Alphabet Fell on Its Best-Ever Quarter. The Moat Is Fine; the Capex Bill Is the Question.

Alphabet just reported what was, by any ordinary reading, its best quarter ever. Revenue rose 24% to $119.8 billion. Google Cloud, the division that decides whether the AI build-out pays off, grew 82% to $24.8 billion. Its own quarter. The stock fell anyway — and it has kept falling, roughly a fifth off its 52-week high, to the point where options dealers now call the calls cheap and bargain hunters are drifting in.

I keep coming back to one question: what more do investors want from a business growing faster than nearly every megacap on the board that they are pricing like the rest of the market? The answer, when you dig into the quarter the way the market did, is not that Alphabet's numbers were bad. It is that the number that mattered this time was not on the income statement at all.

The number that moved the stock wasn't revenue

Alphabet's single biggest, most durable advantage in this entire cycle is the balance sheet. It is a business that for two decades minted cash and shipped most of it back to shareholders. So when the company posted a negativefree cash flow of $5.9 billion in the June quarter — the first time in its history it has burned cash in a quarter — the market did not need to see weak sales to sell. It saw a changed identity.

The cause is the AI build-out. Alphabet spent $44.9 billion on capital expenditures in the quarter and now guides to $195 billion to $205 billion for 2026, roughly double last year's $91.4 billion and about six times what it spent in 2022. Operating cash flow of $39.1 billion could not cover the quarter's capex. Layered on top of that are the narrative scars: the flagship Gemini 3.5 Pro model slipped past its June date, two senior AI researchers left, and in early August the stock fell on a reported model delay. Each story fed the same worry — that Alphabet is losing the frontier-AI race, spending enormous sums, and earning nothing extra for it.

Pile those together and you get the bear case in one line: throwing $200 billion a year at a model race it is losing, while Meta's open-weight Llama and other cheap frontier models suggest the model layer itself is becoming a commodity. That is the story that has dragged the multiple down.

Where that fear should show up — and doesn't

This is where I have to be the annoying contrarian and check the story against the revenue-mix data, because the narrative and the numbers are pointing in different directions.

If AI were really strangling Google's core business — if ChatGPT, Meta AI and OpenAI were quietly replacing the search box — the surest place it would appear is the search advertising line. That line grew 17% year over year to $63.3 billion, a full year into the AI Overviews era. Total advertising rose 14.5% to $81.6 billion. The point is not that competition doesn't exist; it plainly does, and it can get worse. The point is that "competition exists" and "the moat is breached" are different claims, and the revenue data supports only the first.

The same logic cuts the other way on the single biggest counter-signal in the quarter. Yes, capex crushed free cash flow for a quarter — but free cash flow is a function of spending in this story, not of demand. Cloud revenue grew 82% and the company keeps raising capex because, in its words, demand keeps outrunning capacity. That is the classic sign of monetization arriving faster than expected, not slower. A $200 billion annual spend is only a problem if the revenue attached to it does not show up; right now it is showing up in a way the company calls its best cloud quarter ever.

Priced for the fear, not the forecast

So where does that leave valuation? This is the part the selloff has quietly transformed. Alphabet now trades at a trailing price-to-earnings ratio in the mid-to-high teens — roughly 17x — against roughly 24x for Meta, 27x for Microsoft, and 27x for Nvidia, and its forward multiple has compressed toward the broad market's while this business still grows revenue more than twice as fast as the average large company. When a high-growth company's forward multiple converges with the S&P 500, the burden of proof shifts onto the bears. They now have to defend a scenario — AI hollows out search and the cloud build-out fails to pay — that the last three quarters of revenue data have not started to support.

I want to be honest about what this case is not, though. This is not the clean "dirt cheap, market is simply wrong" setup the options chatter implies, and the cheap options themselves tell you why: low premiums mean the debate has gone quiet, and a quiet debate is a positioning signal, not a thesis. The capex and negative-free-cash-flow issue is a thesis-based concern about returns on invested capital, not noise. If AI demand disappoints, Alphabet is stuck with fixed obligations — its disclosed purchase commitments ballooned from roughly $150 billion at the end of 2025 to $811 billion by mid-2026 — that dilute returns and shrink buybacks for years.

That is exactly why the value is real but contingent, and why I would not tell a holder to chase the bounce or a watcher to back up the truck. The setup has arguably overcorrected on the valuation side — the moat is intact on the data and the multiple is near the market's on a business growing at multiples of it — but the resolution lives in one place: whether Google Cloud continues to monetize the build-out faster than the bill grows. The invalidation is concrete, not abstract. If free cash flow stays negative into next year, or cloud growth rolls over while capex keeps climbing, the bear case stops being priced-maybe and becomes priced-right. Until then, the honest read is that the market is paying a market price for a company that no longer deserves one.