Alphabet: the best quarter and the worst reaction
Alphabet's revenue was $127.6 billion, up 25 percent from a year ago, and Sundar Pichai said the company is now on "a half-trillion dollar revenue run rate." Operating income was $40.6 billion, up 30 percent. Google Cloud brought in $30.7 billion, up 103 percent, and Cloud's operating income tripled to $11.1 billion. That's the number everyone wanted, and Alphabet delivered it. (Alphabet's release, via the SEC)
The stock still fell about 8 percent after hours, from a close of $324.89 to under $300. (StockAnalysis) Three reasons, and only one of them is about the business.
- Earnings per share were $2.12 against about $3 expected. That looks like a disaster. It mostly isn't. Alphabet took an $8.7 billion loss on the stocks it holds in other companies, versus a $10.7 billion gain the same quarter last year. The release says that swing alone cost 55 cents a share. Operating income, which ignores those marks, grew 30 percent. If you own Alphabet for Search and Cloud, this line is noise. It's loud noise.
- Search grew 12 percent to $63.4 billion, about $1.5 billion short of estimates. This one is real. Search is still the profit engine that pays for everything else, and it's growing slower than the Street modeled.
- Capital expenditures were $51.4 billion in three months. Operating cash flow was $42.3 billion. So free cash flow was negative $9.1 billion, the second negative quarter in a row after minus $5.9 billion in the spring. Interest expense went from $143 million a year ago to $967 million. That's what borrowing to build data centers looks like on a cash flow statement.
Nine months into the year Alphabet has spent $132 billion on capex. Its standing guidance from July is $195 billion to $205 billion for 2026, with a "significant increase" in 2027 and no number attached. The release didn't update that. The call, which started at 4:30 Eastern, is where it would come up, and I wrote this before it ended. If the 2027 number is north of $300 billion, which is where the analyst chatter is, that explains the reaction better than any of the above.
Microsoft: a beat everywhere, and new labels
Microsoft's revenue was $92.3 billion, up 19 percent, against about $90.7 billion expected. Earnings were $4.89 a share on the adjusted basis, against about $4.72. Microsoft Cloud was $63.9 billion, up 30 percent. Azure grew 48 percent, which beat Microsoft's own guidance of 44 to 45 percent and is faster than last quarter. Commercial backlog, the contracts signed but not yet billed, is $677 billion, up 73 percent. Amy Hood's line was that results "exceeded expectations across revenue, operating income, and earnings per share," and for once that's just a description. The stock rose about 4 percent after hours. (Microsoft, Yahoo Finance)
One housekeeping note if you read these reports yourself. Microsoft reorganized its segments. "Intelligent Cloud" is gone. The business now reports as "Agents and Infra," $76.8 billion, up 25 percent, and "Devices and Consumer," $15.5 billion, down 3 percent. Any summary still quoting Intelligent Cloud is working from an old template.
Microsoft's capex was $39.4 billion on the cash flow statement, double a year ago. Hood had guided to "over $50 billion" including leases, and the release doesn't break leases out, so I'm not going to call that an undershoot until the all-in number shows up on the call. Two days before the report, Microsoft reportedly offered buyouts to as many as 60,000 eligible employees. The data centers aren't the only place it's finding money.
What the two reports say together
Put them side by side and the AI trade looks like this. The cloud revenue is real and it's accelerating: Google Cloud doubled, Azure sped up to 48 percent. Customers are paying. But the companies are spending faster than the revenue arrives. Alphabet is cash-flow negative for the second quarter running. Meta was cash-flow negative last week for the first time. Microsoft is still positive, and the market rewarded the one company on the list that can fund the buildout out of pocket.
That's also what the IMF said two weeks ago, in plainer language than I expected from the IMF. Its October financial stability report lists an AI stock correction as one of three big risks, notes that the hyperscalers went from under 1 percent of global investment-grade bond issuance to 10 percent since 2024, and counts 14 "circular" deals where chip makers, model companies and cloud providers invest in each other's revenue. One IMF official said "the debt may be difficult to service." (CNBC) Alphabet's interest expense going up seven times in a year is that sentence with numbers in it.
And the money got more expensive this month. The 10-year Treasury is just above 5 percent, after topping 5.1 percent earlier in October for the first time in 19 years, and the Fed held rates on Wednesday. (Trading Economics) Every bond a hyperscaler sells to build a data center now competes with a government bond paying 5 percent. That's the environment, and it's the opposite of 2023's.
What I was sent, and what was wrong with it
- It was a preview, and the results were out an hour after it arrived. The results are above.
- "Alphabet directed nearly $45 billion toward AI infrastructure this year." $44.9 billion was one quarter, the spring one. The year so far is $132 billion.
- "Alphabet is only up 3 percent this year." It closed Wednesday up about 4 percent for 2026. At the after-hours price it's down for the year.
- "The 10-year Treasury yield hovering near 4.99 percent." It's just over 5. Close enough, but the source cited doesn't exist as far as I can find.
- "Warnings from central banks over debt-financed computing buildouts." I found one such warning, from the IMF, which isn't a central bank. Nothing from the Fed, the ECB or the Bank of England this month. I keep getting handed this sentence and I keep failing to find a source for it.
- "Azure vs. Google Cloud metrics" and "Intelligent Cloud." Microsoft doesn't report that segment anymore.
What I'm doing
Nothing, again. I own both through broad index funds, and an 8 percent after-hours drop on a quarter where operating income grew 30 percent is the kind of move I've learned to read about rather than trade. Amazon reports Thursday after the close. (Correction, October 31: I originally wrote that Apple also reported Thursday. Apple reports Monday, November 2.) Amazon is the other company that's cash-flow negative on AI spending, and its number will finish the picture Meta started last week.
If you're tempted to do something, here's the one filter I'd use. Ask whether the thing that scared you is a mark on a stock portfolio, a slower Search quarter, or a 2027 capex number. Only the third one changes what these companies are worth in five years, and we will know it by Thursday morning. I'm not a financial advisor, and this isn't advice for your situation. It's how I'm thinking about mine.
Verified October 28, 2026, from the companies' own releases before their earnings calls: Alphabet's Q3 2026 release and Microsoft's fiscal Q1 2027 release. Consensus estimates and after-hours moves from 24/7 Wall St, Yahoo Finance and StockAnalysis; after-hours prices move, so treat them as the picture at about 4:30 p.m. Eastern. IMF findings from CNBC's coverage of the October 2026 Global Financial Stability Report. Capex guidance figures are the companies' July statements; anything said on Wednesday's calls came after this was written.
Comments
Be kind and stay on topic.
No comments yet. Be the first.