Alphabet expands revenue as AI spending hits cash flow

Alphabet’s reported second-quarter figures, if confirmed in an official filing, would show a company converting exceptional revenue and cloud growth into an unprecedented infrastructure spending cycle that has begun to consume more cash than its operations generate. The more consequential issue is not the reported 24% revenue increase or the claimed expansion of Google Cloud, but whether Alphabet can sustain capital expenditure approaching $200 billion a year without materially weakening its balance-sheet flexibility or disappointing markets expecting AI spending to produce durable returns.

The figures supplied in the report have not been accompanied by an Alphabet earnings release, Securities and Exchange Commission filing, earnings-call transcript, or other primary document that would allow independent verification. That limitation is material. The account includes several unusually large financial claims, including quarterly capital expenditure of $44.9 billion, negative free cash flow, nearly $70 billion in financing, and a $514 billion cloud backlog. Until Alphabet publishes source documents, these figures should be treated as unverified claims rather than confirmed results.

The same caution applies to the article’s conclusions about cryptocurrency markets. Its assertion that large technology-company borrowing will drain capital from digital assets, while boosting decentralized computing tokens, is not supported by evidence presented in the material. Corporate bond issuance, AI infrastructure spending, token valuations, and cryptocurrency flows can be related through financial conditions, but the supplied article does not establish a measurable causal connection among them.

Reported earnings point to a widening gap between profit and cash generation

According to the supplied account, Alphabet generated $119.8 billion of second-quarter revenue, a 24% increase from a year earlier, while operating profit rose 30% to $40.8 billion. Those growth rates would indicate that Alphabet’s advertising, cloud, and infrastructure businesses continued to expand at a pace well above that of many mature global companies.

The report also states that Alphabet shares fell nearly 3% in after-hours trading. The claimed market reaction reflects a familiar concern surrounding the largest AI infrastructure builders: accounting profits can remain strong while cash generation deteriorates because data centers, servers, specialized chips, networking equipment, and energy capacity require major upfront commitments.

The supplied figures suggest that this gap became pronounced during the quarter. Operating cash flow reportedly reached $39.1 billion, but capital expenditures totaled $44.9 billion, producing negative free cash flow of $5.855 billion. Free cash flow is generally calculated as operating cash flow less capital expenditures, and it is a key measure of how much internally generated cash remains after a company funds its physical and digital infrastructure.

If those numbers are accurate, Alphabet would have crossed an important financial threshold. It would not mean the company lacks profitability or faces an immediate liquidity problem. A company with Alphabet’s scale can finance large investments through existing cash, operating income, debt, or other securities. It would mean, though, that the company’s AI expansion is demanding more capital in the near term than its operations are providing after infrastructure investment.

That distinction matters because markets often value a company’s future returns on invested capital rather than its current revenue growth alone. The central question is whether the assets Alphabet is building will generate enough incremental cloud revenue, AI-related advertising demand, enterprise software sales, or productivity gains to justify their cost over time.

Anthropic stake reportedly complicates the earnings comparison

The supplied account says Alphabet’s 14% equity stake in Anthropic lifted reported earnings per share to $9.11, while adjusted earnings per share excluding that gain were $2.85. It further says the adjusted result fell below a market estimate of $2.95.

No methodology is provided for the alleged Anthropic-related gain, the valuation used, the accounting treatment, or the source of the market expectation. Those details are necessary to determine whether the reported $9.11 earnings-per-share figure reflects a recurring operating improvement, a mark-to-market adjustment, a realized gain, or another non-operating item.

The distinction is not cosmetic. Alphabet’s core operating performance would be better assessed through revenue, operating income, operating margins, cash flow, and segment results than through a headline per-share figure potentially affected by changes in the value of a private-company investment. Private-company holdings can produce substantial accounting gains or losses without corresponding cash receipts in the reporting period.

The claimed adjusted earnings miss, if accurate, would also help explain why traders might react negatively despite reported revenue and operating-income growth. A market response to an earnings release can reflect several competing expectations at once: future spending, margins, cloud demand, financing needs, and whether adjusted profitability matches consensus forecasts. The supplied material does not identify the analysts behind the $2.95 estimate or provide a range of estimates, so the size and significance of the purported shortfall cannot be independently evaluated.

Google Cloud is presented as the growth engine

Google Cloud was reportedly Alphabet’s fastest-growing business, with revenue of $24.77 billion, up 82% from the prior year. The report attributes the growth to AI infrastructure demand and enterprise-scale AI products, while stating that remaining performance obligations reached $514 billion.

Remaining performance obligations typically represent contracted revenue that has not yet been recognized. They can provide a useful indication of future demand, particularly in cloud computing and enterprise software, but they are not equivalent to immediately realizable revenue. Their quality depends on contract duration, cancellation terms, delivery obligations, customer concentration, and the portion tied to multi-year infrastructure agreements.

The supplied account says more than half of the claimed $514 billion in remaining performance obligations is expected to be recognized within two years. That would imply a very large near-term revenue conversion opportunity, but Alphabet would need to disclose the composition of those obligations for analysts to judge their reliability. A backlog can include commitments for services that require substantial additional infrastructure spending before revenue and cash are recognized.

The reported scale of cloud growth also needs context that the material does not supply. It does not provide Google Cloud operating profit, operating margin, customer retention data, workload mix, capacity utilization, or the amount of revenue associated specifically with AI services. Without those metrics, it is not possible to determine whether AI-related cloud growth is expanding profitability at the same rate as revenue or whether it is being purchased through lower-margin infrastructure commitments.

Alphabet reportedly said Google Services generated $94.54 billion in revenue, up 15% year over year. Search and other services allegedly produced $63.27 billion, a 17% increase, while YouTube advertising revenue rose 13% to $11.06 billion. If confirmed, those numbers would show that Alphabet’s established advertising businesses remain the primary cash-generating base supporting its AI investment program.

That funding relationship is strategically important. Google Cloud may be the fastest-growing operation, but search and advertising would remain the larger source of revenue and likely the main internal support for capital-intensive infrastructure expansion. Any assessment of Alphabet’s ability to finance AI spending should therefore examine whether advertising growth remains resilient as spending commitments rise.

Capital expenditure guidance would reset the scale of the AI buildout

The report says Alphabet raised full-year capital expenditure guidance to between $195 billion and $205 billion after spending $44.9 billion in the second quarter. Such a target would place annual investment at a level that requires close scrutiny of asset life, depreciation, capacity use, power procurement, and expected revenue returns.

Data centers and AI hardware are not interchangeable long-lived assets. Buildings, networking systems, power equipment, conventional servers, and high-performance accelerators have different useful lives and replacement cycles. A company can report strong earnings while the economic life of its computing equipment shortens, because depreciation policies and capital expenditure timing do not always move in lockstep with changes in technology.

The report attributes the spending increase to data centers, servers, and networking equipment. It does not specify how much is devoted to land and construction, how much goes to chips and servers, whether the company has long-term purchase obligations beyond capital expenditures, or what portion is intended for internal products versus external cloud customers. Those omissions limit any effort to calculate prospective returns on the investment program.

The account also says Alphabet raised about $49.6 billion through convertible preferred stock and another $20.3 billion through unsecured bonds. No official transaction documents, pricing details, maturities, coupon rates, conversion terms, or purchaser information are provided. The phrase “convertible preferred stock” is itself unusual enough in the context of Alphabet’s public capital structure that it requires direct confirmation.

Debt financing would not automatically signal financial distress. Companies may issue debt to preserve liquidity, manage taxes, refinance existing obligations, or fund long-duration assets without repatriating cash. Yet the reported combination of negative free cash flow and nearly $70 billion in new financing would represent a meaningful departure from the image of Alphabet as a company that can fund growth primarily from internally generated cash.

Anat Ashkenazi, Alphabet’s chief financial officer, is identified in the supplied report as saying cash flow would remain under pressure because of AI infrastructure spending. The article does not provide a transcript, exact quotation, conference-call recording, or filing in which that statement appears. It should therefore not be treated as a verified public comment until the underlying record is available.

Gemini adoption figures require clearer definitions

The report says Gemini’s app reached 950 million monthly active users and that nearly 90% of Fortune 100 companies adopted Gemini Enterprise. It also claims Gemini processes about 22 billion tokens per minute.

These figures may indicate broad use, but the terminology needs definition before it can be treated as evidence of commercial success. Monthly active users can include occasional users, users accessing features embedded in other Google products, or individuals who use free services. The figure does not show paid conversion, revenue per user, retention, or inference costs.

Similarly, enterprise adoption does not necessarily mean an organization has deployed Gemini across its workforce or committed substantial spending. A Fortune 100 company may be running a limited pilot, using a small number of licenses, integrating an API into a single workflow, or purchasing a broader cloud contract that includes AI tools. Alphabet would need to disclose the meaning of “adoption” and the associated contract value to make the claim analytically useful.

Token-processing volume measures system activity, not necessarily economic value. Tokens are units of text processed by an AI model. High volume could reflect customer demand, internal testing, product features, automated workloads, or repeated model interactions. Without revenue, cost, utilization, and customer data, the reported 22 billion-token figure cannot establish that Gemini is producing attractive margins.

The supplied report also refers to a postponed Gemini 3.5 Pro release but does not explain the reason for the delay, the original release date, or whether the delay affected customers. Product timing is relevant because AI infrastructure spending is being justified partly by the need to compete with other developers. Delays could matter if they reduce a company’s ability to monetize capacity already being built, though no evidence in the material shows that such an effect has occurred.

Crypto conclusions are not supported by the evidence presented

The final section of the supplied material shifts from Alphabet’s finances to claims about digital-asset allocation and decentralized computing networks. It argues that technology companies’ AI-related bond issuance will pull capital from “riskier digital assets,” while creating an opening for blockchain-based networks that aggregate idle computing hardware.

That argument is presented as a recommendation rather than reporting, and it lacks the evidence needed to support either conclusion. Corporate bond issuance can affect financial conditions if it contributes to changes in yields, credit spreads, or liquidity, but a single sector’s borrowing does not mechanically determine demand for Bitcoin, other cryptocurrencies, or computing-related tokens. Digital-asset pricing is influenced by many variables, including leverage, derivatives positioning, macroeconomic conditions, regulation, stablecoin liquidity, token emissions, protocol revenue, and market structure.

The supplied article claims that major technology companies issued roughly $244 billion in corporate bonds by July 2026 for AI hardware and could issue $570 billion by year-end. It does not identify the companies included, the bond database used, which proceeds were specifically allocated to AI infrastructure, or the methodology behind the year-end projection. These numbers cannot be assessed without an identifiable primary source or recognized independent fixed-income data provider.

Its claim that the decentralized computing sector has surpassed a $50 billion market capitalization has the same problem. No index provider, token universe, valuation date, or methodology is identified. Market capitalization can also be a weak measure of a decentralized physical infrastructure network’s competitive position because it does not establish actual hardware availability, utilization rates, customer revenue, service reliability, geographic coverage, energy costs, or the ability to meet enterprise security requirements.

Decentralized networks may offer lower-cost computing in certain workloads, particularly where jobs are flexible, geographically distributed, or less dependent on strict uptime guarantees. That does not demonstrate that they can replace hyperscale data centers for enterprise AI training or inference. Major cloud customers often require contractual service levels, data-governance controls, predictable capacity, technical support, compliance certifications, and integrated software tools that tokenized hardware marketplaces may not yet provide.

The measurable issue to monitor is not a generalized expectation that AI borrowing will redirect money into computing tokens. It is whether Alphabet’s eventual primary disclosures show that cloud and AI revenue, operating margins, and contracted demand are rising fast enough to cover an annual capital expenditure program reportedly approaching $200 billion while free cash flow remains under pressure.