Apple's Q3 FY26 earnings landed on July 30, 2026 with a June-quarter record — $109.4 billion in revenue, up 16% year-over-year — and the stock still fell roughly 6% in after-hours trading. Beats on revenue, earnings, and margin were not enough to offset a September-quarter guide that came in below consensus and a chip-supply squeeze that Tim Cook, on his final call as CEO, blamed on Apple underestimating its own demand.
The quarter matters beyond Apple. It closed a 48-hour window in which Microsoft, Meta, Amazon, and Apple all reported — and Apple was the only one of the four to guide below expectations while also being the only one not running a $100B-plus-a-year AI infrastructure buildout. Whether that is discipline or exposure is now the central question hanging over the stock.
This breakdown covers the record numbers and their one-time asterisk, the guidance math that triggered the selloff, the supply constraint Cook described in unusually candid terms, Apple's outlier position on AI capex, and the one genuinely dated commitment on the call: new Siri AI capabilities shipping this fall — unevenly across markets.
- 01A record quarter could not carry a weak guide.Revenue of $109.4B (+16% YoY) and EPS of $2.02 (+29% YoY) both beat consensus, yet the shares dropped ~6% after hours because Apple guided September-quarter revenue growth to 9–11% YoY against the ~12% analysts had modeled.
- 02The constraint is supply, not demand.Cook named capacity on the advanced nodes that fab Apple's chips as the primary limit — a demand-forecast miss, in his words, because iPhone and Mac outsold Apple's own plans. Supply headwinds are guided to increase significantly into September.
- 03Apple is the AI capex outlier.Apple's fiscal-2026 capex is estimated around $13–14B while the four biggest hyperscalers have guided a combined $700B-plus for 2026 — roughly fifty times more. Apple rents and partners for AI compute instead of building it.
- 04Siri AI ships this fall — but not everywhere at once.Cook confirmed the redesigned Siri arrives for users this fall, with the EU unresolved for iPhone and iPad and China further behind. Apple also signaled a possible paid iCloud+ tier tied to heavy assistant usage.
- 05An era ended on the call.This was Tim Cook's 90th and final earnings call as CEO. John Ternus, who fielded a live analyst question, takes over on September 1, 2026 — inheriting the supply squeeze, the Siri launch, and the capex question in his first month.
01 — The NumbersA record quarter with one asterisk.
The headline results were unambiguous beats. Revenue of $109.4 billion rose 16% year-over-year and topped the roughly $108.7–108.9B consensus. Diluted EPS of $2.02 rose 29% year-over-year against a consensus near $1.89. Net income was $29.8 billion, operating cash flow hit a June-quarter record of $34.4 billion, and blended gross margin reached 50.1%, up from 46.5% in the year-ago quarter. Operating expenses grew faster than revenue — up 23% year-over-year to $19.1 billion — which CFO Kevan Parekh attributed to R&D investment.
The asterisk: reporting citing MarketWatch put a one-time tariff refund — tied to the Supreme Court's ruling against the prior administration's global tariffs — at roughly $2.19 billion, contributing about 11 cents of the $2.02 EPS and more than 2 percentage points of the 50.1% gross margin. Strip the refund out and EPS lands around $1.91 — still a hair above consensus, but the margin and earnings beats are less clean than the headline suggests, and the benefit does not repeat.
Q3 FY26 revenue mix · share of the $109.4B total
Source: Apple Newsroom, MacRumors, 9to5Mac — Jul 30, 2026. Each bar is the segment's share of the $109.4B reported total; segment figures rounded.The mix tells a more balanced story than the record headline. iPhone (+22% year-over-year) and Mac (+29%, credited by Cook to the MacBook Neo and MacBook Pro) both set June-quarter records. But two of five segments underperformed: iPad revenue fell 6% year-over-year against a consensus near $6.9B, and Services — the high-margin engine at a 75.6% gross margin — grew 12% but missed the ~$31.2B estimate. Greater China rose about 22% year-over-year to roughly $18.7–18.8 billion, yet also came in under the ~$19.5B consensus, even as Cook called out June-quarter records for iPhone there and the best Mac quarter in company history for the region.
02 — The SelloffWhy a record quarter still fell ~6% after hours.
The selloff was about the next quarter, not the last one. Parekh guided September-quarter total revenue to grow 9–11% year-over-year — below the roughly 12% analysts polled by LSEG had modeled — and gross margin to 47–48%, down from Q3's refund-boosted 50.1%. He named two drags: a foreign exchange headwind of about 2.5 percentage points sequentially, and supply constraints that he said would increase significantly moving from June to September, hitting iPhone, Mac, and iPad.
YoY growth guided
Below the ~12% YoY growth analysts polled by LSEG had modeled. For a stock that had gained roughly 25% year-to-date and briefly crossed a $5 trillion market cap, adequate was not enough.
down from 50.1% in Q3
The step-down partly reflects the non-repeat of the one-time tariff refund that flattered Q3, plus rising component costs Apple expects to keep paying beyond September.
down from 22% in Q3
Explicitly attributed to FX and supply headwinds rather than softening demand — Cook's argument is that Apple cannot build enough, not that customers stopped buying.
Context sharpened the reaction. Reported after-hours declines ranged from about 4.5% to 6.65% off the prior close depending on the snapshot — call it roughly 6%. The bar was unusually high going in: the stock had gained about 25% year-to-date and briefly crossed a $5 trillion market cap before the report, so guidance merely in line with a supply-constrained reality read as disappointment. The same week, Amazon posted its own record quarter the same day — the contrast in market reaction across the two reports is the week's cleanest illustration of how differently investors are pricing AI-infrastructure sellers versus AI-infrastructure renters.
03 — The Supply SqueezeAdvanced nodes, memory costs, and a candid admission.
Cook was unusually direct about what is limiting Apple: capacity on the advanced semiconductor nodes that fabricate Apple's own system-on-chip processors. His framing matters — this is not a supplier failing Apple, it is Apple having under-forecast its own product demand.
"The primary issue is advanced nodes that we run our SOCs on. That's the primary supply constraint now, and the root cause of it is not a regular supply issue. It's a demand forecast issue, to be candid, where the iPhone and the Mac are both doing remarkably better than we thought they would do."— Tim Cook, CEO, Apple Q3 FY26 earnings call, Jul 30, 2026
A second squeeze is coming from memory. Apple raised prices on Mac — including the MacBook Neo — and iPad during the quarter, citing memory and processor-manufacturing-capacity shortages, and Cook said Apple expects to pay even higher memory costs beyond September as market pricing continues to rise. Industry coverage widely attributes that memory crunch to hyperscalers pulling memory-chip supply toward AI datacenter buildouts — a framing worth noting is the press's, layered on top of the call, not a causal claim Apple itself made.
If that attribution is right, there is a real irony here: Apple is the one major platform company not spending on AI datacenters, yet its costs, its device prices, and now its guidance are being squeezed by the AI datacenter boom anyway. Apple is paying the AI-capex tax without capturing the AI-capex upside — which makes the strategic bet in the next section the most important thing on the call.
04 — The Strategic BetThe AI capex outlier.
Every earnings report this week was covered in isolation. Put them side by side and Apple's position is singular. Analyst compilations put Apple's fiscal-2026 capital expenditure around $13–14 billion, up modestly from a reported $12.7 billion in fiscal 2025 — while the four largest hyperscalers have guided a combined $700 billion-plus for 2026. Both figures are secondary-aggregated estimates rather than same-basis SEC filings, so treat the ratio as illustrative — but even with generous error bars, Apple is spending on the order of fifty times less than its Big Four peers combined.
| Company | 2026 capex guide (est.) | Basis | Multiple of Apple's guide |
|---|---|---|---|
| Apple | ~$13–14B | FY2026, analyst-compiled | 1× (baseline) |
| Alphabet | up to ~$205B | CY2026, tracker estimate | ~15× |
| Amazon | ~$200B | CY2026, tracker estimate | ~15× |
| Microsoft | ~$190B | CY2026, tracker estimate | ~14× |
| Meta | up to ~$145B | CY2026, tracker estimate | ~11× |
| Big Four combined | >$700B | aggregated, mixed windows | >50× |
Sources: The Motley Fool (Jul 27, 2026) and third-party capex trackers; secondary-aggregated estimates, not same-basis SEC filings. Multiples computed against the ~$13.5B midpoint of Apple's estimated guide and rounded — illustrative only. Fiscal-year windows differ, and company guidance is not always quantified: on its July 29 call Microsoft said only that FY2027 capex will grow year-over-year, alongside roughly $175B for calendar 2026 and more than $50B in the September quarter — a different basis from the calendar-2026 tracker figure shown here.
Asked directly on the call whether Siri AI changes Apple's capital intensity, Cook confirmed a hybrid model — some third-party cloud, some Apple-owned data centers — and said Apple has been growing its operating expenses and spending "quite a bit more" on AI, while declining to commit to how compute costs would be recovered. In other words: the outlier position is deliberate, and the spend that does exist is showing up in opex, not capex.
The market's read is evolving. What was framed for most of 2025 as Apple missing the AI wave is increasingly being framed as capital-efficiency discipline — Apple lets others fund the infrastructure and rents the results.
For the full five-company comparison — including how Microsoft, Meta, and Amazon each justified their own guides this week — see the earnings-week AI capex scorecard, which pairs naturally with Alphabet's own Q2 earnings breakdown from earlier in July.
05 — Siri TimingSiri AI ships this fall — unevenly.
The most operationally useful sentence on the call was a date. Cook confirmed that the redesigned Siri's new AI capabilities reach users this fall, noting the assistant has been in developer beta since immediately after WWDC and moved to public beta a few weeks before the call.
"We're looking forward to bringing these new capabilities to users this fall, in addition to amazing new updates across our operating systems."— Tim Cook, CEO, Apple Q3 FY26 earnings call, Jul 30, 2026
One attribution point worth being precise about: on the call itself, Cook and Parekh referred to the product only as "Siri AI" and "Apple Intelligence" — the word "Gemini" never came up. The Gemini connection comes from separately reported coverage of Apple's January 2026 deal with Google, which pegged the licensing cost at roughly $1 billion a year (per Bloomberg's Mark Gurman) and up to $5 billion multi-year (per Deepwater's Gene Munster) — analyst estimates never confirmed by either company; the full mechanics are covered in our breakdown of the Gemini-powered rebuild of Siri.
Just as important: the fall launch is not global. In the EU, Cook said Apple is working closely with the European Commission and has not yet reached a solution for iPhone and iPad (the Mac is exempt from the relevant regulation). In China, Apple received approval only the week before the call to ship the original, pre-Siri-AI Apple Intelligence features, with Cook saying more work is required down the road for Siri AI — Apple's phased China rollout of Apple Intelligence just gained a fresh, dated data point.
There is also an early monetization signal. Cook said plans for recovering Siri AI's compute costs are still "early going," but pointed to "some kind of upgrade possibilities on iCloud Plus" — the first public hint of a paid tier tied to heavy assistant usage, sitting on top of a paid-subscriptions base Cook put at 1.5 billion, a figure that includes iCloud and App Store subscription cuts.
06 — The TransitionNinety calls, and a handoff in progress.
This was Tim Cook's 90th and final earnings call as CEO. John Ternus — announced as his successor back in April 2026 — was on the call and took a live analyst question, an unsubtle signal of how deliberately Apple is staging the handoff before Ternus formally takes over on September 1, 2026. Cook's own framing: "The transition is going seamlessly, and I am beyond excited for John to step into his new role and lead Apple into its next era."
The timing hands Ternus a dense first month: the September quarter Apple just guided below consensus, the supply squeeze Cook says will intensify into that same quarter, the fall Siri AI launch, and the open question of whether the rent-don't-build AI posture survives contact with rising compute demand. His first earnings call as CEO will be the scorecard on all four.
07 — The PlaybookWhat this quarter means for brands and marketers.
Strip out the investor drama and four practical threads fall out of this report for operators — each with a date or a price attached rather than a vibe.
Get answerable before fall
Siri AI now has a shipping window, not a rumor cycle. Structured product data, clean feeds, and content an assistant can actually cite stop being optional the day hundreds of millions of iPhones get a conversational front door to discovery.
Plan market by market
US and most markets get Siri AI this fall; the EU is unresolved for iPhone and iPad; China trails further. Brands operating across those regions should stage assistant-readiness work in the same order the launch actually lands.
Device costs are rising
Apple raised Mac and iPad prices mid-quarter on memory and processor-capacity shortages, and Cook expects memory costs to keep climbing beyond September. Teams planning device refreshes should budget for higher prices, not wait for relief.
Watch the monetization model
Apple is betting it can monetize AI through distribution — 1.5 billion paid subscriptions and a hinted iCloud+ tier — while renting the underlying compute. If that works, it is the template for every company weighing build-versus-partner on AI.
The through-line is that assistant-mediated discovery just got a calendar. Making a brand answerable to an AI assistant — feeds, schema, content structured for citation rather than clicks — is exactly the ground our agentic SEO practice covers, and for teams working out where assistant channels fit in a broader stack, our AI transformation engagements start with precisely this kind of dated-window planning. Looking forward, the pattern to watch through fiscal 2027 is whether Apple's opex-heavy, partner-supplied AI posture holds once Siri AI is live at scale — if usage-driven compute costs spike, the company that skipped the capex cycle may end up paying for it through the income statement instead.
08 — ConclusionA record priced for more.
Apple beat everything, guided below the bar, and put a date on Siri.
The June quarter was genuinely strong — record revenue, record operating cash flow, records for iPhone and Mac — and the market still marked the stock down roughly 6% after hours because the September guide of 9–11% growth came in under the ~12% consensus. After a ~25% year-to-date run and a brush with a $5 trillion valuation, Apple is being priced for acceleration, and supply physics just got in the way.
The strategic story is the one worth carrying forward. Apple spends an estimated $13–14 billion a year on capex while its four biggest peers guide a combined $700 billion-plus, rents its frontier AI through partners, and — per the irony this quarter made concrete — still absorbs the AI boom's costs through memory prices and fab capacity it does not control. Whether that reads as discipline or exposure depends entirely on how the fall Siri launch performs.
That launch now has a date, a staggered map, and a hinted price. For operators, the actionable takeaway is not the stock move — it is that assistant-mediated discovery on the world's most valuable consumer platform starts this fall in the US, later in the EU, later still in China. The brands that treat that as a deadline rather than a headline will be the ones an AI Siri can actually find.