Sundar Pichai had a number ready on Alphabet’s earnings call: $514 billion. That, the chief executive said, is the backlog of cloud orders the company is sitting on, driven by demand for the artificial-intelligence infrastructure Google now sells to some of the largest companies in the world. Alphabet reported second-quarter revenue up 24 percent from a year earlier, beating analyst estimates, and the stock moved higher.
The quarter gave investors a look at how the AI buildout is changing Google’s economics. Google Cloud revenue rose 82 percent, the fastest growth the unit has recorded, and Pichai said nearly 90 percent of Fortune 100 companies have deployed Gemini Enterprise, the company’s AI productivity suite. The backlog figure — orders signed but not yet delivered — is the metric executives pointed to as evidence that the demand is real and durable, not a bubble in the making.
Alphabet also raised its 2026 capital-spending target to $205 billion, up from the $185 billion it guided three months ago. The increase follows a similar move by Microsoft, which has said it will spend more than $200 billion this year, and by Meta, which raised its own AI budget. The three companies together are committing roughly $600 billion to data centers, chips and networking this year, a sum that has become the central bet of the technology industry.
The spending is producing results, at least in the cloud business. Google Cloud’s operating margin has widened as AI workloads fill capacity, and executives said the unit’s backlog grew faster than revenue in the quarter, meaning the pipeline for future quarters is expanding. The company said it expects the buildout to keep paying off as customers move from testing AI models to running them in production.
The question hanging over the call was sustainability. Analysts pressed executives on how long customers will keep paying for AI capacity, and on whether the industry is overbuilding ahead of demand. Pichai’s answer was that orders are backed by signed contracts and that enterprises are still early in adopting AI tools. He conceded that some customers are experimenting, but said the conversion of trials into paid contracts has been faster than expected.
The earnings also arrived against a regulatory backdrop. The European Commission fined Google €890 million earlier in the week, the first penalty under the Digital Markets Act. Executives did not discuss the fine in detail on the call, but the timing underscored the tension Alphabet now manages: record spending on AI in one hand, a growing stack of regulatory rulings in the other.
For the market, the quarter sharpened the divide between the companies building AI infrastructure and everyone else. Alphabet, Microsoft and Meta are spending at levels that would once have seemed reckless; their shares have held up because revenue is growing to meet the spending. The risk, analysts said, is a quarter where growth slows before the investment pays off. That is the scenario the next round of earnings will be tested against.
Pichai closed the call with a prediction: AI infrastructure will be the defining investment of the decade, and Alphabet intends to be the largest supplier of it. The $514 billion backlog is the evidence he offered. Whether that backlog converts into profit at the pace investors expect is the question the market will keep asking every quarter until it does.
The rest of Alphabet’s business moved in familiar patterns. Search advertising grew at a single-digit pace, slower than the double-digit rates the company once delivered, as AI features change how users find answers. YouTube ad revenue grew at a solid clip, and the company’s “Other Bets,” the collection of experimental businesses that includes Waymo, narrowed its losses. The cloud unit, however, was the story, and its growth lifted Alphabet’s overall operating margin even as spending on AI infrastructure climbed.
The $205 billion capital-spending figure deserves context, executives said. A growing share of that budget now goes to building out the company’s own custom chips, the TPU line, which Alphabet designs in-house and uses alongside Nvidia hardware. Executives argued that owning the silicon gives Alphabet better economics per unit of computing than rivals that rent everything from outside suppliers, and that the backlog of cloud orders justifies the pace.
Wall Street’s response was measured. The stock rose on the results, but several analysts noted that Alphabet’s valuation now assumes the AI spending produces durable returns for years. The company trades at a premium to its historical average, and the premium rests on the same question investors ask of Microsoft and Meta: can the AI buildout keep compounding, or does it eventually hit a wall of diminishing returns? Pichai’s $514 billion backlog is the strongest answer any of the three companies has offered so far, analysts said.
The call also carried a note of caution about the pace of change. Pichai said the company is reorganizing teams around AI products faster than some employees would like, and that some traditional roles are being redefined. The message to investors was consistent: Alphabet is spending, hiring and reorganizing for one outcome, and it intends to be the company that supplies the AI era’s computing foundation.
