NEW YORK — Alphabet shares dropped roughly 5% in after-hours trading Wednesday after the company lifted its 2026 capital expenditure guidance to a record $205 billion during its second-quarter earnings call, unsettling investors who had just watched the parent of Google post what was otherwise a strong beat. Revenue reached $119.8 billion for the quarter, a 24% jump year on year, and Google Cloud expanded at 82% annually, the fastest rate the division has reported in at least three years.
The reaction was swift and negative. Alphabet’s stock had been holding steady through the headline numbers, then fell sharply when chief financial officer Anat Ashkenazi confirmed the new full-year capex target. The prior guidance had been roughly $75 billion for the year. Raising it to $205 billion in a single call, with the second half of 2026 only beginning, amounted to telling analysts that the next two quarters would require capital commitments that dwarfed what Alphabet had historically spent on infrastructure across multiple years combined.
The company’s explanation rested on the assumption that AI-driven cloud demand will justify the investment. Google has been releasing successive generations of Gemini models at an accelerating pace, and the theory is that every enterprise customer who adopts Gemini ultimately needs Google Cloud infrastructure to run it. That theory has now produced a 82% revenue surge in the cloud division, the fastest organic growth since Alphabet began breaking out the segment separately. The question analysts were asking on the call was whether the 82% rate can last long enough to absorb a capex program that now rivals the GDP of countries.
“We are seeing strong demand signals across all verticals,” said Sundar Pichai, Alphabet’s chief executive, on the call. He cited healthcare, financial services, and government as the enterprise sectors where Gemini adoption is accelerating fastest, and said the company expects the AI infrastructure build-out to remain a “sustained multi-year investment cycle.” What that means in practice is that $205 billion in 2026 is not a ceiling.
YouTube, often treated as an afterthought in Alphabet’s earnings narrative, posted $11 billion in advertising revenue for the quarter, a 13% increase over the same period last year. The growth rate is healthy by media standards but modest by Alphabet’s own historical benchmarks, and it reflects a structural shift in how YouTube makes money: direct-response advertising, the category that connects an ad to an immediate purchase, is now the largest revenue contributor, with brand advertising and connected-television deals growing behind it. That diversification matters because it makes YouTube’s revenue base less vulnerable to any single advertising category cooling.
The tension in the results is not between a weak and a strong business. Nearly every line in Alphabet’s Q2 report was ahead of analyst estimates. Search revenue grew, Cloud grew faster, YouTube grew steadily, and total operating income expanded. The tension is between scale and belief. The broader AI infrastructure build-out has already produced record backlogs across server and networking suppliers, signaling that demand is real and enterprise customers are committing capital. But the returns on that demand have been slower to appear in the income statements of companies doing the most AI spending.
Alphabet is currently spending more on capital investment in a single year than it earns in net income over roughly the same period, which means the company is self-funding an AI infrastructure arms race with cash generated by search and advertising businesses that are not themselves growing at AI speeds. That is a stable arrangement if the cloud and AI lines eventually grow large enough to be self-sustaining. It becomes unstable if the cloud growth rate normalizes before the capex wave does.
The closest parallel in Alphabet’s own history is the period from 2017 to 2020 when the company began investing heavily in data centers for what would eventually become Google Cloud. Analysts at the time consistently questioned whether the investment would pay off against the dominance of Amazon Web Services and Microsoft Azure. It eventually did, though it took years. The current AI capex cycle is larger, faster, and more concentrated in a shorter time window, which leaves less room for a slow-burn outcome.
Wall Street has already spent parts of 2026 questioning whether the AI spending cycle is outpacing the revenue it generates, and Alphabet’s call crystallized the concern. Three of the four largest US technology companies have now reported second-quarter results, and the pattern is consistent: strong beats on revenue and earnings followed by after-hours drops on capital expenditure figures that were larger than expected. The market is not yet convinced that the AI winners can be identified from among the AI spenders.
What Wednesday’s results do not reveal is whether the $205 billion will look prescient or extravagant three years from now. Alphabet’s cloud growth rate of 82% is real, and the enterprise AI pipeline Pichai described is visible in customer contracts. But the returns on infrastructure at this scale take time to materialize, and the analyst call ended with a question no one could answer cleanly: at what point does building the AI future start paying for itself, and how much is Alphabet prepared to spend while it waits to find out.
