Record $114.5B quarter, record iPhone, record Services — and the stock fell anyway. New CEO John Ternus said memory pricing is "unprecedented" and still rising. This is a cost squeeze, not a demand break.
中文版 Chinese edition →HOLD / WATCH Apple at $329.08. Scale only $300–312. Target $365–375. Invalidation: daily close below $285. 1–3 months. Conviction: WATCH.
Beat on sales ($114.5B, +12%) and EPS ($2.10) was erased by December gross-margin guide of 45–46% vs 46.3% expected — almost all memory.
Consensus calls this a one-quarter digestion. Ternus said pricing keeps rising beyond December and carry-in inventory benefit fades — this lasts 2–3 quarters.
December gross margin printing inside 45–46% with iPhone Duo still constrained = pricing power intact. Below 45% = elasticity broke.
The takeaway: Apple beat everything except the two numbers that price the stock — next-quarter margin and Services growth.
Q4 FY2026 (ended Sept. 26): revenue $114.5B, +12% YoY, vs ~$113.4B expected. EPS $2.10, +14%, vs $1.99. iPhone $57.6B, +17% — September-quarter record despite supply constraints and FX headwinds. Mac $10.8B, +24%. Services $31.6B, +10%, all-time record. Net income $30.7B record. Full-year revenue $478.8B (+15%) and EPS $8.98 (+20%) — best year ever.
Then the guide: December revenue +7% to +10% YoY (would be Apple's best quarter ever, ~$154B at midpoint, above 6.1% consensus), but gross margin 45–46% vs 46.3% expected, and Services +8–9% vs ~10.7% expected. CFO Kevan Parekh said most of the sequential margin fall is memory, plus Siri AI ramp costs. After-hours AAPL fell to $328.81, ~$139B of value gone in minutes; it traded $329.08 (‑2.7%) intraday Tuesday. The details are in the Apple Q4 press release, the full earnings-call transcript, and the call highlights summary.
The takeaway: every growth engine except supply-constrained iPad is firing; Duo demand exceeds supply with no balance date.
iPhone upgraders hit a September-quarter record globally, in the U.S., mainland China and India. Mac had its best fiscal year ever for units, gaining PC share per IDC. Installed base: 2.5B+ active devices, paid subscriptions 1.5B+, double-digit growth. COO Sabih Khan: iPhone Duo is constrained and "really hard to predict when supply and demand will be in balance"; December Mac/iPad constraints will be "primarily memory-driven" on higher-than-expected demand. September constraints were advanced-node SoC (easing on 2nm); December is memory. That distinction matters: SoC tightness is Apple-specific ramp, memory tightness is industry-wide and priced in dollars per gigabyte.
The takeaway: strip the one-time tariff refund and the entire sequential margin drop is memory — and the cheap-inventory cushion is burning off.
September gross margin was 47.9%, including ~1 point of tariff refunds ($0.05 of EPS; "vast majority" of refunds now received). Ex-refund, June→September fell ~1 point entirely on memory, partly cushioned by carry-in inventory and cheaper non-memory parts. December guide midpoint 45.5% implies another ~140 bps sequential fall ex-refund, most of it memory. On a ~$154B December quarter, every 100 bps of margin is ~$1.5B of gross profit — the 240 bps headline drop (47.9→45.5) is roughly a $3.5B+ gross-profit headwind, partly offset by the June Mac/iPad and September iPhone price hikes.
Three accelerants, per TechCrunch's inventory reporting and Quartz's margin breakdown:
The takeaway: Apple sits on the wrong side of the AI buildout — hyperscalers' HBM demand crowds out the commodity DRAM/NAND in every iPhone.
Suppliers prioritized high-margin HBM for AI accelerators; commodity DRAM/NAND supply was starved just as on-device AI (bigger models, larger working memory) lifted content per phone. Industry checks cited in 2026 outlooks put DRAM revenue +50% and NAND +45% YoY, with DRAM contract prices +40%+ in a single quarter at points this year. Micron at $860 (+3.2% Tuesday while Apple fell) is the mirror: Apple's margin dollar is Micron's revenue dollar. Until HBM capacity additions free wafer supply — widely not expected before late 2027/2028 — Apple pays the AI tax without collecting AI-cloud revenue to match.
The takeaway: Apple already raised Mac/iPad prices in June and iPhone in September; December tells us if buyers blinked.
Parekh: "we did reluctantly make the decision to change pricing" and "it's really early to really assess any elasticity impact." Translation: price hikes cover part of the hole, but Apple does not yet know the demand cost. China ($17.54B vs ~$18.75B expected) and Services deceleration (8–9% guide on FX + licensing + App Store link-out ruling) are the two places elasticity or regulation shows first. iPad ‑10% to $6.2B (missed $6.7B) and Wearables ‑8% to $8.3B were framed as supply + hard compares — plausible, but also the most price-sensitive categories.
Bulls get: record revenue guide, Duo backlog, 2.5B installed base, $137B cash vs $80B debt, $24B quarterly buyback. Bears get: rising purchase prices into fading inventory benefit, Siri AI inference costs hitting Services COGS before paid AI tiers contribute, and opex $20.4–20.8B (+11–13%) growing faster than the 7–10% revenue guide.
The takeaway: at ~38× trailing earnings with margin falling, Apple is priced for the demand it has and the margin it no longer has.
AAPL at $329 is ~36–38× trailing EPS ($8.98 FY2026), ~44× forward per handler data, 0.5% yield, 52-week range $243–$352 (high Oct. 23). A 200–250 bp sustained gross-margin impairment on ~$480B+ revenue is ~$10B+ of gross profit — roughly $0.60–0.70 of annual EPS before offsets. Price hikes + mix + buyback can cover perhaps half. My fair-value math (opinion): $310–330 on $9.00–9.50 normalized EPS at 34–36×; $365+ requires margin back above 46.5% and Services re-accelerating to double digits. That is a show-me, not a buy-the-dip.
| Trade | Detail |
|---|---|
| Rating | HOLD / WATCH — no chase above $335 |
| Entry | Scale $300–312 (prior breakout + 50-day zone); starter only above $320 |
| Target | $365–375 base (margin stabilizes + Duo ramp); stretch $395 if March GM ≥46% |
| Invalidation | Daily close below $285 (breaks Sept. breakout + prices structural margin loss) |
| Horizon / Conviction | 1–3 months / WATCH (one signal confirmed — margin guide — elasticity unconfirmed) |
| Position | Half-size max until January print; pair long memory (MU) or hold cash — see below |
The takeaway: base case is sideways-to-down until January; the tails are defined by two prints.
| Scenario | Weight | Trigger | Price |
|---|---|---|---|
| Bull: pricing power wins | 25% | Dec GM ≥46% + Services ≥9% + Duo lead times shrinking by January | $375–395 |
| Base: squeeze lingers | 50% | Dec GM 45–46% as guided, Services 8–9%, memory pricing flat-to-up into March | $300–340 chop |
| Bear: elasticity breaks | 25% | Dec GM <45% or revenue miss on units; China down YoY; March GM guided lower again | $270–290 |
The takeaway: I am bearish on margin duration, bullish on demand — either leg can falsify the stance.
The takeaway: if you must be long this cycle, be long the seller of the shovels, not the buyer.
| Item | Value |
|---|---|
| AAPL intraday Nov 3 ~16:00 UTC | $329.08, ‑2.73%, range $325.33–334.50, prev close $338.32 |
| Q4 FY2026 revenue / EPS | $114.48B +12% / $2.10 +14% (incl. $0.05 tariff refund) |
| iPhone / Mac / Services / iPad / Wearables | $57.6B +17% / $10.8B +24% / $31.6B +10% / $6.2B ‑10% / $8.3B ‑8% |
| Gross margin Sept / guide Dec | 47.9% (incl. ~1pt refund) / 45–46% vs 46.3% consensus |
| Dec revenue / Services guide | Total +7–10% YoY / Services +8–9% vs ~10.7% expected |
| Inventory / purchase obligations | $14.54B (+154.3% YoY, +31.1% QoQ) / +$33B QoQ |
| Cash / debt / return | $137B / $80B / $28B ($24B buyback + $3.9B div); $0.27 div payable Nov 16, record Nov 12 |
| Fundamentals (handler) | P/E TTM 38.2, forward ~44.0, rev growth 14.2%, 52w $243.42–$351.73, yield 0.51% |
| Context | SPY $756.04 +0.5%, MU $860.22 +3.2%, MSFT $516.51 +1.9%, BTC $77,674; Election Day Nov 3 |
Simplified Dec-quarter bridge (opinion model): $154B revenue at 45.5% GM = $70.1B gross profit. At 47.9% = $73.8B. Gap ≈ $3.7B pre-offset. Assume price hikes + mix + non-memory savings recover ~$1.5B, leaving ~$2.2B net hit ≈ $0.13–0.15 quarterly EPS drag. Annualized with Q1-weighting ≈ $0.50–0.70. Load-bearing assumption: DRAM/NAND purchase prices stay flat-to-up through March 2027. If wrong (prices roll over on demand destruction or HBM capacity conversion), the entire margin thesis reverses and fair value moves $30+ higher. Watch TrendForce monthly contract prints and MU's next gross-margin guide as the independent check.
Opinion section is the author's judgment, not investment advice. Prices intraday Nov 3, 2026 (~16:00 UTC); Tuesday, Nov 3, 2026 verified weekday. No blocked domains used. Financial-data handler for levels; page extraction for narrative only.