Alphabet's second-quarter numbers settle one argument about artificial intelligence but intensify another. AI features have not yet broken Google Search's commercial engine: Search and other revenue grew 17% year over year. Google Cloud also accelerated sharply. The harder question is now financial rather than existential — how much cash must Alphabet commit before that demand produces an acceptable return?
That distinction matters because the quarter contains two very different stories. The operating business expanded, while the income statement also recorded an enormous investment gain that does not represent advertising or cloud cash generation. At the same time, capital expenditure reached a level that pushed quarterly free cash flow below zero. Investors therefore need to read past both the record profit headline and the initial excitement around AI demand.
Search strength changes the AI question
Alphabet reported second-quarter revenue of $119.8 billion, up 24%, and operating income of $40.8 billion, up 30%, according to its official results. Google Services remained the largest engine. Within it, Search and other revenue rose 17% to about $63.3 billion, while YouTube advertising grew 13%.
Those figures do not prove that generative AI can never weaken traditional search economics. They do show that displacement is not visible in the current revenue line. Management says AI Overviews and AI Mode are increasing query activity, and CEO Sundar Pichai said AI Mode had passed one billion monthly active users. In his published earnings remarks, he also said the company is sending billions of clicks to websites each week through AI-powered Search features.
The mechanism is important. If AI interfaces lead users to ask more or more complex questions, Google can protect search volume even as the format changes. But revenue growth alone does not reveal whether an AI-assisted query carries the same margin as a conventional one. Generating answers requires more computing infrastructure, and some responses may leave less room for paid links. The quarter therefore supports the demand side of the case, not yet the full unit-economics case.
The headline profit is not the operating scorecard
Alphabet reported net income available to common shareholders of $112.1 billion, almost four times the prior-year level. Treating that figure as evidence that AI spending has already paid for itself would be a serious analytical error. The company recorded $98.0 billion of other income, primarily a net unrealized gain on equity securities. The Associated Press reconciliation likewise notes that investment gains — mostly linked to SpaceX after its June listing — drove the exceptional bottom line.
Operating income is a cleaner measure of the businesses' current performance, although even that does not capture the cash needed to build infrastructure. Alphabet spent about $44.9 billion on capital expenditure during the quarter. Reuters reported that this contributed to negative free cash flow of $5.9 billion, the company's first negative quarter on that measure.
One quarter of negative free cash flow is not the same as a broken model. Data-center construction and server deliveries are lumpy, while the assets will be used for years. Yet the scale changes the valuation debate. Management raised full-year 2026 capital-expenditure guidance to $195 billion-$205 billion from $180 billion-$190 billion. The midpoint is more than double Alphabet's 2025 capital expenditure. Depreciation will follow the spending into future income statements even after the initial cash outflow has passed.
Cloud demand is the strongest defense of the build-out
The best evidence supporting that budget is Google Cloud. Revenue rose 82% to $24.8 billion, operating margin expanded to 35.6%, and management put Cloud backlog at $514 billion. Backlog is not the same as recognized revenue — it depends on contract timing, customer usage and delivery — but it offers more concrete support than a broad promise that AI will be transformative.
Management also says demand still exceeds available capacity. If true, added servers and data centers can unlock revenue that Alphabet is currently unable to serve. Cloud is especially useful as a test because customers pay directly for infrastructure and AI services. That creates a more observable link between investment and revenue than consumer AI features funded indirectly through advertising.
There is nevertheless a concentration risk in the logic. Alphabet's infrastructure serves Search, Gemini, YouTube and internal workloads as well as external Cloud customers, so no simple division of capex by Cloud revenue is meaningful. The company is also building before the industry knows the long-run price of AI inference or how quickly hardware becomes obsolete. Faster demand today can coexist with disappointing returns if competition drives prices down or if each dollar of revenue requires repeated hardware refreshes.
The burden of proof moves to utilization
The bullish interpretation is coherent: Search is growing, Cloud is capacity constrained, Cloud margins are expanding, and a $514 billion backlog provides visibility. Under that view, temporarily weak free cash flow is the cost of bringing forward infrastructure that will support several businesses for years.
The skeptical interpretation is equally legitimate. Alphabet's cash-generating advertising franchise is financing an investment program of unprecedented size just as AI changes the cost structure of its core product. A strong quarter cannot establish the lifetime return on data centers that are still being built. Nor can an unrealized investment gain fund recurring operations in the same dependable way as customer cash.
The evidence that would change the analysis is now measurable. Over the next several quarters, investors can compare Cloud revenue and operating income with deployed capacity; track whether Search growth persists without a material erosion in Services margin; and watch whether operating cash flow again covers capital expenditure as delivery timing normalizes. A sustained rise in depreciation without matching operating-profit growth would weaken the case. Continued Cloud acceleration, resilient Search margins and a return to positive free cash flow would strengthen it.
Alphabet has passed the immediate demand test. Its next test is harder: demonstrating that AI infrastructure can become a productive asset base rather than a permanently rising admission price for defending Search.