December-quarter gross margin guide
45-46%
Guided Nov 2, 2026, versus 47.9% reported in the September quarter
Services net sales, fiscal Q4 2026
$31.57B
Up about 10% year over year, an all-time record; 8-9% growth guided for December
Products gross margin
37.2%
September quarter 2026, against 76% for Services
iPhone 18 Pro memory bill
+~400%
TrendForce estimate for 3Q26 versus a year earlier
Micron Mobile & Client gross margin
90%
On $13.11B of revenue in the quarter ended Sept 3, 2026
Apple inventory
$14.5B
Sept 26, 2026 balance sheet, up from $5.88B on Dec 27, 2025
The November 2 earnings call
One Call, Two Stories: A Margin Hole and a Revenue Hint
Apple's fiscal fourth quarter looked like a record on the surface: revenue of $114.48 billion, up 12%, and diluted earnings of $2.10 per share. One call later, the December-quarter gross margin guide of 45-46% put a hole in the outlook that memory prices carved, and CFO Kevan Parekh said AI offers \"opportunities across third-party apps and partnerships\" that can \"contribute to revenue growth over time.\" No take rate, fee schedule or launch date was disclosed, so the toll is inference rather than policy. The pairing is still deliberate: the margin guide is the hole, the third-party app language is the candidate fix.
| Metric | Fiscal Q4 2026 (ended Sept 26, 2026) | Context |
|---|---|---|
| Total net sales | $114.48B | Up 12% year over year; fiscal 2026 revenue of $479B was a record |
| iPhone | $57.58B | Above the $56.54B consensus |
| Services | $31.57B | Up about 10%, an all-time record; 8-9% growth guided for December |
| Company gross margin | 47.9% | Includes roughly 1 point of tariff refunds |
| Products gross margin | 37.2% | Services gross margin 76%, up 40bps sequentially |
| Diluted EPS | $2.10 | Up 14%, including $0.05 from tariff refunds |
| Greater China | $17.54B | About 15% of quarterly net sales |
| December-quarter guide | Gross margin 45-46%; services +8-9% | Consensus gross margin was 46.3% |
- Ternus on the cost side: \"What we're seeing with memory pricing right now is unprecedented.\"
- Parekh on the arithmetic: excluding tariff refunds in both June and September, company and products gross margin \"would be down about 1 point sequentially, primarily driven by higher memory costs.\"
- On the revenue side, Parekh pointed to \"opportunities across third-party apps and partnerships as well as first-party opportunities,\" and to a record number of paid subscriptions with double-digit growth for third-party apps on the App Store.
- What was not said: any take rate on AI app transactions, any fee on AI app subscriptions, any launch date, and any confirmation that a third-party AI app programme exists at all.
From Mac to iPhone to iPad, we are proud that our products are the best platforms in the industry for AI for both first-party and third-party applications, and we're excited about the opportunities that creates.
Where the margin went
The Hole Is Upstream: $1.6 Billion a Quarter, and the Suppliers Already Took It
Strip out the roughly one point of tariff refunds and Apple's underlying September gross margin was about 46.9%. A 45-46% December guide implies roughly 140 basis points of sequential compression, which an analyst on the call framed as \"at least\" that much, with management attributing most of it to memory costs and adding Siri AI ramp expense and hardware deferrals on top. Applied to the September revenue base, that is a hole of about $1.6 billion to $1.8 billion of gross profit a quarter. The money did not vanish. It moved upstream, to the companies that set the price of the DRAM and NAND in every iPhone.
| Company | Latest reported quarter | Revenue | Profit metric | Margin |
|---|---|---|---|---|
| Apple | Fiscal Q4 2026 (ended Sept 26, 2026) | $114.48B total | Products gross margin 37.2%; Services 76% | Under pressure |
| Micron | Fiscal Q4 2026 (ended Sept 3, 2026) | $54.23B | $47.05B gross profit | 86.8% gross margin; Mobile & Client 90% on $13.11B |
| SK hynix | Q3 2026 | ₩104.62T | ₩83.33T operating profit | 79.7% operating margin |
| Samsung Electronics | Q3 2026 | ₩195.25T | ₩107.44T operating profit | Device Solutions: ₩147.1T revenue, ₩106.5T operating profit |
Apple's company gross margin by fiscal quarter, with the December-quarter guidance midpoint
Gross margin computed from reported gross profit and net sales. September includes roughly 1 point of tariff refunds; the December bar is the midpoint of guidance.
Unit: %
FQ1 2026
Quarter ended Dec 27, 2025
48.2%
FQ2 2026
Quarter ended Mar 28, 2026
49.3%
FQ3 2026
Quarter ended June 27, 2026
50.1%
FQ4 2026
Reported, including tariff refunds
47.9%
Dec quarter guide midpoint
Midpoint of 45-46% guidance
45.5%
- Apple's own inventory balance more than doubled, from $5.88B at the end of December 2025 to $14.54B at the end of September 2026. That pattern is consistent with pre-buying memory before contract resets, though Apple attributes cost pressure to component supply constraints rather than stockpiling.
- Apple's Q3 fiscal 2026 Form 10-Q says the company is \"experiencing a period of supply constraints and increasing costs for components,\" naming \"advanced semiconductors, storage (NAND) and memory (DRAM),\" and expects those trends to \"intensify.\"
- SK hynix now expects the market structure \"in which demand exceeds supply to persist through 2030,\" with HBM4 ramping into volume and server DDR5 absorbing capacity.
- Micron guided fiscal Q1 2027 to $61.5B of revenue at about 86% gross margin, keeping handset and client pricing power intact into Apple's December build.
Apple's inventory balance, fiscal 2026 quarters
Reported inventories rose through fiscal 2026 as memory contract prices reset.
Unit: $ billions
Dec 27, 2025
5.9
Mar 28, 2026
6.8
Jun 27, 2026
11.1
Sep 26, 2026
14.5
TrendForce says that Apple won't be passing on the full impact to customers, and will be increasingly looking to Services revenue to offset higher hardware manufacturing costs.
The toll and the law
Why a 30% Take on AI Apps Is the One Fee Apple Cannot Easily Collect
The instrument Apple would need to refill a $1.6 billion quarterly hole is a cut of third-party AI app revenue. Its own filings describe how tight that path has become. The April 2025 ruling in Epic Games v. Apple enjoined Apple from imposing any commission on purchases made outside an app and from restricting how developers steer customers to those purchases; the Ninth Circuit modified parts of that in December 2025 while allowing some commission on link-out purchases; Apple petitioned for Supreme Court review in May 2026, and the Supreme Court granted review on June 30, 2026 of the legal standard for civil contempt. In Europe, Apple's own rewritten terms took effect Oct 1, 2026, replacing a per-install fee with a 5% Core Technology Commission on digital transactions in apps distributed outside the App Store.
| Route to revenue | What it costs today | Constraint |
|---|---|---|
| US in-app purchase subscriptions | 30% standard commission | Applies to AI apps today, but a new AI-app-specific surcharge would be tested against the same standards now before the Supreme Court |
| US link-out purchases | No commission since the April 30, 2025 injunction | Supreme Court is reviewing the contempt ruling; the standard for such fees is unsettled |
| EU distribution outside the App Store | 5% Core Technology Commission from Oct 1, 2026 | Apple's own remedy under the Digital Markets Act |
| EU alternative payment processing | 20%, or 10% reduced | Digital Markets Act framework |
| Paid first-party AI tier | To be priced; Apple says expanded Siri AI access will be available for a fee | None of substance; Apple keeps the whole margin |
| Model partnership economics | Revenue share, not a store commission | Already the model, with Gemini powering Siri |
There is a second reason to doubt the simple version of the story. The gatekeeper changed hands just before the call. Phil Schiller stepped back from leading the App Store and Apple's product events in September 2026 and became an Apple Fellow, with the App Store returning to Eddy Cue inside Services. That places the App Store, the Services margin problem and the new revenue hunt under one profit-and-loss statement, with a first-time chief executive asking questions. It also means the person who publicly defended the commission model is no longer running it.
The labs' counter-move
The Labs Are Building Storefronts Specifically Designed to Route Around a Toll
The labs Apple hopes to tax are not waiting for the fee to be written. OpenAI is turning ChatGPT into a place where software is discovered, launched and used, with app suggestions inside the conversation, in-chat launch, an enterprise marketplace with more than 30 partners at launch, and agentic assistants running in their own cloud infrastructure against an ecosystem of more than 4,000 apps. OpenAI announced no billing or revenue-sharing system, and noted that users could instead subscribe on an app maker's own website. Meta went further on distribution: Muse launched Sept 8, 2026, overtook ChatGPT as the leading free iOS app in the US within days, and has been free at the top of the chart on a paid tier of $20 or $100 a month.
| Surface | Reach | Monetization | Exposure to an Apple AI-app toll |
|---|---|---|---|
| Apple App Store | Nearly 1B average weekly visitors; over 1.5B paid subscriptions | Commission on subscriptions and in-app purchases | Lowest risk; Apple is the toll collector, but its fee options are legally constrained |
| ChatGPT | More than 800M weekly users as of OpenAI's October 2025 disclosure | Direct subscriptions plus app-maker web checkout | High; OpenAI is explicitly building a storefront and web billing |
| Meta's Muse | More than 2.5M downloads since the Sept 8 debut; 730,000 in roughly five days | Free app with monthly tiers at $20 and $100 | High on price, low on distribution, since Muse acts across the web |
| Claude | About 400,000 downloads in the comparable post-launch window | Direct subscription | High |
| Grok | About 200,000 downloads in the comparable post-launch window | Direct subscription | High |
- In the same 13-day post-launch window, ChatGPT took 3.1M downloads while Muse took 2.5M, ahead of Claude at 400,000 and Grok at 200,000, so the new entrant was already within reach of the incumbent on Apple's own chart.
- ChatGPT had reached roughly $3 billion of cumulative consumer spend on mobile apps, per Appfigures, which is the revenue pool a 30% take would apply to.
- Apple's own Services line, at $31.57B in the September quarter, is already growing more slowly than the apps it would tax: 14% in fiscal 2025, about 10% in the September quarter, and 8-9% guided for December, with roughly half a point of sequential FX headwind and mobile gaming softness also cited.
- The strategic point is enforcement. Meta's Muse and OpenAI's agents act across the open web, where an in-app commission has no jurisdiction, so a fee that works only inside the App Store collects less each quarter as agents do more of the work.
Catalysts
Two Horizons: What the December Print Decides, and What the Next Three Years Decides
| Timing | Event | What it tests |
|---|---|---|
| Late January 2027 | December-quarter results | Whether 45-46% holds, whether services growth holds 8-9%, and whether inventory keeps building |
| Pending | Supreme Court ruling on the civil contempt standard in Epic | Restores or removes the option of charging on US link-out transactions |
| In force since Oct 1, 2026 | EU 5% Core Technology Commission on out-of-store distribution | Whether Apple can extract 5% from apps that sell outside its store |
| First half 2027 | DRAM and NAND contract prices for the iPhone 18 build | Whether the memory hole widens or narrows, and whether the toll motive survives |
| Unscheduled | Pricing and launch of paid expanded Siri AI access | Whether Apple monetises its own AI first, which would be the lower-risk path |
| Fiscal 2026 Form 10-K | Apple's annual report | Whether the Google search payment and component cost exposure get quantified |
- Short term, days to quarters: memory suppliers are the certain winners of Apple's cost problem, while the January print decides whether Apple has to escalate the services hunt or can coast on price increases.
- Short term: a services-side revenue surprise is the only credible path to offset 140 basis points of product margin compression without hurting units, which is why the third-party app remark matters more than the Siri commentary.
- Long term, one to three years: SK hynix expects demand to exceed supply through 2030, so the memory drag on Apple's bill of materials is more likely to become structural than to fade, keeping the toll option alive.
- Long term: if agents continue to transact on the open web rather than inside apps, the taxable base for any in-app commission shrinks every quarter, which argues Apple monetises its own models and surfaces rather than taxing other people's.
- Long term: the honest counter-case is that memory prices normalise, Apple finishes passing through the cost in premium and storage-tier pricing, and the third-party app remark stays what it was on Nov 2: an option with no price attached.
Conclusion
The Toll Is an Option Apple Signalled, Not a Fee Apple Is Ready to Enforce
Apple's memory problem is upstream and its fix has to be downstream, which is why the third-party AI app remark appeared in the same breath as a 45-46% margin guide. That is a real signal: the search for services revenue is now a board-level priority under a new chief executive, the App Store sits inside Services, and the labs' app economics are visible on Apple's own charts. It is not yet a fee schedule. The instrument that would fill a $1.6 billion quarterly hole with a 30%-era take is precisely the one now under Supreme Court review, and the EU has already been pushed down to a 5% commission on out-of-store distribution.
The more likely near-term path is the one Apple controls: raising premium and storage-tier prices, accepting a smaller hardware margin, and selling expanded access to its own Siri AI, where Apple keeps the entire margin and faces no fee regulators. The third-party app language matters as an opening bid in that negotiation, and as a tell that Apple believes it has leverage over labs while the labs still need iOS. For investors, the tradeable consequences are upstream for now, and defensive for Apple: the memory makers are paid today, and the toll is a 2027 question, not a November one.
What Apple did not disclose on November 2
Take rate on third-party AI app transactions
Not disclosed
CFO cited only \"opportunities across third-party apps and partnerships\"
Fee on AI app subscriptions
Not disclosed
No tier, schedule or exception described
Timing
Not disclosed
No launch date or fiscal period given
Surfaces involved
Not disclosed
Siri, Spotlight, App Store and Apple Intelligence not specified
Existing partnership economics
Not quantified
Google search licensing and Gemini arrangements remain undisclosed
Google search payment
Not broken out
Services is reported as a single line in the earnings release
Where the money moves: memory makers are paid today, the toll is argued over tomorrow
- December-quarter gross margin guided down to 45-46% from 47.9%, with memory costs the main driver and Siri AI ramp expense additive.
- Near term, days to quarters: the January print tests whether 45-46% holds; a miss turns the services-revenue hint into urgency and price increases into the last lever.
- Long term, one to three years: with memory tight through 2030, a services-side toll or paid first-party AI tier becomes structural rather than opportunistic.
- Services gross margin of 76% and 8-9% guided growth leave room to fund AI investment without further margin damage.
- Mobile and client revenue of $13.11B at a 90% gross margin prices the memory squeeze straight into the P&L of Apple's largest handset customer base.
- Fiscal Q1 2027 guidance of $61.5B of revenue at about 86% gross margin extends the cycle through Apple's December build.
- Near term: client and smartphone demand is the swing factor as data-centre demand absorbs wafer capacity; long term, HBM4 supply discipline underwrites the price.
- Q3 2026 operating profit of ₩83.33T on ₩104.62T of revenue puts the operating margin at 79.7%.
- Management expects demand to exceed supply through 2030, keeping handset DRAM contract prices elevated and Apple's bill of materials expensive.
- Near term: HBM4 shipments began in earnest and are expected to surpass HBM3E in fourth-quarter volume; long term, capacity additions are the main risk to the margin.
- Device Solutions operating profit of ₩106.5T on ₩147.1T of revenue concentrates the profit pool in memory.
- Group operating profit of ₩107.44T on ₩195.25T of revenue shows how little the handset business contributes while memory prints records.
- Near term: DRAM and NAND contract prices set Apple's cost curve; the same tightness that helps memory division margins is a second-order drag on device pricing.
- Muse passed ChatGPT for the top free iOS slot in the US with 730,000 downloads in five days, proving labs can own iPhone distribution without Apple.
- Opposing force: a reinstated 30%-era take on AI app subscriptions would hit Meta's $20 and $100 monthly tiers and its own app economics first.
- Q3 2026 revenue of $64.60B at a 79.3% gross margin leaves room to absorb a fee, while operating margin fell to about 20% on AI infrastructure spending.
- Reported near $1 billion a year for Gemini inside Siri gives Alphabet a paid seat inside Apple's install base.
- Opposing force: Apple's own 10-Q says Google search-licensing arrangements are under government investigation and that remedies could materially reduce what Apple earns, and therefore what Alphabet is paid.
- Near term: an AI-app toll would tax Alphabet's own Gemini app; long term, an Apple-owned answer layer can disintermediate paid search on the same device.
