Every business eventually reaches the moment when growth stops being free.
For most of the past decade, the biggest American technology companies looked like they had found a way around that rule. They generated cash faster than they could spend it, bought back their own stock, and pushed into new markets without ever asking shareholders to fund the expansion out of current earnings.
Investors got comfortable. A generation of retirement accounts was quietly built on the assumption that these companies could grow enormously without bleeding.
That assumption is being stress tested in public now, one quarter at a time, by a handful of firms that have decided artificial intelligence is worth almost any price. The outlays are staggering. The payback remains theoretical. And the market has spent most of 2026 trying to work out how much patience it actually has left.
Wednesday, July 22, in the afternoon, brought the biggest test of that patience yet. Alphabet (GOOGL) posted second-quarter results that beat Wall Street on revenue, delivered the fastest cloud growth in company history, and then watched its own shares slide anyway.
Why Google Cloud carried the whole quarter
The setup mattered. Going into the print, the bear case on Alphabet had almost nothing to do with advertising and almost everything to do with arithmetic.
More AI Stocks:
- GE Vernova’s Al power trade has one weak link
- Jim Cramer reveals 4 surging chip stocks he likes best
- Zuckerberg’s stock rebound made him world’s 5th richest at $222B
Search held up. Google Search and other advertising revenue climbed 17% to $63.3 billion, and YouTube advertising added $11.1 billion, up 13%, according to Alphabet’s earnings release. The feared collapse in query volume from chatbots has not shown up in the numbers, a point our coverage flagged ahead of the print in Google stock price faces major AI test ahead of earnings.
Cloud was the swing factor, and it swung hard. Google Cloud revenue jumped 82% to $24.8 billion, with segment operating income reaching $8.8 billion against $2.8 billion a year earlier.
Analysts had been watching the margin line closely, since cloud margin estimates for the back half of the year had already come down roughly 300 basis points from the prior quarter, according to S&P Global.
Total revenue landed at $119.80 billion against the $116.93 billion analysts expected, reported CNBC. Operating income rose 30% to $40.8 billion.
By any ordinary standard, that is a strong quarter. This was not an ordinary quarter.
What the capital spending guidance changed for Alphabet investors
The stock held steady through the numbers and broke when chief financial officer Anat Ashkenazi got to guidance.
Alphabet lifted its 2026 capital expenditure range to between $195 billion and $205 billion, up from $180 billion to $190 billion three months ago, and repeated that 2027 spending would increase significantly from there.
Shares dropped nearly 5% in extended trading before recovering part of the loss, reportedStocktwits.
I ran the capital spending figure against the revenue line before writing a word of this, and the ratio is what makes the reaction rational rather than emotional. Alphabet spent $44.9 billion on capital projects in a quarter that produced $119.8 billion in revenue. That is roughly 37 cents of every dollar coming in the door going straight back out into data centers, chips, and power.
Free cash flow went negative as a result, landing at negative $5.9 billion, reported Seeking Alpha. Adjusted earnings came in at $2.85 a share against the $2.89 analysts modeled.
Here is the quarter in five numbers:
- Google Cloud revenue rose 82% to $24.8 billion, according to Alphabet’s earnings release.
- Cloud backlog reached $514 billion after growing more than $50 billion in three months, reported Seeking Alpha.
- Capital expenditures hit $44.9 billion for the quarter, roughly double the year ago figure, according to Alphabet’s earnings release.
- Analysts polled by Visible Alpha had modeled about $188 billion of 2026 capital spending, reported CNBC.
- Alphabet is now guiding to as much as $205 billion, according to Alphabet’s earnings release.
Management is not apologizing for it. “We’re still in a supply-constrained environment,” Ashkenazi told analysts, according to CNBC.
How a $98 billion gain reshaped the profit headline
Then there is the number that made every headline and tells you almost nothing about how Google is running.
Net income came in at $112.1 billion, up 298%, with diluted earnings of $9.11 a share. Almost none of that came from selling ads or renting servers.
Other income reflected “a net gain of $98 billion,” driven primarily by unrealized gains on equity holdings, Alphabet said in its earnings release. The equity gain alone added $6.26 to diluted earnings per share and $77.1 billion to net income.
Those gains trace primarily to Alphabet’s stakes in privately held AI developer Anthropic and in SpaceX, which went public in June, reported Fortune. Bank of America had flagged the accounting quirk days earlier, as our preview of the analyst note laid out.
What struck me when I read the release is how cleanly the two stories separate. Strip out the paper gains and you get a company growing operating income 30% while spending like it is at war. That is the business. The rest is a mark-to-market on somebody else’s valuation, and marks move both ways.
What Alphabet investors should watch over the next two quarters
The useful question is not whether Alphabet had a good quarter. It did. The question is what the reaction tells you about how the market now prices this entire category.
For most of the past three years, rising capital spending was read as confidence, and stocks went up on it. That reflex is gone. The same guidance that would have added market value in 2024 subtracted it on Wednesday, and Alphabet is the most profitable company on the tape.
If you hold an S&P 500index fund in a 401(k), you own this argument whether you follow it or not. The four largest AI spenders carry enough index weight that their return on invested capital is functionally a decision about your retirement math, made by finance chiefs you will never meet.
Three things worth tracking into the fall. Whether cloud operating margin holds near current levels once the newest capacity comes online. Whether free cash flow returns to positive territory by the fourth quarter. And whether the 2027 capital number, when Alphabet finally puts a range on it, lands closer to analyst models or blows through them the way this one did.
The efficiency of that spending and the timeline on returns remain the variables that separate the bull and bear cases.
The backlog says the demand is real. The cash flow statement says the bill arrives first.
Related: Intel and Google deepen AI ties for chip design
