Business

Know the Business

Apple is not a smartphone company. It is a platform-rents collector wearing hardware skin: 74% of revenue is hardware that exists to acquire and retain a 2.5-billion-device active install base, and the other 26% is the high-margin business — Services at a 76% gross margin. The right debate is no longer "is Apple a great business" but "what is the platform layer worth, what does regulation chip away, and how much of the iPhone install base inflects to AI-driven replacement and Services ARPU expansion in FY2026–28?" Two underweighted points: the China + India inflection (Greater China +38% in Q1 FY2026 after −4% in FY2025) and Services ARPU compounding. One overweighted point: iPhone unit growth — the 23% Q1 FY2026 print was supply-chase off lean channel inventory, not a durable volume base.

1. How This Business Actually Works

Apple monetizes the same user three times: once when they buy the device, every month for as long as they own it, and again when they upgrade. The first transaction is one-shot and gross-margin-capped at ~37% by the bill of materials. The second is recurring — App Store commissions, AppleCare, iCloud, content subscriptions, advertising, payments, and the search-distribution payment from Google (estimated ~$20B/year) — running at a 76% gross margin. The third is replacement, paid for by trade-in credit, carrier promotion, and 36-month financing that Apple does not carry on its own balance sheet.

The cost structure is fixed-cost-light at the top of the value chain because manufacturing is outsourced. Apple does not own the fabs or assembly lines — Foxconn, Pegatron, Wistron, Luxshare do. What Apple owns is the IP (Apple Silicon's A- and M-series chips), the OS family (iOS / iPadOS / macOS / watchOS / visionOS), the App Store, the brand, and direct distribution. R&D is 8.3% of revenue ($34.6B). Capex is 3.1% — extreme for a hardware company with $416B in revenue. The result: return on invested capital is 48% and the cash conversion cycle is −43 days: Apple is paid by the channel weeks before it pays suppliers, so working capital throws off cash on growth instead of consuming it.

The picture is asymmetric. iPhone is half of revenue but only ~40% of gross profit; Services is a quarter of revenue and roughly 42% of gross profit. iPhone exists to put a screen in a pocket so Services can be sold against it. That is the business model in one sentence.

2. The Playing Field

There is no single peer. Apple competes against three sets of companies in three different economic registers, and the right peer table puts them all on one page so the gap is visible.

What the peer set reveals. HPQ and DELL — Apple's nominal "PC competitors" — trade at 0.6–0.8× sales, single-digit P/Es, 11% FCF yields. That is what pure consumer-hardware economics earns. MSFT, GOOGL, and META — the platform peers — trade at 8–13× sales, 28–37× P/E, 2–3% FCF yields. That is what platform economics earns. Apple sits on top of the platform cluster on EV/Sales (9.1×) and EBITDA (26×) despite hardware being three-quarters of its revenue. Two implications: (1) the market is paying Apple a platform multiple for the consolidated business, so any erosion at the Services layer (regulation, search-deal loss) is the leveraged risk; (2) hardware peers offer the cheapest tail-hedge on PC-cycle exposure, so DELL/HPQ valuations are the floor — not the comparable.

China is the one to watch. It was the only weak line in FY2025 (−4%) and printed +38% YoY in the December 2025 quarter on the iPhone 17 cycle. If that proves durable rather than a single-quarter pull-forward, it removes the largest open headwind on the consolidated story.

3. Is This Business Cyclical?

Yes — but the cycle hits a different P&L line for each segment, which is why headline revenue looks smoother than the underlying volatility.

The pattern is clear. Apple is cyclical at the hardware unit and gross-margin lines, structurally non-cyclical at the Services line, and uncorrelated with the broader macro at the cash-return line. iPhone unit volume contracts in down cycles (FY2016 was −8% revenue on iPhone 6S air-pocket; FY2019 was −2% on the China/trade-war shock; FY2023 was −3% on post-COVID device hangover). What does not contract is Services and what does not contract is buybacks — Apple has bought back stock every year for the last 13 years regardless of the cycle, retiring 44% of shares from FY2012 to FY2025.

The 2026 cycle is supply-driven, not demand-driven. Memory (DRAM/NAND) is being diverted to AI data-center customers, and Apple flagged on the Q1 FY2026 call that memory will be "a bit more of an impact" to Q2 FY2026 gross margin. Customer pull on iPhone 17 was strong enough to deplete channel inventory in a single quarter. The risk is margin, not units.

4. The Metrics That Actually Matter

Asset turnover, current ratio, inventory days are mechanical for a company that outsources manufacturing and runs a negative cash-conversion cycle. They tell you nothing about whether Apple is winning or losing. The numbers below explain almost all the value-creation and value-destruction in this business going forward.

The equation: Services revenue = install base × ARPU. Both are growing. Install base ~7% CAGR (2018→2025) × ARPU ~6% CAGR = ~14% Services revenue CAGR. If either slows materially — install base from a smartphone air-pocket, ARPU from regulation — the growth equation breaks. If both keep compounding for another 5 years, Services alone would be a >$200B business by FY2030.

5. What Is This Business Worth?

Apple is best valued as a sum of two parts plus a call option, not as one business at one multiple. The hardware franchise and the Services franchise have wildly different gross margins (37% vs 76%), capital intensity, growth profiles, and competitive sets. Pretending the consolidated 32% operating margin and 34× P/E are the "right" lens for either piece is what causes investors to either anchor too high (against PC peers) or too low (against software peers).

The arithmetic gives a SOTP floor of roughly $2.7–2.8 trillion (hardware ~$1.0T + Services ~$1.7T + net cash $54B). Apple trades at ~$4.2T. The implied ~$1.4T premium is the market's price for: (a) the integration scale that lets hardware drive Services in a way no other company has matched, (b) the Apple Intelligence + Siri-overhaul AI option, (c) capital-return discipline (FY2025: $97B in buybacks + $15B in dividends), and (d) Vision Pro / spatial computing optionality.

Roughly half of Apple's EPS growth has come from share-count reduction. This is the second engine of the model and is structurally tied to Free Cash Flow staying near $100B; if FCF stalls, buyback pace drops mechanically.

6. What I'd Tell a Young Analyst

Stop modeling iPhone units in isolation. The iPhone unit is the customer-acquisition cost of the Services franchise. A flat iPhone unit number with a growing install base and growing Services ARPU is a win, not a stagnation. Track install base + ARPU; let units be derivative.

Re-segment the company yourself. Apple reports revenue by geography and by product. Build your own P&L that separates Hardware vs Services, allocate opex by the disclosed gross margins, and value the two pieces against their actual peer sets. The consolidated multiple hides the answer to almost every interesting question.

Watch three things weekly, not the share price. (1) Memory spot pricing on TrendForce / DRAMeXchange — the FY2026 gross-margin variable. (2) Greater China shipment data from IDC and Counterpoint — the largest open headwind that just inflected. (3) EU Commission, DOJ, and Japan METI dockets for App Store rulings — the largest open risk. Two of three turning negative would change how the market values earnings independent of the print.

Take the buyback yield seriously. Three percent of market cap returned annually with 50% incremental return on invested capital is a real number. It is also why Apple can grow EPS in years when revenue is flat or down — and why long-term holders have been rewarded even through the 2016, 2019, and 2023 air pockets.

The thesis that breaks Apple is not iPhone fatigue. It is regulatory take-rate compression at the App Store + the search-distribution payment going to zero + Services growth slowing below install-base growth (which would mean ARPU stalled). All three would have to happen together to break the model. One alone is absorbable. Two would compress the multiple. Three would re-rate the stock to a hardware-plus-modest-services multiple — that is the bear case worth knowing precisely what would trigger.