Business

Know the Business

Apple is a premium consumer-electronics company whose economic gravity has quietly shifted: the iPhone still sets the rhythm, but Services — at 26% of revenue and 42% of gross profit — is now the real compounding engine. The market gets the installed-base flywheel right; what it under-appreciates is how little incremental capital Apple deploys to grow it, and what it over-appreciates is that a 34x P/E can survive a hardware-flat year, a shrinking China, and an AI race Apple is visibly not leading.

1. How This Business Actually Works

The simplest way to see Apple: it sells a hardware subscription that renews every 3–4 years, and monetizes the gap between renewals through a software tax on its own installed base.

The iPhone is half of revenue and pulls every other segment along: it is the entry ticket to the ecosystem, sets the ASP ceiling, and gates App Store, advertising, cloud, and payment revenue that follows. Products — iPhone, Mac, iPad, Wearables — carry a reported 36.8% gross margin. Services carry 75.4%. That is the whole story of operating leverage at this company: every incremental dollar of revenue that comes from a developer taking a 70/30 App Store split, an iCloud subscriber paying $0.99/month, or a third-party ad placed against an App Store search is a dollar that arrives with roughly twice the contribution margin of selling another phone.

The scarce asset is the installed base, not the chip. Apple does not need to discover new customers — it needs existing iPhone owners to stay on iOS, upgrade to Pro tiers when ASPs allow, and spend more on Services each year. That explains the capital allocation: $89.3B repurchased and $15.4B paid in dividends in FY2025 against $12.7B of capex. Apple's "factory" — silicon design, OS, App Store, payments rails — is already built, and the marginal cost of the next App Store transaction is approximately zero. The bottleneck is not capital; it is ecosystem trust (Epic, DOJ, EU DMA) and replacement-cycle demand.

2. The Playing Field

Apple sits in the Big Five hyperscaler peer set, not in a consumer-electronics peer set — its economics long ago left Samsung and Lenovo behind. Within that club, Apple is the lowest-R&D, highest-buyback, most-hardware-exposed name, and the AI race is re-ranking the group.

The peer table exposes Apple's unusual shape. Its 32% operating margin is mid-pack — Microsoft and Meta both convert revenue to profit more efficiently — and its gross margin of 47% is structurally capped by the hardware mix. The flattering numbers are that Apple runs on 8.3% R&D intensity versus 11–29% for its pure-software peers; Apple effectively outsources foundation-model research to TSMC, ARM, and OpenAI, and historically that lean cost structure has been the right answer. In the current AI cycle it has become a live question: Meta is spending 28.5% of revenue on R&D, and Alphabet $61B a year, while Apple's Siri and Apple Intelligence have visibly trailed. The most honest read of the P/E spread — Apple at 34.3x versus Microsoft at 26.5x — is that Apple is priced on the durability of the installed base, not on incremental growth the way its peers are.

What "good" looks like in this peer set is clear: Microsoft's mix of recurring software and cloud with 46% operating margins; Meta's ability to force-multiply R&D through a single ad-targeting model. What Apple does better than anyone is capital returns and gross-profit-per-dollar-of-hardware — no one else gets 75% margins on software running on somebody else's store.

3. Is This Business Cyclical?

Moderately. Apple is not commodity-cyclical, but the iPhone replacement cycle, China consumer demand, and now tariffs create visible troughs every 3–5 years. The pattern: one flat/down year every few years, followed by a product-cycle rebound. Services smooths but does not eliminate it.

Three cycle facts that matter. First, FY2016 (−8%) and FY2023 (−3%) are the template: iPhone cycle fatigue plus China weakness. FY2019 (−2%) had the same fingerprint. Second, Greater China has now declined three fiscal years in a row — from $73B (FY2023) to $64B (FY2025), or roughly 150 bps of company revenue per year — and this has been a structural, not cyclical, move (domestic champions, government procurement preferences, consumer nationalism). Third, the new exposure is tariffs: the FY2025 MD&A flags tariff costs as a fresh drag on Products gross margin, with a Section 232 semiconductor investigation specifically threatening the company's ability to price through. Unlike past cycles, this one is not self-correcting on a 12-month view.

Working capital holds up through downturns — inventory is tiny (1.4% of revenue) and Apple actually collects from customers before paying suppliers, so free cash flow stays positive even in −8% years. The pressure point in a downturn is not liquidity; it is gross margin, because channel discounting and component mix shifts compress Products margins, and ASP concessions in China reverberate across the services attach rate.

4. The Metrics That Actually Matter

Most headline metrics on Apple (P/E, ROE, EPS growth) are second-order. The few that actually drive the valuation:

Two metrics that look central but are misleading. ROE (reported 152%) is an arithmetic artifact of a near-zero equity base — Apple has bought back so much stock that the denominator is $74B against $112B of net income. ROA of 24% is a cleaner single-year read; the real economic return on the business is closer to ROIC in the 50–60% range, which is world-class but not the three-digit number the screen shows. EPS growth is likewise distorted upward by the buyback — revenue grew 6% but EPS grew roughly 12% in FY2025; the cleaner picture is to look at net income growth and remember buybacks are doing half the work.

5. What I'd Tell a Young Analyst

Four things.

1. Follow Services, not iPhone. The iPhone number is useful for headlines and for triangulating the replacement cycle, but the marginal dollar of value in this company is being created in Services — the App Store, advertising, AppleCare, and the cloud subscriptions that ride on top of the installed base. If Services growth slips below 10%, the thesis is broken in a way no iPhone cycle can fix.

2. Watch China as a structural story, not a quarter-to-quarter one. Three consecutive years of decline in Greater China is not a cycle — it is share loss to domestic brands and a political ceiling on how much Apple can sell there. Model China as a declining $60B line, not a $70B line coming back. Tariffs make this worse, not better.

3. The market is underpricing the tariff pass-through question. Apple's FY25 MD&A explicitly says tariffs are already hitting Products gross margin, and the Section 232 semiconductor review is unresolved. This is a real, mathematical $5–15B gross-profit risk depending on outcome, and consensus is still treating it as a footnote.

4. The market is over-pricing the AI race as a known negative. Apple is behind on foundation models. That is correct. What is also correct is that the installed base is 2+ billion active devices, and Apple controls the hardware distribution channel into which any serious on-device AI product has to ship. The thesis that breaks Apple is not "Apple Intelligence is mediocre" — it is "developers and consumers care enough to switch platforms." Historically, they do not. Watch switching metrics, not model benchmarks.

What would actually change my view: (a) Services growth decelerating below 10% for two consecutive quarters, (b) a DOJ or EU ruling that forces App Store take rates below 20%, (c) China down another 5%+ year, or (d) Apple materially slowing buybacks — because that would signal management sees reinvestment opportunities the rest of us don't. Short of those, this is a compounding hardware-software annuity trading at a premium the market has been willing to pay for 15 years and will probably keep paying unless one of those four things breaks.