Alphabet Inc. GOOGL

Key data

The business

Alphabet sells three things. The first is attention: the search box, YouTube and the ads stitched into both, which is still about 79% of revenue and threw off nearly all the profit. The second is Google Cloud, renting computing and AI tools to businesses, now growing fast enough to matter. The third is a scatter of moonshots, Waymo and the rest, small enough to ignore in the numbers. The moat is distribution nobody can buy back: Search is the default on billions of devices, and the same fleet of data centers that answers a query also trains the models it rents out. The thing a customer holds is invisible and constant, the answer that appears the instant they hit enter, and the whole bet turns on whether an AI-written version of that answer still carries an ad.

The numbers

The sequence matters more than any single line. Revenue climbed from $96.4B to $119.8B across the last five quarters, and the growth rate itself is accelerating: Q2 2026 was up 24% on the year, the fastest in the set, lifted by a US market now 51% of sales.

The profit line below is a trap. Net income tripled to $112.2B in Q2, but $98.0B of that was an unrealized mark-up on stakes Alphabet holds in other private companies, paper gains that reverse as easily as they arrive. Operating income, the money the business actually made, rose a steadier 30% to $40.8B. Read the EPS column knowing the last two rows are flattered.

Across five years the shape is steady growth compounding into a giant. Revenue roughly doubled since 2021 and operating income grew about 13% a year. The 1H 2026 net income below carries the same paper-gain distortion, so weigh the revenue, not the bottom line.

The real story of the last eighteen months is capital, not earnings. Operating cash flow climbed 62% over the two years from 2023 to 2025, reaching $164.7B, while free cash flow stayed flat near $73B all three years because capital spending tripled. Management guided this year's spend to a range near $200B, raised twice. Every dollar is a wager that AI demand shows up in the $513.9B of signed cloud contracts not yet delivered, over half of which converts to revenue within two years.

What this memo believes the market underrates is the quality of that backlog: it is contracted, not hoped-for. The single print that settles the argument is Cloud's growth, up 82% in Q2 year-over-year.

Management

For years the reflex was returning cash. That reflex just reversed. Buybacks fell from $62.2B in 2024 to $45.7B in 2025 and to exactly zero in Q1 2026, while long-term debt went from $12B two years ago to roughly $98B by June. Alphabet raised equity capital in June and issued 6.25% mandatory-convertible preferred. The capital-raising turn was already visible in the June quarter, and it has only deepened since with senior notes issued in August. For the first time in memory Alphabet is a net raiser of cash, all of it feeding the AI build. Insider selling is trivial and routine, $3.4M over twelve months with no purchases and plan status not disclosed, nothing to read into at a $4T company.

How it fails or surprises you

Capex outruns the return. Free cash flow is already flat at about $73B while operating cash flow soared, because spending tripled. With this year's capex running near record levels, if cloud revenue converts the $513.9B backlog too slowly, free cash flow turns down and the paused buyback stays paused. The print: full-year 2026 free cash flow against the capex line.

Search monetization erodes. Services is still 79% of revenue and grew 14.5% in Q2, the slower engine funding everything. If AI-written answers, Google's own included, pull queries away from paid links, that base decelerates and the antitrust remedy narrows distribution further. The print: Services growth slipping toward single digits, plus the remedy's terms.

Cloud re-rates the whole company (right tail). Cloud is compounding fast against a half-trillion-dollar backlog. If its margins scale the way AWS once did and Gemini monetizes, the mix tilts from a decelerating ad company toward a growing AI utility, and a 17x multiple is the wrong one. The print: Cloud operating margin and organic revenue growth.

Closing thoughts

This is a mostly-priced name where one specific print still matters. Thirty-five analysts cover it, every segment is disclosed, and the balance sheet is a fortress, $242B in cash and securities against manageable debt, so survival is never the question. The edge is not information, it is patience with a number the market cannot yet see cleanly: how fast Cloud grows, and whether record capex earns its keep. The person on the other side is selling a decelerating search monopoly with a frightening capex bill; you are buying the option that the same spend builds the next franchise. I would judge the right tail modestly fatter than the left, because the cash floor limits how badly the downside cuts while the cloud re-rate is uncapped.

The bet is still that Google keeps owning the box billions of people type their questions into, rents the computers behind it to other companies, and that the AI answers people get straight from Google do not kill the ads. What breaks it is the pair to watch every quarter: Services growth sliding under 10% while free cash flow falls under the capex load. Hold both lines and the compounding is intact; lose both at once and you own an expensive utility, not a franchise.