Market Pulse
Mixed observation times, Oct 27–Oct 30, 2026 (3 days apart)
Stocks · 7 min

Alphabet fell 12.64% in a day. The release shows capex above cash flow

Alphabet fell 12.64% on October 29, 2026, larger than all but 6 of 5,583 sessions since 2004. Revenue rose 25%; capex of $51,355m topped operating cash flow.

◆ The Signal
30-second read
What happened
Alphabet shares closed 12.64% lower on October 29, 2026, the first session after the company released its third-quarter results dated October 28, 2026. The site's Market Memory page counts 6 sessions out of 5,583 since 2004 with a move that large or larger. The release reports revenue of $127,577m, up 25%, and Google Cloud revenue of $30,709m, up 103%.
Why
The same release shows purchases of property and equipment of $51,355m in the quarter, against operating cash flow of $42,290m, which left free cash flow at −$9,065m. An $8.7bn loss on equity securities took diluted EPS to $2.12 from $2.87. The income statement and the cash-flow statement tell different stories, and the share price moved after both were published.
The numbers that matter
−12.64%Alphabet share price move, October 29, 2026 close (Market Memory, data date Oct 29, 2026)
$51,355mPurchases of property and equipment, quarter ended September 30, 2026 (vs $23,953m a year earlier)
−$9,065mFree cash flow, quarter ended September 30, 2026 (Alphabet, non-GAAP reconciliation)
14.47%Rise needed to recover a 12.64% fall (calculation)
Why it matters
A 12.64% fall needs a 14.47% rise to get back to even; the 29.5% drawdown from the high needs 41.8%. Losses and recoveries are not symmetric.
What happens next
Alphabet's quarterly dividend of $0.22 per share is payable December 14, 2026 to holders of record on December 7, 2026. The 10-Q for the quarter ended September 30, 2026 follows the release on EDGAR.
Go deeper ↓
Alphabet
Alphabet purchases of property and equipment by quarter, Q3 2025 to Q3 2026

Quarterly capital spending more than doubled in a year and, at $51.4bn, exceeded the $42.3bn of operating cash flow in the quarter ended September 30, 2026.

Alphabet purchases of property and equipment by quarter, Q3 2025 to Q3 2026
Q3 202524.0
Q4 202527.9
Q1 202635.7
Q2 202644.9
Q3 202651.4
US$ billions, per quarter; marker is operating cash flow in Q3 2026Source: Alphabet Inc., Q3 2026 earnings release, free cash flow reconciliation, dated October 28, 2026
Go deeper

Alphabet closed 12.64% lower on October 29, 2026. In 5,583 trading sessions since 2004, the site's Market Memory page for Alphabet finds only six moves that large or larger in either direction. The two nearest analogues are March 16, 2020 (−11.63%) and September 29, 2008 (−11.61%), both days of system-wide stress. October 29, 2026 was not one of those: the site's Market Memory page for the S&P 500, which uses the SPY ETF as its proxy, puts the index at −0.38% on the same date.

The session followed the company's own third-quarter release, dated October 28, 2026. Read as an income statement it is a strong quarter: revenue up 25%, Google Cloud revenue up 103%, operating margin 31.8%. Read as a cash-flow statement it is something else: purchases of property and equipment of $51,355m against operating cash flow of $42,290m, and free cash flow of −$9,065m. This piece walks through what the filing says, the historical context, and what a fall of this size means in loss-recovery arithmetic. It does not say whether the move was right.

What the income statement says

Data
Revenue $127,577m, up 25%; Google Cloud $30,709m, up 103%; operating margin 31.8%
Alphabet's release for the quarter ended September 30, 2026 reports total revenues of $127,577m, against $102,346m in the quarter ended September 30, 2025, a rise the company states as 25%. Google Cloud revenue was $30,709m against $15,157m; the release says Cloud “revenues increased 103% to $30.7 billion”. Google Services revenue was $96,302m against $87,052m. Income from operations was $40,576m against $31,228m; the release prints the operating margin as 32% against 31%.
Wealth Signal calculation
Operating margin to one decimal: 31.8%, up from 30.5%
$40,576m ÷ $127,577m = 31.8%. A year earlier, $31,228m ÷ $102,346m = 30.5%. Total costs and expenses rose from $71,118m to $87,001m, or 22.3%, slower than the 24.7% rise in revenue computed from the same two figures ($127,577m ÷ $102,346m − 1). Google Cloud operating income rose from $3,594m to $11,121m.
Data
Net income fell to $26,468m and diluted EPS to $2.12, with an $8.7bn equity-securities loss
Net income was $26,468m against $34,979m a year earlier, and diluted net income per share was $2.12 against $2.87. The reason is below the operating line: “Other income (expense), net” was −$8,221m against +$12,759m. The release's footnote states that the net effect of the loss on equity securities of $8.7 billion decreased the provision for income tax, net income and diluted net income per share by $1.9 billion, $6.7 billion and $0.55, respectively.
Wealth Signal calculation
Without the $0.55, diluted EPS would have been about $2.67, still below the year-ago $2.87
$2.12 + $0.55 = $2.67. The year-ago quarter's $2.87 itself included “other income” of $12,759m. Mark-to-market gains and losses on equity securities flow through net income each quarter, so the EPS comparison is dominated by them rather than by advertising or Cloud. See adjusted earnings.

What the cash-flow statement says

The cash-flow statement is where the quarter changes character. Operating cash flow fell while capital spending more than doubled, and the gap between them went negative for the second quarter in a row.

Alphabet, quarter ended September 30, 2026 versus quarter ended September 30, 2025
Line (US$ millions)Q3 2025Q3 2026Change
Total revenues102,346127,577+25%
Google Cloud revenues15,15730,709+103%
Income from operations31,22840,576+30%
Net income34,97926,468−24%
Diluted EPS$2.87$2.12−26%
Net cash provided by operating activities48,41442,290−6,124
Purchases of property and equipment23,95351,355+27,402
Repurchases of stock11,5040−11,504
Number of employees190,167204,757+7.7%
Source: Alphabet Inc., Q3 2026 earnings release dated October 28, 2026, income statement, cash-flow statement and segment tables. Percentage changes computed from the figures shown; the release states +25% and +103% for revenue and Cloud.
Data
Free cash flow was −$9,065m in the quarter and $19,747m over twelve months
The release's non-GAAP reconciliation shows net cash provided by operating activities of $42,290m for the quarter ended September 30, 2026, against $48,414m a year earlier, and purchases of property and equipment of $51,355m against $23,953m. Free cash flow, defined by the company as the difference, was −$9,065m. The same table gives the four most recent quarters: $24,551m (Q4 2025), $10,116m (Q1 2026), −$5,855m (Q2 2026) and −$9,065m (Q3 2026), summing to $19,747m for the trailing twelve months. Trailing operating cash flow was $179,551m and trailing capital spending $159,804m.
Wealth Signal calculation
Capital spending was 121% of operating cash flow in the quarter, against 49% a year earlier
$51,355m ÷ $42,290m = 1.214, so the company spent $1.21 on property and equipment for every $1.00 its operations generated. In the quarter ended September 30, 2025 the ratio was $23,953m ÷ $48,414m = 0.495. Over the trailing twelve months it was $159,804m ÷ $179,551m = 0.890. The ratio has risen in each of the five quarters the release shows: 49.5%, 53.1%, 77.9%, 115.0%, 121.4%.
Interpretation
Investors buying the stock are buying the spending programme as much as the advertising business
Free cash flow is what is left for dividends, buybacks and the balance sheet after the company has paid for its own growth. In the quarter it was negative, and repurchases of stock were $0 against $11,504m a year earlier. Cash, cash equivalents and marketable securities nonetheless rose from $126,843m at December 31, 2025 to $265,098m at September 30, 2026, and the release lists a dividend on mandatory convertible preferred stock, which points to financing rather than operations as the source. The 10-Q carries that detail; the site's guide to reading an earnings release explains which document holds what.

How rare the move was

The Market Memory page computes the daily move from closes of $324.89 on October 28, 2026 and $283.82 on October 29, 2026: $283.82 ÷ $324.89 − 1 = −12.64%. Its data date is October 29, 2026, and its own text carries the caveat the site applies to every such page: historical context, not a forecast.

−12.64%
Alphabet, October 29, 2026
99.9th percentile of daily moves since 2004
6 of 5,583
Sessions since 2004 this large or larger
either direction; among declines alone, none larger
−29.5%
Drawdown from high at Oct 29, 2026 close
20-day trend −16.1%; 20-day realised volatility 49.1%
Data
Volume was 11.4 standard deviations above its 20-day average, and only two past days qualify as analogues
The page reports a volume z-score of 11.4 σ against the prior 20 sessions, and 20-day realised volatility of 49.1% annualised, higher than 93% of the record. It finds two similar days within 4.19 percentage points of the move: March 16, 2020 (−11.63%) and September 29, 2008 (−11.61%). After those two days the stock was higher one day later in both cases, lower one week later in both, and split one against one at one month (+17.18% and −3.22%) and at three months (+32.41% and −21.94%). The page itself says two sessions are “too few for a meaningful summary; treat the figures as anecdote.”
Interpretation
The two analogues were index-wide crashes; October 29, 2026 was a single-stock event
March 16, 2020 and September 29, 2008 were days when nearly every large stock fell hard. On October 29, 2026 the S&P 500 proxy fell 0.38%, with a drawdown from its high of 1.2%. The nearest comparisons for the size of the move are not comparisons for its cause, one more reason to read the two-row outcome table as anecdote. The site's volatility piece covers why a single stock's volatility can detach from the index's.

What it means for your money

A reader does not need a view on Alphabet to use the arithmetic of this move. The percentages a fall and a recovery require are not the same number, and the gap widens as the fall deepens. The site's loss-recovery tool does the sums for any figure; the three that matter here are below.

Wealth Signal calculation
A 12.64% fall needs a 14.47% rise to get back to even
Start with $10,000 of Alphabet stock at the October 28, 2026 close. After −12.64% it is worth $10,000 × (1 − 0.1264) = $8,736. To return to $10,000 it must gain $1,264, and $1,264 ÷ $8,736 = 14.47%. The general formula is 1 ÷ (1 − loss) − 1. Checked against the closes: $324.89 ÷ $283.82 − 1 = 14.47%.
Wealth Signal calculation
A 29.5% drawdown from the high needs a 41.8% rise
The Market Memory page puts the drawdown from the high at −29.5% as of the October 29, 2026 close. $10,000 held from the high is worth $10,000 × (1 − 0.295) = $7,050. Returning to $10,000 requires $2,950, and $2,950 ÷ $7,050 = 41.8%. For comparison, a 10% loss needs 11.1%, a 20% loss needs 25.0%, and a 50% loss needs 100%. The drawdowns glossary entry explains the convention.
  • Concentration. The S&P 500 proxy moved 0.38% on the day Alphabet moved 12.64%. A household holding an index fund experienced a small fraction of the move; one holding the stock directly experienced all of it. See position sizing and diversification.
  • Volatility and horizon. Realised volatility of 49.1% annualised is about 3.1% a day (49.1% ÷ √252). Volatility is not the same thing as risk, but it is uncertainty about the path, and a six-month horizon feels it more than a ten-year one.
  • Taxes. In a taxable account a loss realised in 2026 can offset gains under the rules in the site's wash-sale piece; an unrealised loss changes nothing on a tax return. Neither is a reason to act.

The dividend is unchanged in the release: $0.22 per share on Class A, B and C stock, payable December 14, 2026 to holders of record on December 7, 2026. At the October 29, 2026 close of $283.82 that is $0.88 a year, or 0.31%, which is small relative to a daily move of 12.64% and to the 41.8% recovery the drawdown requires.

What this piece does not claim

The filing, the Market Memory pages and the arithmetic above are all this article uses. It does not use a consensus figure and does not compare the quarter with what anyone expected; it compares it with the company's own prior year. The same method applied to another company is in the site's Nvidia price-tag piece and before-you-buy checklist.

This article does not say whether Alphabet's shares were mispriced on October 28 or October 29, 2026, does not say what they do next, and offers no view on whether the capital-spending programme is sound. The two historical analogues on the Market Memory page are anecdote, as the page itself states, not a pattern. All company figures are from Alphabet's Q3 2026 earnings release dated October 28, 2026; the share-price statistics are from the site's Market Memory page with a data date of October 29, 2026; the loss-recovery figures are arithmetic from those inputs. The October 30, 2026 session had not closed when this was written and no figure from it appears here. Nothing here is investment advice.

Sources

Wealth Signal prefers primary sources — regulators, statistical agencies and company filings. Named secondary sources are used where a primary document does not exist or is not public. Our source standards.

Editor and publisher · Markets, economics, personal finance

Wealth Signal is researched, written, edited and published by one person, under the Wealth Signal name. Every figure links to the document it came from, and errors are corrected on the page and logged publicly.

Related analysis

Cite this: Wealth Signal NYC (October 30, 2026). “Alphabet fell 12.64% in a day. The release shows capex above cash flow.” https://www.wealthsignalnyc.com/stories/alphabet-worst-day-since-2004-what-the-filing-says/ — the figures are the issuing agency’s, listed with their dates in the sources above; quote the agency, link the explainer.

This article is general information and analysis, not investment, tax or legal advice; read the financial disclaimer. Spotted an error? Corrections are made on the page and logged on the corrections page — tell the editor.