Alphabet's $9 billion cash burn: the quarter AI spending outran search

A record cloud quarter, punished by the market anyway

Google parent Alphabet reported its third-quarter results after the market closed on Wednesday, October 28, 2026, and the market's verdict arrived within minutes: shares fell more than 10% in after-hours trading, sliding toward $288 after closing above $324. It was a strange kind of punishment. Revenue rose 25% from a year earlier to $127.6 billion, Google Cloud more than doubled, and operating income climbed 30% to $40.6 billion. Yet the two numbers investors watch most closely both missed: total revenue came in just below the roughly $128 billion Wall Street expected, and earnings per share of $2.12 fell far short of the $3.02 consensus. The miss, and the spending behind it, is the story of this quarter.

The trigger was not a collapse in the business. It was a collision between two truths Alphabet now lives with: its artificial-intelligence buildout is working as a technology story and straining as a financial one. The company spent $51.4 billion on capital expenditure in three months, up 114% from a year earlier, and burned $9.07 billion in free cash flow; the second straight quarter in negative territory, after a $5.9 billion outflow in the prior quarter that was itself the first negative reading in Alphabet's 22-year history as a public company.

The ledger, line by line

The revenue miss sat squarely in advertising. Google's advertising business generated $82.01 billion, about $1.41 billion short of expectations, according to LSEG data. Google Search and other revenue rose 12% to $63.35 billion; solid in isolation, but below the roughly $65 billion analysts expected, and a slowdown from 17% growth in the previous quarter. YouTube advertising grew 9% to $11.2 billion, also slower than the 13% of the prior quarter. Network revenue was nearly flat at $7.5 billion. In total, Google Services revenue rose 11% to $96.3 billion, below expectations of about $98.5 billion.

Cloud was the exception that proved the rule. Google Cloud revenue surged 103% to $30.71 billion, blowing past the $28.8 billion expected, and its operating income more than tripled to $11.1 billion. Chief Executive Sundar Pichai said the company is now on a half-trillion-dollar revenue run rate, and noted that search query growth is running at its fastest pace in five years. The business underneath is still expanding at a remarkable rate; the argument is about what that expansion costs and how quickly it pays.

Below the operating line, the earnings miss has a specific cause. Alphabet booked $8.22 billion in net other expenses, against $12.76 billion of other income a year earlier, after a paper loss on equity securities that the company attributed to its investment in SpaceX. Net income available to common shareholders fell 25% to $26.17 billion. On an adjusted basis, earnings per share of $2.67 also missed the $3.02 target. Management held its 2026 capital-expenditure forecast at $195 billion to $205 billion, but warned that spending will increase significantly in 2027, without putting a number on it, and flagged near-term margin pressure from using third-party cloud capacity.

Why does AI Mode pay less per query than classic search?

The central unanswered question is monetization. Google's chief business officer, Philipp Schindler, told analysts that queries in the new AI Mode are monetized at a lower rate than in traditional search, even as overall query growth hits five-year highs. The company says it is expanding AI Mode monetization now, but it has not said how fast that gap closes, or what a mature AI search result is worth compared with the classic list of links and ads. Until that number moves, every quarter will invite the same argument: more usage, less revenue per use.

Nor do we know the size of the 2027 bill. Ashkenazi confirmed the direction, significantly higher, but not the figure, leaving investors to guess how far beyond the $195–205 billion planned for 2026 the company will go. We do not know whether the SpaceX paper loss reverses, how much of the cloud surge reflects one-off hardware sales of Google's in-house TPU chips rather than recurring cloud contracts, or how long third-party capacity costs will weigh on margins. Each of those unknowns cuts in a different direction, which is why the same set of results can support both a bull case and a 10% selloff.

Search pays for the AI race; so search slowing is the whole story

Alphabet matters beyond its own shareholders because its search advertising has, for two decades, been one of the great cash machines of the technology industry; the engine that funds everything else, from moonshots to the AI race itself. When that engine decelerates while spending more than doubles, it tests the central bargain of the AI era: that today's enormous infrastructure bill will be repaid by tomorrow's AI revenues. Alphabet is not alone in that bargain. Together with Amazon, Microsoft and Meta, the big hyperscalers are projected to spend well over $700 billion in 2026 on data centres, chips and networking. A wobble at the company with the strongest advertising franchise therefore reads as a warning for the whole trade.

It also matters for everyone who lives inside Google's ecosystem rather than its stock: advertisers who buy search placement, publishers who depend on search traffic, and the hundreds of millions of users whose daily queries are now answered by AI summaries. If AI Mode answers questions without sending users onward, the economics of the open web shift; and the quarter's numbers are the first hard evidence of how that shift feels in dollars.

What would prove the spending is paying off?

Strip out the paper loss and Alphabet's operating performance is genuinely strong: a 31.8% operating margin, cloud operating income tripled, query growth at a five-year high, the Gemini app past one billion users. The bull case writes itself; the company is buying the infrastructure of the next decade while its rivals scramble, and its own TPU chips are becoming a product in their own right, with media reports this week that OpenAI plans to buy some. That last detail is the most strategically interesting in the whole release: Google selling chips to the company whose chatbot threatens its search franchise.

The bear case is simpler and, on the night, louder: revenue per query is falling just as the cost per query is exploding. AI answers are computationally expensive, and the advertising that pays for them has not yet caught up. Management's refusal to quantify 2027 spending reads, to a nervous market, as an open-ended commitment. The test for the next two quarters is narrow and measurable: search revenue growth re-accelerating toward the high teens, cloud growth holding above 50% without TPU one-offs doing all the work, and free cash flow returning toward break-even. Miss on two of the three, and the “AI is eating search” thesis hardens from fear into fact.

The ripples: chips, advertisers, and the other bets

The second-order effects reach well beyond Mountain View. If Google's TPUs genuinely challenge Nvidia's dominance in AI accelerators, the chip hierarchy of the AI era becomes a contest rather than a coronation, with Samsung, this week's other record-setter, supplying the memory that both sides need. For advertisers, slower search growth and AI answers mean the auction for attention gets re-priced: fewer clicks per query, more competition for the clicks that remain. For Alphabet itself, the “Other Bets” division, including the Waymo robotaxi business; lost $2.0 billion on just $372 million of revenue, a reminder that the company is funding several futures at once. And for regulators, a quarter in which a $9 billion cash outflow barely dents a balance sheet built on a search monopoly will sharpen, not soften, the question of what that monopoly is worth.

The economy around AI is also repricing risk. A $51.4 billion quarterly capital bill, repeated across four hyperscalers, is a macroeconomic event in its own right; a construction boom in data centres, power and chips that shows up in electricity demand, trade flows and employment long before it shows up in AI profits.

Thursday's session, then the 2027 number

The immediate test is Thursday's regular trading session: whether the after-hours fall of more than 10% holds, deepens, or is bought as an overreaction to a paper loss. Beyond that, three dates matter. The first is any update on 2027 capital expenditure; the number Ashkenazi declined to give. The second is the fourth-quarter report, which will show whether AI Mode monetization is closing the gap Schindler described. The third is the cloud backlog: whether the 103% surge sustains once TPU hardware sales are stripped out. Investors have given Alphabet the benefit of the doubt for years because search always paid. This quarter, for the first time in a long while, they asked to see the receipt. The finance world will now watch whether the company can produce one.