Why Alphabet lost about $500 billion last week

Search ads still grew. The stock fell because they grew more slowly, AI Mode earns less per query, and the AI buildout spent the cash.

Alphabet reported third-quarter results after the close on Wednesday 28 October 2026. On Thursday the shares fell about 13%, the worst session since the 2004 IPO. CNBC put the loss at roughly $500 billion of market value and noted the previous record drop was 11.63% in March 2020. The BBC measured the company at $3.97 trillion before the session and $3.46 trillion at the end of it.

The headlines blamed Google's own AI summaries. The transcript is more specific, and the cash-flow line mattered as much as the ads.

What the quarter showed

Consolidated revenue was $127.6 billion, up 25%, the thirteenth straight quarter of double-digit growth. Adweek said that sat under a $127.93 billion consensus. Google Services, which holds ads, subscriptions, and devices, rose 11% to $96.3 billion. Advertising was $82.0 billion. Adweek said analysts had looked for $83.5 billion.

The line investors circled was Search & other: $63.35 billion, up 12% from $56.57 billion a year earlier. In the second quarter the same line was $63.27 billion and the growth rate was 17%, which Alphabet stated on that quarter's call. Search ads barely moved from one quarter to the next.

The rest of the report pulled in two directions. Google Cloud was $30.7 billion, up 103%. Operating margin was 31.8%, and earnings were $2.12 a share. A loss on equity securities of $8.7 billion cut diluted earnings by $0.55. The filing records the loss and does not name the holdings. Purchases of property and equipment were $51.4 billion, and free cash flow was negative $9.1 billion, after negative $5.9 billion in the second quarter. CNBC said analysts had expected a deficit of about $2.7 billion.

Chief financial officer Anat Ashkenazi left 2026 capital-spending guidance at $195 billion to $205 billion and said 2027 spending should still rise significantly. CNBC reported a miss on revenue and profit.

Why search growth slowed

On the call, J.P. Morgan's Doug Anmuth asked why query growth was the fastest in five years, including commercial queries, while search revenue growth had eased. Were there simply fewer ads in AI Overviews and AI Mode?

Philipp Schindler, chief business officer, answered in order. Foreign exchange was the single largest driver of the slowdown versus the second quarter. Some Asia-Pacific retail advertisers pulled back after the European Union removed its de minimis exemption on low-value imports. Media and entertainment spending cooled after a second quarter inflated by the World Cup, which put more matches, and more ad budgets, in June than in July.

Then he named the product:

"It's worth noting that the strong growth in AI Mode queries had an impact because of the lower monetization we were seeing in AI Mode."

He did not say how much lower, or how many dollars that represented. He said ad experiments looked promising and that Google feels good about the path. Chief executive Sundar Pichai framed the sequence the way he framed mobile a decade ago: build the search experience first, monetize after.

AI Mode queries run about three times as long as a traditional search, and one in five uses voice, an image, or a document. Schindler said many of those questions are upper-funnel, and that most would not have been asked on the old results page. They can become inventory later. This quarter they diluted the average.

AI Overviews are not that line

AI Overviews and AI Mode get collapsed into one phrase, "AI summaries," in a lot of the coverage. Google treats them as one search experience. It does not claim they earn the same amount.

AI Overviews are the snapshot at the top of a normal results page. AI Mode is the conversational surface, with follow-up questions. At Google Marketing Live in May 2025, Shashi Thakur, who runs search ads, told reporters including Search Engine Journal that AI Overviews were generating revenue "at the same rate" as those same searches without an overview. That claim was about Overviews. It did not cover AI Mode. On last week's call, the monetization drag Schindler named was AI Mode.

Google is still putting more ads into both. Schindler said ads in AI Overviews and AI Mode are scaling to more users and more queries. The same day, Thakur wrote that AI Mode ads are expanding to ten more countries, including Australia, Canada, and India, with Brazil, Mexico, and the United Kingdom planned for the fourth quarter. In the United States, a follow-up question inside the same conversation can now carry an ad. Ads are labeled Sponsored. Google's U.S. data for 20 August through 3 September says click-through on sponsored product listings inside AI Mode already matches organic product listings. That comparison is product cards, not every link in the answer.

Source: Google Lens and AI Overviews for marketers, Google.

The illustration is the business Google is protecting: a search, then a product. Overviews were built to answer and still send the click. AI Mode answers in a thread, and the ads are catching up to it. Existing Search, Shopping, and Performance Max campaigns are eligible for those placements by default. There is no separate opt-out.

Kim Scott, who once led Google's ads business, told the BBC the company would rather eat its own lunch than hand the question to someone else. A slower search-ad quarter is the cost of keeping that question on Google, while OpenAI says ChatGPT already handles about a fifth of the lookup volume. Why that claim is not yet a fifth of the ad revenue is the companion piece.

The cash left through the other door

Search still grew. It grew more slowly, and Google put AI Mode behind currency and two category swings as a reason. The session that took off about $500 billion was also about cash. Free cash flow has been negative for two quarters. This year's spending guide did not come down, and next year's is higher. Ashkenazi said the third-quarter deficit also reflected inventory for TPU system sales and payment of a European Commission fine.

Forrester analyst Tracy Woo told the BBC the quarter shows how hard it is to fund the AI build while the old business stays as strong as it was. Query growth at a five-year high, with search revenue nearly flat versus spring, is the number to keep.

ThinkReview posts the review on the diff itself, on GitHub, GitLab, Azure DevOps, and Bitbucket. The model is a choice in Model selection.