AI Is Making Meta Richer and Google Poorer. So Why Did Both Stocks Fall in October? The Two Charts Behind Last Week’s US$5 Trillion AI Sell-Off

For twenty years Google and Facebook were the same trade: two companies that sold advertising, grew every year, and printed cash. Last week the AI era pulled them apart. Meta’s advertising grew 26% in the third quarter, its fastest in years, because AI now writes, targets and tunes the ads. Google’s search advertising grew 12%, its slowest in years, because people are asking chatbots instead of clicking blue links. The gap between them, 14 percentage points, is the widest in Meta’s favour since 2021.

So you would expect Meta’s shares to be up and Alphabet’s down. Alphabet’s are: a 12.6% fall on Thursday, the worst day in its history, erased about US$500 billion. But Meta’s shares fell 14% in October too, their worst month since 2022. If AI is making one richer and the other poorer, why are both stocks falling? The answer is the most useful thing a first-time investor can learn from this week, and it has nothing to do with who has the better AI.

📌 Key points

• Meta’s ads grew 26% to US$63.1 billion; Google’s search ads grew 12% to US$63.4 billion. The two businesses are now the same size, and Bernstein expects Meta to pass Google Search before the year ends.
• AI helps Meta because it makes ads work better: nearly 10 million advertisers used Meta’s AI tools last quarter, and those who followed its recommendations got almost 5% more conversions and spent 9% more.
• AI hurts Google because chatbots and Google’s own “AI Mode” answer the question without a click, which Google admits earns less per query.
• Both companies are spending more than they earn: Meta’s capital spending was 106% of its operating cash flow last quarter, Alphabet’s free cash flow was minus US$9.1 billion. Both are selling shares to fund it. That, not AI, is why both stocks fell.
• Forty-seven AI and chip stocks have lost about US$5 trillion since 6 October while the rest of the US market gained. For a Malaysian trader the lesson is to separate “good business” from “good stock”.

Two charts, two different stories

	Meta (Q3 2026)	Google Search (Q3 2026)
Ad revenue	US$63.1bn, +26%	US$63.4bn, +12%
What drove it	Impressions +10%, price per ad +14%	Record queries, but less money per query
AI’s role	Writes, targets and bids ads; lifts conversions	Answers questions without a click; “lower monetisation”
Capex this quarter	US$34.9bn (106% of operating cash flow)	US$51.3bn
Free cash flow	−US$2.1bn	−US$9.1bn
2026 capex plan	US$135–145bn	US$195–205bn
Share price, October	−14%	−12.6% in one day, −31% from high

Why AI makes Meta’s business better

Meta sells one thing, advertising, and it is 98% of revenue. An advertiser’s only question is whether the ad produces a sale. AI has made Meta’s answer to that question better in three ways. Its ranking models decide which ad each of its 3.6 billion daily users sees; the latest version, GEM, lifted conversions on app ads by 9.6% in a quarter. Its Advantage+ system automates the targeting and budget, and Meta says advertisers using it earn US$4.52 for every dollar spent, 22% more than campaigns set by hand. And its Meta AI business assistant now builds campaigns for small advertisers: chief financial officer Susan Li said those who took its recommendations “drove almost 5% more conversions” and then “increased their advertising investment by 9%”.

That last sentence is the whole model. Better results make advertisers spend more, which is why the average price of a Meta ad rose 14% in a year even as the number of ads rose 10%. Nearly 10 million advertisers used at least one of these AI tools in the quarter. Mark Zuckerberg’s goal is ad creation “fully automated by the end of 2026”, which means a Malaysian kedai owner typing a sentence and a budget and letting the machine do the rest. Bernstein, the research house, calculated in August that Meta took nearly half of every new dollar spent on digital advertising and was on track to pass Google Search in ad revenue before the end of the year. On last week’s numbers it is almost there.

Why AI makes Google’s business worse

Google’s search business runs on a 25-year-old bargain: you type a question, Google shows ten links and some ads, and you click. Chatbots break the bargain. Ask ChatGPT, Claude or Google’s own Gemini a question and you get an answer, not a list, and nothing to click. Google has responded by putting AI answers inside search itself, through AI Overviews and a conversational “AI Mode”. It keeps the user, but its chief business officer Philipp Schindler admitted on the earnings call that AI Mode queries are eating into traditional search and produce “lower monetisation”. The company’s own data makes the point: chief executive Sundar Pichai said query growth was the highest in five years, yet search revenue grew only 12% against the 15% analysts expected. More questions, less money per question.

Nate Elliott of the research firm Emarketer put the fear plainly: “people’s search habits might change for the first time in a generation”, which would be “potentially catastrophic for Google’s ad business”. UBS said the quarter would “bring Search cannibalization fears right back to the forefront”, and Wedbush wrote that “GOOGL shares are unlikely to work until the search narrative changes”. Google’s defence is real: it is now putting ads inside AI Mode, deploying Gemini “across our entire ads infrastructure”, and its cloud division doubled to US$30.7 billion selling computing power to AI companies including Anthropic. Advertising is now less than two-thirds of Alphabet’s revenue. But search is still the engine, and the engine slowed.

So why did Meta fall too?

Because the market has stopped paying for growth and started asking for cash. Meta’s capital spending reached US$34.9 billion in the quarter, more than its entire operating cash flow, and its free cash flow turned negative. It plans to spend US$135 to 145 billion this year, and S&P expects more than US$225 billion next year, which is why the rating agency put Meta’s credit outlook on negative. To pay for it Meta is selling US$37.5 billion of new shares and preferred stock, has created a new non-voting class of shares to sell up to US$60 billion more from January, and cut 15% of its staff. Alphabet is doing the same at a larger scale: US$51.3 billion of capex in the quarter, a share sale of US$85 billion announced in June, and more than US$1 trillion of future commitments for data centres, chips and power.

Dec Mullarkey of SLC Management described the mood: “Markets are carefully calibrating the trade-off between capex and free cash flow.” On that scale, Meta’s better AI story bought it a smaller fall, not a rise. The whole sector was repriced: by Yahoo Finance’s count, 47 AI and chip stocks lost about US$5 trillion between 6 October and the end of the month, with Nvidia and Alphabet losing about US$834 billion each, while the other 625 large US stocks gained US$1.3 trillion. The Nasdaq fell 2.9% on Friday alone, its worst day since June, and 5.2% for the month. Apple, the one giant that is not building data centres, is now worth more than Nvidia and is the only one of the “Magnificent Seven” beating the index this year.

What it means for a Malaysian trader

• A better business is not automatically a better stock. Meta had the best advertising quarter of the big four and still fell 14%. Price is set by what investors will pay for future cash, and both companies are spending that cash now. Judge the stock on cash flow, not on whose AI is cleverer.
• If you advertise, this is good news. Malaysian small businesses spend most of their marketing on Facebook and Instagram. The tools that lifted conversions 5% for Meta’s advertisers are the same ones in your Ads Manager. Google’s AI Mode, meanwhile, means fewer clicks from search, so a business that relies on Google traffic should watch its numbers.
• Google’s spending lands here. Alphabet’s capex includes its first Malaysian data centre and cloud region in Selangor. The spending the market is punishing is the spending Malaysia is courting; a real cut would show up in projects like that.
• Trade the index with the sector in mind. Nasdaq-100 moves last week were driven by a handful of AI names. The futures opened 0.9% higher on Monday morning Malaysian time; Apple and Palantir report after the US close tonight, the midterms are on Tuesday, and the jobs report on Friday. Four known events in five days is a week to size positions small.
• Valuation now differs sharply. Alphabet trades at about 15 times trailing earnings after the fall, Meta at about 21. The cheaper one has the bigger question over its core business; the dearer one has the clearer AI payoff and the bigger bill. That is the trade the market is still deciding.

FAQ

Why did Alphabet’s stock fall 12% in one day?

Google Search grew 12% in the third quarter of 2026 against 15% expected, the company said AI Mode queries earn less than traditional search, free cash flow was minus US$9.1 billion, and 2026 capex guidance is US$195 to 205 billion. Shares fell 12.6% on 29 October, the worst day since its 2004 listing.

Is AI good or bad for Meta?

Good for the business so far. AI tools lifted ad conversions, raised the average price per ad 14% and drove 26% ad revenue growth. The cost is capital spending of US$135 to 145 billion in 2026, which pushed free cash flow negative and led Meta to sell new shares.

Why did Meta’s stock fall in October 2026 if its results were strong?

Investors are pricing the cost of AI, not just the benefit. Meta’s capex exceeded its operating cash flow, it announced a US$37.5 billion share and preferred stock sale, and S&P put its credit outlook on negative. The stock fell 14% in October after rising 27% in September.

Is ChatGPT taking users from Google?

Google says search queries are at a record, so users are not leaving. What is changing is how questions are answered: chatbots and Google’s own AI answers remove the click that earns money. Industry estimates put ChatGPT at under half of AI assistant usage, with Google’s Gemini second.

What does this mean for Malaysian investors?

For traders, last week showed the Nasdaq is being driven by a few AI stocks and their spending, not by the wider market, which rose. For business owners, Meta’s AI ad tools are working, while Google’s AI answers mean fewer clicks from search. Google’s Malaysian data centre is part of the capex now under scrutiny.

Summary

AI is pulling the two advertising giants apart. Meta’s ads grew 26% because AI makes them work better and advertisers pay more for results; Google’s search grew 12% because AI answers questions without the click that earns the money. Yet both stocks fell in October, Alphabet by 12.6% in a day and Meta by 14% for the month, because both are spending more than they make on data centres and selling shares to cover it. The market has moved from rewarding growth to demanding cash, and 47 AI stocks have lost US$5 trillion since early October while everything else rose. For a Malaysian trader, the lesson is to separate the business from the stock; for a Malaysian business, Meta’s tools are the ones to learn and Google’s clicks are the ones to watch.

Disclaimer

This article is for education only and is not investment advice. Figures are from company reports, analyst notes and market data as of 2 November 2026, before the US market open; prices will have moved. Do your own research or consult a licensed adviser before investing.

Related reading

• Google lost US$480 billion in one day for spending too much on AI: what Big Tech’s results say about the Nasdaq
• Why is AI surging, and why is crypto becoming a payment asset?
• Nvidia’s newest chip, Rubin: a game changer for AI tokens?
• Investing is not the same as trading
• America votes on Tuesday: why a US midterm election can move your stocks