Loom/BlogGet the app →
November 2, 2026Alphabet · GOOGL · Alphabet · GOOG · Marvell Technology · MRVL4 min read

Why Alphabet fell 12.6% in one day: AI costs and a shrinking search share

Alphabet lost 12.6% on Oct 29 after its third-quarter report. The report set off the drop, but the two worries behind it, AI costs and search erosion, had been building in the news for a month.

Alphabet had its worst day as a public company on Oct 29

On Oct 29, 2026, Alphabet's Class A shares (GOOGL) fell 12.6%, from $324.89 to $283.82. The Class C shares (GOOG) fell 12.3%, from $321.89 to $282.32. Bloomberg put the lost market value at about $500 billion. Chipmaker Marvell (MRVL) fell 13.2% the same day, from $251.65 to $218.34. I come back to Marvell later. Its link to this story is looser.

Oct 1 → Nov 2
−12.6%Alphabet (GOOGL), $324.89 → $283.82, Oct 29
−12.3%Alphabet (GOOG), $321.89 → $282.32, Oct 29
−13.2%Marvell (MRVL), $251.65 → $218.34, Oct 29
Daily closing prices

The trigger was Alphabet's third-quarter report, released after the close on Oct 28. The FT reported that a $51bn AI outlay deepened cash burn. Search growth slowed. Alphabet kept its full-year spending plan in place.

One idea helps here: free cash flow. It is the cash a business brings in from its operations, minus what it spends on long-lived things like data centers and chips. When it goes negative, the company is spending more than the business produces and has to cover the gap some other way. Alphabet also disclosed that it sold $12.4 billion of its own stock in the quarter, issuing 36 million shares. Selling new shares is one way companies raise cash. It also spreads ownership thinner.

A month of stories about AI costs and the value of a search click

On Sept 26, a Barchart piece argued that Meta's Muse could be the biggest threat Google has faced in 20 years. The idea: an AI agent that makes buying decisions replaces the search, the click and the ad. GOOGL had last closed at $343.92.

Daily close, indexed to 100 on Oct 1
GOOGL, GOOG and MRVL daily closes, indexed, Oct 1 to Nov 2, 2026
608010012012345Oct 1Oct 12Oct 22Nov 2
MRVLGOOGLGOOG
  1. 1Oct 5Airbnb CEO says ChatGPT traffic converts better than Google's
  2. 2Oct 19Gartner analyst stands by view that AI returns are too small
  3. 3Oct 21Pichai says Google Cloud is giving up revenue for AI capacity
  4. 4Oct 27Reports: Google to sell its TPU chips to OpenAI and Oracle
  5. 5Oct 29Q3 report: $51B AI outlay, negative cash flow, slower search
Daily closing prices, Oct 1 – Nov 2.

On Sept 30, a Goldman Sachs note said AI spenders need about $300 billion a year in AI revenue just to break even. Meta fell about 4%. The note named Meta's spending, but the math applies to every large AI spender, Alphabet included.

On Oct 5, Airbnb's CEO said ChatGPT traffic converts better than Google's. That matters because advertisers pay Google a premium for clicks from people ready to buy. If another source sends better buyers, that premium is in question.

Then the cost side came back. On Oct 19, a Gartner analyst stood by his view that AI returns are not yet large enough to justify the build-out. On Oct 21, Alphabet slipped 2% while the Nasdaq rallied, after Sundar Pichai said Google Cloud was giving up revenue to keep scarce computing capacity for frontier AI. On Oct 22, a Meta sell-off dragged tech lower after Meta forecast hundreds of billions of dollars in AI spending.

I followed these through Loom, a news app that links each story to the stocks it's likely to move. Few of these stories were about Alphabet's own numbers. They were about Meta, Airbnb, Goldman and Gartner. What tied each one to Alphabet was that it touched either the cost of AI or the value of a search click.

The other side: chip deals, a new model and a bounce into the report

The stock did not slide in a straight line. Between late September and the report, GOOGL traded as high as $354.97 and as low as $316.80. It closed at $353.78 on Oct 13, a week after the Airbnb comment. By Oct 26 it was at $316.80, down 10.5%. It then rose 2.6% to $324.89 by Oct 28.

That bounce had reasons. On Oct 27, The Information reported that Google would sell its AI chips to OpenAI and Oracle. These are TPUs, processors Google designs in-house, and OpenAI has long bought from Nvidia. CNBC said Google was enlisting partners to sell more of them. Google also launched Gemini 4 Carbon, its second major model in a month. And data-center supplier Vertiv beat sales estimates and raised its outlook. All of that pointed to real demand for what Alphabet was building.

Some of the report itself was strong. The Information wrote that Google Cloud was closing in on Azure, helped by chip sales.

The report put numbers on both worries at once

The reaction was fast. CNBC's Investing Club said the results disappointed at nearly every turn. By midday CNBC reported Alphabet was on course for its biggest ever post-earnings drop, citing larger-than-expected negative free cash flow.

One detail matters. Spending itself did not come in above forecasts. CNBC reported that both Alphabet's and Microsoft's capex fell short of estimates. The problem was what the spending was doing to cash, with no cut to the plan.

Search was the second blow. Yahoo Finance reported that search fell below half of revenue for the first time. Bank of America valued Google search at a lower multiple than fossil fuel stocks. The same afternoon, Reddit said Google search traffic was shrinking.

Marvell is harder to read. After the close on Oct 28, The Information reported Marvell was in talks to buy Lumentum, an optics maker worth more than $80 billion. The next day MarketWatch said investors were worried about the premium and a long China approval. But the whole AI chip group was under pressure that day, after Microsoft's lower capex hit neocloud stocks. I cannot separate the two. Marvell kept falling, to $178.07 by Nov 2. GOOGL closed at $283.19 that day.

What I take from it

The report was the trigger. But the size of the reaction only makes sense against the month before it. By Oct 28, two questions were already in the open: is AI spending paying off, and is AI eating search. The report answered both on the same night, with numbers.

The good news was real too. Chip demand and Cloud growth came in the same release. On this day the cash and search story weighed more. Stories like these tell you which questions a stock is being judged on. They do not tell you when the answer arrives, or which way it will land.

Sources

Prices are daily closes. This is not investment advice.

← All posts