing the same Schindler quote, and it is the only direct admission.

A similar signal came from the other side. The head of Reddit said AI Mode "does not send traffic." According to him, search referrals fell by more than 1 million a week in October, and the company is moving users toward logging in with an account. For sites that live on search traffic, this is the same mechanism seen from their side of the screen.

The bad news did not come only from search

Alphabet's drop coincided with worsening sentiment around AI in the stock market. Several facts from the same news cycle:

• OpenAI is reportedly raising money at a $1.25 trillion valuation instead of the expected $1.4 trillion. For a company preparing for one of the largest IPOs in tech history, that is an unwelcome signal.
• Anthropic has filed an S-1 to list on Nasdaq under the ticker "ANT." According to a summary of the document, revenue for the first half of 2026 was $16.3 billion, compared with $1.2 billion a year earlier. Compute costs fell from about 6.5 times revenue to about 0.5 times revenue. These figures come from a summary, not from the document itself.
• Intel guided fourth-quarter revenue of $15–16 billion, against market expectations of about $17.1 billion. It cut its gross margin outlook from 44.1% to 38%. Data center and AI revenue grew about 80% to $7.3 billion, but is constrained by supply shortages through the first half of 2027. Management expects the PC market to shrink by about 12% in 2026 and by 10% in 2027, because AI demand is pushing up memory prices.

Separately, a SemiAnalysis breakdown of Anthropic's finances, relayed by one of its authors on social media, is telling (the estimate is unconfirmed). By this estimate, subscriptions bring in about 18% of revenue but consume more than half of compute resources. API margins, meanwhile, are around 92%. In other words, AI companies share the same problem: the product people like most earns the least per unit of compute.

What this means

Google's situation is paradoxical. Its technical results are strong: in the Arena rankings, the Gemini 4 Carbon model took first place in text tasks and second among agents. But the gain in answer quality costs more than the gain in revenue. The old "link plus ad" model works better than the new "answer without a click."

For practitioners, there are three takeaways:

• Website owners should assume that search traffic will keep falling and think about direct audiences, subscriptions, and logins.
• Companies building AI products need an answer to the question of how to earn money from an answer with no click. Ads inside the text, paid plans, and enterprise contracts currently look like the three main options.
• Investors and engineers calculating model economics should remember that growth in query volume does not equal revenue growth if revenue per query is falling.

One quarter is not enough to talk about an "AI bear market." But Google has, for the first time, said out loud the price it pays for moving from links to answers.