Record Revenue, Flat Earnings: December's Margin Guide Points FY27 Below FY26 as the Stock Sets a New High; Downgrading to Underperform

Results vs. Consensus

This is a quarter where the reported numbers and the guide point in opposite directions, and the guide carries more weight. On the September print itself there is little to argue with: Apple landed above the top of its July revenue range, at the top of its gross margin range, inside its opex range and in line on Services, and it set records for the quarter and the year. The tariff-refund help more than halved from June, to $0.05 of EPS, so the headline is cleaner than the June quarter's. The beat is not where the downgrade comes from.

Q4 FY26 Scorecard

Quality of Beat/Miss

• Revenue: Real and narrow. The $12.0B year-over-year increase came almost entirely from iPhone (+$8.6B) and Mac (+$2.1B), with Services adding $2.8B and iPad and Wearables subtracting $1.4B. Apple absorbed nearly 2 points of sequential FX headwind and supply constraints on three product lines. The iPhone Duo contributed nothing, because it shipped after the quarter closed. Part of the growth is price: Mac and iPad carried increases from the end of June, and iPhone prices rose in September.
• Margins: Better than the July bridge but still lower. Reported gross margin of 47.89% rose 71bps year over year, but about 1 point was tariff refund, so the underlying ~46.9% was roughly 30bps below the year-ago level and about 120bps below June's underlying ~48.1%. The CFO attributed the sequential decline "entirely" to memory. Services gross margin of 75.95% held up. Operating margin fell 52bps to 31.13% because operating expenses grew 20.6%, almost twice revenue growth, on a 32.3% rise in R&D.
• EPS: $2.10 (+13.5%) rests on net income up 11.8% and diluted shares down 1.6%. The year-ago tax rate was 16.3% and this quarter's was 15.0%, worth about $0.03 of the year-over-year growth. Strip the refund and the tax-rate difference and underlying EPS growth was nearer 9%, against 11.7% revenue growth. In other words, operating deleverage had already begun in September, before the December margin step.

Year-over-Year Comparison

Sequential Comparison (vs. Q3 FY26, June quarter)

Fiscal 2026 in Full

The four reported quarters sum to the fiscal-year totals on every line in the table, to the million.

Segment Performance

The product cut and the geographic cut tell the same story from two angles this quarter. By product, iPhone and Mac supplied all of the incremental growth and then some: the two together added $10.6B of year-over-year revenue, against a $12.0B increase for the whole company, while iPad and Wearables shrank by a combined $1.4B. By geography, Greater China grew at twice the company rate and Japan was the only segment to decline. The common thread is that Apple sold every high-end phone and laptop it could build and gave ground almost everywhere the product was mid-cycle or supply-starved.

By Product Category

By Geography

iPhone

iPhone grew 17.5% to $57.58B, a September-quarter record, and ran ahead of the mid-teens rate management guided in July. This is the fourth consecutive quarter of double-digit growth and the slowest of the four, down from 21.7% in June and 23.3% in December. The launch quarter carried two new form factors at once. The iPhone Duo, Apple's first foldable, was introduced at the September event but began shipping after the quarter closed, so it contributed nothing to September revenue. It is supply-constrained "as of today," and the COO declined to say when supply will meet demand. The iPhone 18 Pro and Pro Max carry a new A20 Pro chip, a variable-aperture camera and, by management's account, the biggest battery-life increase in iPhone history. The standard iPhone 18 is absent from the fall lineup, and when asked why, the CEO pointed to the iPhone 17, iPhone Air and 17e as the rest of the range and declined to discuss the roadmap.

"iPhone revenue was $57.6 billion for the September quarter, up 17% from a year ago despite supply constraints and foreign exchange headwinds. We grew strong double digits in the U.S. and Greater China to achieve September quarter revenue records. We also set a September quarter record for upgraders worldwide as well as in the U.S., China Mainland and India, helping to deliver our best fiscal year ever for upgraders globally." — John Ternus, CEO

Two details matter for FY27. First, Apple raised iPhone prices in September, so part of this quarter's growth is price rather than units, and the CFO said it is "really early to really assess any elasticity impact." Second, the supply constraint changed shape. The September constraint came from advanced-node silicon availability. The COO said those constraints are easing as the lineup moves to 2-nanometer nodes, and the December constraint he named covers Mac and iPad on memory, not iPhone. The Duo is the exception and it remains constrained.

Assessment: The iPhone pillar holds, and the Duo gives the cycle a genuine new product to sell into FY27. Our concern is not demand. It is that iPhone revenue now carries a price increase taken to defend margin, so the reported growth rate overstates the volume trend by an amount Apple will not disclose. December's total-company guide of 7% to 10% sits against a December-quarter iPhone comp that grew 23.3% a year ago, which is the hardest base of the cycle. A foldable that cannot be built fast enough helps the narrative more than it helps the December quarter.

Mac

Mac grew 23.7% to $10.79B, its second straight quarter above 20%, on MacBook Neo and MacBook Pro. Apple sold more Macs in fiscal 2026 than in any prior fiscal year, gained global share in the quarter and the year by management's account, and set a September-quarter record for customers new to Mac. The education channel keeps producing large deployments: the Philippines Department of Education will deploy 170,000 MacBook Neos, Apple's biggest education deal ever for Mac. The fall refresh adds touch to the MacBook Pro for the first time alongside M6 in the 14-inch MacBook Pro and iMac.

"Mac revenue was $10.8 billion, up 24% year-over-year despite supply constraints. This performance was driven by the strength of the MacBook Neo and MacBook Pro. We saw double-digit growth in both developed and emerging markets with all-time records in markets like India and Southeast Asia." — Kevan Parekh, CFO

Assessment: Mac is the cleanest growth story in the print, and it is the one most exposed to the cost problem. Mac carries the price increase Apple took at the end of June, it is named as memory-constrained for December, and a Mac carries far more DRAM and storage than an iPhone. Mac demand looks real and institutional. Mac margin is where memory inflation lands hardest, and Apple does not disclose category margins.

Services

Services grew 9.8% to $31.57B, an all-time record in dollars and the first single-digit growth rate in our coverage. It landed in line with the roughly 9.5% the July framework implied. The CFO put the March-to-September foreign exchange drag at about 4 percentage points and said December carries another half point. The more consequential disclosure was about the App Store: its September-quarter performance was "similar to the prior year," mobile gaming headwinds continue into December, and the US link-out ruling and App Store business-model changes in certain countries remain in place. Records came from video, cloud services, payment services and advertising.

"During the September quarter, the performance on the App Store was similar to the prior year, and we had a record number of paid subscriptions with double-digit growth for third-party apps on the App Store. At the same time, we did see some headwinds in mobile gaming that we've talked about before and expect to see those continue into the December quarter." — Kevan Parekh, CFO

December guides Services to 8% to 9%, a further step down, with "third-party licensing" named as a new primary sequential headwind next to FX. Services gross margin of 75.95% rose 33bps sequentially, but December brings a new cost into the line: the cloud inference costs of Siri AI will be recognized in services cost of sales.

Assessment: The Services pillar said mid-teens on roughly 77% margins. The line now grows high single digits, the App Store engine is flat, an unexplained licensing headwind is arriving, and the AI cost line moves into the same segment. FX explains part of the slowdown and will reverse. The rest is structural. This pillar breaks this quarter.

Greater China and the Geographic Cut

Greater China grew 21.0% to $17.54B, a September-quarter record, and finished fiscal 2026 at $82.38B, up 28.0%. iPhone drove it, with record upgraders in China Mainland, and Mac grew very strong double digits. The quarterly trajectory continues to normalize: 37.9% in December, 28.1% in March, 22.4% in June and 21.0% in September, with revenue down 6.8% sequentially. Japan was the outlier, down 1.4% and the only segment management left out of its list of September records. Europe slowed to 11.5% from 22.4% in June.

"Revenue was up 21% during the September quarter, which was a quarterly record. The performance was primarily driven by iPhone, where we set a September quarter record and by Mac, where we saw very strong double-digit growth. And taking a step back, Greater China was up 28% for the full year." — John Ternus, CEO

Assessment: China stays a confirmed pillar. Four straight quarters above 20% is a recovery, not a bounce, and the third-party survey data management cited put iPhone and Mac at the top of the China sales rankings. Asked whether the region can sustain this pace, the CEO answered with product quality rather than a framework, so we model a further slowdown as the base laps. Europe halving its growth rate in one quarter, with no explanation offered, is the bigger surprise in the regional table.

iPad and Wearables

iPad fell 10.4% to $6.23B on supply constraints and the A16 iPad compare. Wearables, Home and Accessories fell 7.9% to $8.30B against the prior-year Watch SE 3 and AirPods Pro 3 launches. Both installed bases reached all-time highs, and over half of iPad and Apple Watch buyers were new to the product. The new iPad mini with OLED and the home lineup (HomeView, Apple TV 4K, HomePod mini) ship into the December quarter.

Assessment: Neither category drives the thesis, but iPad answers one of July's open questions. It was the declining category that took a price increase and then became supply-constrained. It is now down double digits. We cannot separate elasticity from supply in that number, and management says it cannot yet either. iPad is again on the December constraint list, so the test stays open for a second quarter.

Key Topics & Management Commentary

Overall Management Tone: Management was polished on the product cycle and consistent on the cost line, but less expansive than in July. It again named memory as the driver of both the September and December margin steps, then answered two direct questions about long-term memory agreements with the same purchase-obligation figure each time. On the first call of the new leadership, strategic questions went to the CEO and every number went to the CFO or the new COO, which made the call orderly and left the hardest questions (memory price terms, supply magnitude, Siri AI cost, elasticity) unanswered.

1. Memory Is Now "Unprecedented," and Adjusted Gross Margin Has Lost 380bps in Three Quarters

In July the outgoing CEO called memory "a 100-year flood." The incoming CEO used a different word for the same thing and then extended the warning past the guide window again. September carried higher memory costs than June, partly offset by carry-in inventory and cheaper non-memory components. December carries "even higher" costs, now partly offset by the price increases on Mac, iPad and iPhone. Beyond December, memory pricing keeps rising and the inventory benefit fades.

"What we're seeing with memory pricing right now is unprecedented… Now for the December quarter, we expect even higher memory costs, partly offset by a few factors… And looking beyond the December quarter, we see the pricing for memory purchases continuing to increase. As we've said before, we do believe we'll see a decreasing benefit from the carry-in inventory on memory over time." — John Ternus, CEO

The CFO said the 1-point sequential decline in September, after removing tariff refunds from both quarters, was "driven entirely by memory impacts," and that the gross memory impact was larger than 1 point before the offsets. The walk on management's own adjusted basis is now 49.3% (March) to about 48.1% (June) to about 46.9% (September) to a 45.5% December midpoint. That is roughly 380bps in three quarters, with each step attributed mainly to memory, and the December step also carries Siri AI costs.

Assessment: September landed better than the July bridge, about 46.9% against a 46.5% underlying midpoint, and we give management credit for that. But the direction did not change, the offsets now include price increases that customers pay, and the December guide puts adjusted gross margin below the level of any quarter in fiscal 2025. December is Apple's largest revenue quarter, and a 45% to 46% guide in that quarter tells us the FY27 margin base is lower than our July model assumed. This is the single input that moves our FY27 estimate, and it moves it down.

2. A $33 Billion Jump in Purchase Obligations Is the Answer to the Supply-Agreement Question

The July checklist asked whether Apple would sign long-term memory supply agreements. Management was asked again, twice, and twice answered with one number: manufacturing purchase obligations rose $33 billion from June to September. The CEO volunteered it in the gross margin answer, and the COO repeated it to each analyst who asked about memory procurement.

"As we work to improve supply, our manufacturing purchase obligations have increased by $33 billion from June to the September quarter, which you'll see in our annual filing when it comes out tomorrow." — John Ternus, CEO

The 10-K, filed the next day, puts manufacturing purchase obligations at $89.7 billion on September 26, up from $57.0 billion in June and $56.2 billion a year earlier. Two details in the filing point to longer-dated commitments. The portion payable after twelve months rose to $7.0 billion from $0.8 billion in June, and the risk factor on purchase commitments dropped the phrase "typically for periods up to 150 days." Component inventory, a separate balance, rose to $10.45 billion from $7.65 billion in June and $2.12 billion a year earlier. The filing names no supplier, gives no price terms and gives no reason for the increase, and it says Apple "expects these trends to intensify for memory and storage components."

Assessment: A $33 billion one-quarter increase in commitments is large against $59.7B of quarterly cost of sales, and it is the clearest evidence yet that Apple is locking in supply. It does not tell us that Apple locked in price. Management framed the commitments as a way "to improve supply," not to cap cost, and in the same answer said memory pricing keeps rising beyond December. We read this as Apple securing volume at elevated prices, which bounds the shipment risk but not the margin risk. July's upgrade trigger asked for an agreement that "bounds the cost exposure." This does not meet it.

3. A Record December Guide That Implies Roughly Flat Earnings

Apple guided December revenue up 7% to 10% year over year, which "reflects our best view of constrained supply" and would be the best revenue quarter in the company's history. The rest of the guide is less generous. Services grows 8% to 9%. Gross margin is 45% to 46%, against 48.2% in the December quarter a year ago. Operating expenses of $20.4B to $20.8B grow 11% to 13%. OI&E is about $300M and the tax rate about 17%.

"We expect our December quarter total company revenue to grow between 7% and 10% year-over-year, which reflects our best view of constrained supply. We expect this to be the best quarter for revenue in the history of the company… We expect gross margin to be between 45% and 46%. This includes primarily the impact of higher memory costs as well as some costs related to the ramp of Siri AI." — Kevan Parekh, CFO

Assessment: Run the midpoints through the income statement and December EPS lands near $2.89, about 2% above the $2.84 Apple earned a year ago, on revenue up 8.5%. Gross margin compression of roughly 270bps year over year and operating expenses growing faster than revenue absorb almost all of the top-line growth. That is the shape of FY27 if memory does not turn: record revenue and flat earnings. The headline record and the earnings arithmetic point in different directions, and the earnings arithmetic is what the multiple is paid on.

4. Siri AI's Cost Line Arrives: Cloud Inference in Cost of Sales and a New Revenue Deferral

In July, management said AI costs sit in cost of sales as well as operating expenses and called the balance "a bit uncertain." This quarter the CFO specified the mechanics. Siri AI's server-side inference costs will be recognized in services cost of sales, and Apple will defer part of each device's revenue for Siri AI and recognize it over the device's life, as it already does for bundled services such as Siri and iCloud.

"As we continue to roll out Siri AI to more customers, the associated cloud costs will be recognized through services cost of sales. And consistent with how we've always accounted for Mac, Siri and iCloud for the devices we sell that include these bundled services, we defer a portion of the revenue and recognize it over the life of the device. So our hardware deferrals as of the December quarter will reflect the addition of Siri AI. Both of these items will have an impact to the gross margin." — Kevan Parekh, CFO

Asked about the on-device versus cloud split of Siri AI queries, the CFO said on-device processing carries "no real costs," that the server side carries inference costs, and that Apple is in "very, very early stages" of understanding the mix and any further compute requirements.

Assessment: This changes the capital-light AI argument we credited in July. The strategy is still light on capex, and fiscal 2026 capex fell 28.9%. But AI now has a recurring cost that scales with usage and lands in the same gross margin line that memory is already compressing. The revenue deferral also moves part of each hardware sale into later periods. Neither effect was quantified. Both start in the December quarter, so the 45% to 46% guide is the first quarter that includes them, and the first data point arrives in January.

5. Siri AI Ships in Beta, and Monetization Starts at the Usage Limit

Siri AI launched in beta for US English users in September and added French, Japanese, Korean, Portuguese and Spanish in October. Management cited more than 2.5 billion Siri requests a day and described the product as request-based today, built on an "orchestrator-driven architecture" that leaves room for agentic capabilities later. The first monetization detail also appeared. Features that rely on server-side models, including Siri AI and intelligent photo editing, will carry usage limits, and users who need more capacity will be able to buy expanded access.

"There's going to be some usage limits on features that rely on server side models, which include Siri AI and intelligent photo editing and some other things like that. And so anyone who needs greater capacity will be able to purchase expanded access in the future. So kind of to your point, in the beginning, iCloud tiers were largely about storage, but we've been adding more capabilities and features to it." — John Ternus, CEO

Assessment: The product timeline the AI bear point once worried about has resolved. Siri AI is in users' hands in six languages. What has not resolved is the economics: the paid tier has no price or launch date, the cost line has no size, and the call did not mention Siri AI availability in the European Union or China Mainland at all, after July framed both as work in progress. Usage caps that push heavy users toward paid iCloud+ tiers are a sensible design. Until they carry a price, we treat Siri AI as a cost in FY27 and an option thereafter.

6. The Supply Constraint Moves From Silicon to Memory

For three quarters Apple attributed its constraints to advanced-node silicon. The new COO said that constraint is easing as the lineup moves to 2-nanometer nodes. December's constraints fall on Mac and iPad and come from memory, with demand again above Apple's own plan. The iPhone Duo is the exception and remains constrained with no date for balance.

"The Q4 constraints were primarily driven by the availability of the advanced nodes our SoCs are produced on. But as we have transitioned more of our lineup to the latest 2-nanometer technology nodes, our constraints on advanced nodes are easing. And for the December quarter, we expect constraints on Mac and iPad, primarily driven by memory as demand has been higher than what we expected." — Sabih Khan, COO

Assessment: Memory is now both a cost problem and a volume problem. In July we gave the demand-forecast explanation full weight in revenue and none in margin. That split still holds, but the binding input has changed to one Apple does not make and, by the July CEO's account, buys from three suppliers. A constraint that comes from a scarce, rising-price input is harder to resolve than a node transition, and it explains the $33 billion of new commitments better than any other reading.

7. Apple Takes iPhone Price, and Elasticity Is Still Unread

Apple raised iPhone prices in September after raising Mac and iPad prices at the end of June. The CFO described the decision as reluctant and balanced across units, revenue and margin, and said it is too early to read elasticity because iPad and Mac were constrained after their increases and the iPhone increase is recent. Asked about affordability, the CEO pointed to trade-in and upgrade programs.

"We did reluctantly make the decision to change pricing in a number of our products, starting with the iPad and Mac in the June time frame and then iPhone more recently in September. I think what we try to do is evaluate in doing all that, we evaluate the unit revenue and margin. And we're really being thoughtful and deliberate at evaluating kind of the short-term and long-term considerations, and we really believe we're balancing this well." — Kevan Parekh, CFO

Assessment: Pricing now offsets part of the memory cost on every major hardware line. That makes the December revenue guide partly a price guide, and it means the true volume test arrives in the March quarter, after the holiday constraint clears. July's elasticity checkpoint is unresolved for a second quarter. iPad fell 10.4% after its increase, but management attributes the decline to supply and the compare, and the data do not let us separate the causes.

8. Services: A Flat App Store and a New "Third-Party Licensing" Headwind

The CFO's Services walk opened with FX, about 4 points of drag on the year-over-year rate from March to September and another half point into December. The rest of the walk was less reassuring. App Store performance in September was similar to the prior year, mobile gaming headwinds continue, and December's 8% to 9% guide names "third-party licensing" as a primary sequential headwind without saying which agreement it concerns.

"As we move ahead, we expect services growth to be between 8% to 9%, as I mentioned in the prepared remarks, in the December quarter, and this outlook contemplates our best estimates on sequential headwinds, primarily from third-party licensing and the additional 0.5 percentage point of foreign exchange I just mentioned." — Kevan Parekh, CFO

Asked twice whether Services can re-accelerate, management answered with long-term attributes rather than a rate: the installed base, nearly 1 billion average weekly App Store visitors, record paid accounts, emerging markets and AI-linked tiers.

Assessment: On the CFO's figure, FX explains about 4 points of the 6.5-point slowdown since March (16.3% to 9.8%), which is the bull case's best number. The bear case has three better ones: a flat App Store, a licensing line that the guide treats as a primary headwind, and an inference cost that starts landing in services cost of sales. A segment at 76% gross margin growing 8% to 9% no longer offsets hardware compression. It slows it at best.

9. The New CEO's First Call: Orderly, AI-Forward, and No Numbers

John Ternus opened his first call as CEO with an extended product tour and an AI framing he will likely repeat: iPhone as "the intelligent personal hub," Mac "having an AI moment," and Apple devices as "the best platforms in the industry for AI." He introduced Sabih Khan, the COO, as a regular participant on future calls. On strategy he was direct but brief. He did not discuss the roadmap, said Apple has no fundamentally different plans for macOS and iOS, and said agentic features must protect privacy and security. Capital allocation and supply questions went to the CFO and the COO.

"It's primarily request-based today. But as we talked about at WWDC, it's built on a modern orchestrator-driven architecture that gives us real room to expand its capabilities over time. As far as agentic capabilities, we think these are promising tools for users. We just want to make sure that users are aware of what they're giving access to." — John Ternus, CEO

Assessment: The transition risk stays contained. The handoff was orderly, the CFO is unchanged, and adding the operations chief to the call puts the person closest to the memory problem in front of investors. July's checklist asked whether the new CEO would offer substantive strategy. He offered a clear AI posture and no financial framing, which is a reasonable first call and not yet an answer.

10. Operating Expense Growth Slows to 11% to 13%, and R&D Was Flat Sequentially

Operating expenses rose 20.6% year over year to $19.19B, in the lower half of the $19.1B to $19.4B guide. R&D of $11,730M was essentially flat with June's $11,729M after growing 32.5% over the first nine months. The December guide of $20.4B to $20.8B implies 11% to 13% growth, down from 21.5% for fiscal 2026. The CFO attributed it to continued AI investment and said the growth rate is "coming down" from fiscal 2026 levels, and the CEO added that Apple is "spending what we need to spend to achieve our road map."

"For the December quarter, we do expect OpEx to grow between 11% and 13% year-over-year, which is driven by increased and continued investment in R&D, in particular, AI. This growth rate is coming down, as you alluded, from the levels of growth that we had in FY '26. But as we've said before, we continue to significantly increase our overall investments in AI while continuing to invest in the product road map." — Kevan Parekh, CFO

Assessment: Slower opex growth is a genuine offset, but even 11% to 13% still runs ahead of the 7% to 10% revenue guide, so operating leverage is negative in December. Some AI spend is also moving out of opex: inference costs now land in services cost of sales. Part of the "moderation" is a reclassification of where AI cost appears rather than a reduction in it.

11. Capital Return Now Equals Operating Cash Flow, and the Inventory Buffer Keeps Building

Apple returned about $28B in the quarter, $24.3B of repurchases on the cash flow statement and $3.9B of dividends, against $28.1B of operating cash flow and $25.9B of free cash flow. September-quarter operating cash flow fell 5.5% from $29.7B a year ago even as net income rose 11.8%, because working capital absorbed cash: inventory rose another $3.45B to $14.54B, two and a half times the $5.72B at fiscal 2025 year-end. Cash and marketable securities fell to $136.9B from $146.5B in June, and total debt fell to $80.1B from $84.3B, so net cash declined about $5.4B in the quarter.

"We ended the quarter with $137 billion in cash and marketable securities and $80 billion in total debt. During the quarter, we returned $28 billion to shareholders. This included $3.9 billion in dividends and equivalents and $24 billion in share repurchases." — Kevan Parekh, CFO

Asked what retiring the net-cash-neutral target means under the new leadership, the CFO said Apple will evaluate cash and debt "independently" for flexibility and keeps its philosophy of investing in the business and returning excess cash.

Assessment: The buyback pillar holds. Fiscal 2026 free cash flow of $136.1B rose 37.8%, and diluted shares fell 1.9% for the year. But two details inside the cash flow statement moved the wrong way. The carry-in inventory that management expects to fade did not draw down. It grew by a further $3.45B, much of it bought while memory prices rose, and that cost will flow through FY27 cost of sales. And with the net-cash-neutral target retired, returns at or above operating cash flow now reduce net cash every quarter.

Guidance & Outlook

The December guide splits cleanly into a revenue line above the Street and a profit line below it. Revenue of +7% to +10% on a $143.76B base implies $153.8B to $158.1B, and even the bottom of the range clears the pre-print consensus. The CFO tied it to "our best view of constrained supply," which means the range is a shipping estimate, and he reminded the call that it sits on top of a December quarter that grew 16% a year ago. Gross margin of 45% to 46% is the problem. It includes "primarily" higher memory costs plus Siri AI ramp costs, no tariff refund, and it compares with 48.2% a year ago and a consensus of roughly 46.3% to 46.9%. Services at 8% to 9% sits below the roughly 10.7% the Street expected, with third-party licensing and another half point of FX named as the sequential headwinds.

Street at. Before the print, consensus had December revenue near $152.5B to $153.6B, gross margin near 46.3% to 46.9%, Services near 10.7% and EPS near $2.89 to $2.90. FY27 consensus EPS stood at $9.54 to $9.60 and was little changed after the print. Our new FY27 range of $8.20 to $9.10 sits roughly 5% to 15% below that, and our high end is close to the lowest individual estimate we found. The gap is the gross margin path. A $9.58 FY27 requires the March through September quarters to grow EPS about 9% on a base that includes a 49.3% March-quarter margin and $0.16 of tariff refunds. With December guided at 45% to 46% and memory pricing rising beyond it, that requires a margin recovery management did not describe.

Guidance style. Apple guides one quarter at a time and has met or beaten its revenue range for five consecutive quarters, including this one. The revenue guide is probably beatable again if supply improves. The margin guide has been accurate rather than conservative: September landed 40bps above the underlying midpoint, and the step-down management guided in July happened. We treat the 45% to 46% as the honest number and the 7% to 10% as the one with upside.

Analyst Q&A Highlights

Why the Standard iPhone 18 Is Missing From the Fall Lineup

The first question of the call asked why the standard iPhone 18 did not launch in September, and whether component supply or the Duo launch drove the decision. It is a fair question in a quarter when memory is scarce: a cheaper model with high unit volume is exactly the product a memory-constrained company might hold back. The answer listed the iPhone 17, iPhone Air and 17e as the rest of the holiday range and declined to discuss the roadmap.

Q: "First, on the iPhone 18, the standard model was left out of the September lineup. I was just wondering if you could share some of the reasons why, whether there were some supply constraints and component-related things that led to that decision or if it had to do with the launch of the Duo?"
— Michael Ng, Goldman Sachs

A: "Look, we were thrilled to introduce the iPhone Duo and 18 Pro earlier this year in September. And we have iPhone 17, iPhone Air and 17e continuing to kind of round out this incredibly strong lineup for the holidays. So obviously, we don't make any comments about future road map, but we're excited about where we're at."
— John Ternus, CEO

Assessment: Apple never discusses unreleased products, so the non-answer is policy. The consequence for the model is still real. The fall lineup skews the new-model mix toward the Pro, Pro Max and Duo, which supports ASP and revenue growth in December but asks a prior-generation phone to carry the volume tier. If the standard model was held back to save memory for higher-margin devices, that is a sensible allocation that also flatters December's price-mix and defers volume into a later quarter.

Walking the December Gross-Margin Step

One question put a number on the December step: excluding tariff refunds, the 45% to 46% guide implies roughly 140bps of sequential decline. The analyst asked management to split it between memory, mix and the Siri AI ramp. Management accepted the magnitude, gave memory "most" of it, and then explained the two Siri AI mechanisms rather than sizing them.

Q: "I think if I exclude all the tariff refunds, you start of implying gross margins will be down about 140 basis points sequentially at least. Can you just separate how much of this is maybe commodity memory inflation versus mix versus -- you called out Siri AI ramp costs as well."
— Amit Daryanani, Evercore ISI

A: "So as you alluded for the December quarter, we talked about a gross margin outlook of between 45% and 46%. As you said, when you remove the benefit of the tariff refunds in September, the midpoint of the range is a decline sequentially in the magnitude you mentioned. We expect most of that decline to be driven by higher memory costs."
— Kevan Parekh, CFO

Assessment: This confirms the 140bp step and the cause, which is the core of the downgrade. "Most" memory leaves the remainder to Siri AI cloud costs and the new deferral, with no split. In July the CFO attributed more than the whole step to memory. This quarter a second, AI-driven cost joins it, and that cost scales with usage rather than with commodity prices.

Whether Apple Is Signing Long-Term Memory Agreements

A recurring line of questioning asked whether Apple has changed how it buys memory now that it competes with well-capitalized AI buyers signing long-term agreements, and whether any agreement secures availability even if price stays a risk. Both times the operations chief answered with the $33 billion increase in manufacturing purchase obligations and did not describe terms, tenor or price.

Q: "Clearly, you're now competing for memory more so versus historically from pretty well capitalized companies who are signing long-term agreements. So my question is, is Apple changing anything in the way that you are procuring memory? How do you think about the progression of memory over time?"
— Multiple analysts incl. Wamsi Mohan, Bank of America; Krish Sankar, TD Cowen

A: "So as you know, we have worked with many of our suppliers for decades, and we deeply value their partnership over the years. We know our products would simply not be possible without them, and we look forward to continuing to work together. We're always working hard to improve supply. And as John said earlier, we -- our manufacturing purchase obligations have increased by $33 billion from the June to the September quarter."
— Sabih Khan, COO

Assessment: The same answer, three times on one call, is a prepared position. It discloses that Apple committed a very large amount of capital to future supply and withholds whether that capital buys a price cap. We read the omission as informative. A company that had fixed its memory cost would say so in a quarter when the stock trades on that cost. July's question is answered on volume and still open on price.

How Large the Gross Memory Hit Is Before Offsets

A follow-up asked whether memory inflation is costing one to two points of margin. The CFO confirmed a 1-point sequential decline in September after tariff refunds and added a detail he had not given before: the gross memory impact was larger than the reported decline, and carry-in inventory, cheaper non-memory parts and pricing offset part of it.

Q: "I understand the memory price inflation. Kevan, it seems like it's probably hike 1 to 1.5 or 2 percentage points. A, is that true?"
— Krish Sankar, TD Cowen

A: "We saw about a 1 point sequential change in margin, and that was driven entirely by memory impacts. Now we did have some offsets because the gross memory impact was larger than that, but we had some offsets as John talked about earlier, driven by the benefit of carrying inventory, some offsets from nonmemory costs as well as some of the benefits of pricing."
— Kevan Parekh, CFO

Assessment: This is the most important sentence for FY27. The net decline understates the gross pressure, and two of the three offsets are temporary. Carry-in inventory fades by management's own account, and pricing is a one-time step until Apple raises prices again. If the gross impact keeps rising while the offsets shrink, the net impact accelerates even if memory prices merely stay high.

iPhone Duo Demand and When Supply Catches Up

Questioning on the Duo asked what Apple learned from the launch, what supply-chain issues remain, and when supply and demand will balance. A follow-up asked management to quantify the constraints for September and December. The COO said the Duo is constrained and the ramp is going well. He did not date the balance or quantify any constraint.

Q: "Maybe on the Duo, can I just ask in terms of maybe your thoughts in terms of our learnings from the launch of the Duo and the demand that you've seen on that front? And what are the sort of supply chain-related issues that you're facing right now that you probably need to iron out if you're trying to meet demand on that front?"
— Samik Chatterjee, JPMorgan

A: "So we're thrilled with the initial response on the iPhone Duo. iPhone Duo is constrained as of today. The ramp is going well, and we're working hard to get this revolutionary new product into the hands of our customers as quickly as possible. But at this time, it's really hard to predict when supply and demand will be in balance."
— Sabih Khan, COO

Assessment: Constrained demand for a new form factor is a good problem, and the bull case will lean on it. It is also the fourth consecutive quarter in which Apple cited supply constraints and declined to size them. We cannot tell how much December revenue the Duo constraint holds back, and so we cannot tell whether the 7% to 10% guide is conservative or simply realistic.

Price Versus Units in the December Guide

One question asked management to separate price from quantity in the December outlook, given industry-wide PC price inflation and Apple's own increases. The CFO declined to give the split, noted the strong year-ago compare, and repeated that elasticity is too early to read.

Q: "Can you help us understand sort of the difference or maybe the delta of quantity versus price across sort of the product lineup? Because it certainly looks like you're looking at double-digit growth across the iPhone and the Mac business. And we all know, I think, in the industry that there's been pretty strong price inflation across the PC Mac ecosystem, and you guys mentioned that you're reluctantly taking price as well."
— David Vogt, UBS

A: "We don't normally provide that level of detail or color as we look at the outlook other than to say, again, we did guide to the 7% to 10% year-on-year. And again, just as a reminder, that is on top of a phenomenal December quarter last year where revenue grew 16% -- and so we don't provide the level of detail, but I would say it's early to really read price elasticity at this point in time."
— Kevan Parekh, CFO

Assessment: Apple does not disclose units, so the refusal is expected. The analytical problem is that every hardware category now carries a price increase. A 7% to 10% revenue guide on top of price increases implies low unit growth or a decline, and constraints alone do not explain the difference. We assume price contributes a meaningful share of December's growth.

Whether AI Agents Threaten the Services Business

A question to the new CEO asked what role Apple wants Siri AI to play among consumer AI agents, and how he views the risk of agents encroaching on Services. The answer covered Siri AI's architecture and privacy safeguards. It did not address the encroachment risk.

Q: "I'd just love to get your perspective on what role you want Apple and Siri AI to play in kind of this consumer-agentic AI landscape? And just how you look at any potential risks of some of these agents trying to encroach on your services business?"
— Erik Woodring, Morgan Stanley

A: "Siri has long been popular with our users with more than 2.5 billion requests every day. And with Siri AI, we really have the kind of the Siri that we always wanted. It's primarily request-based today… As far as agentic capabilities, we think these are promising tools for users. We just want to make sure that users are aware of what they're giving access to. We want to make sure this is done in a way that protects security and privacy."
— John Ternus, CEO

Assessment: In July the same analyst asked whether AI is diverting time from the App Store and got no answer. This quarter the App Store was flat year over year and the question got no answer again. Privacy is Apple's real differentiator, but it does not explain how Services grows if agents complete tasks that once ran through paid apps. Two quarters of silence on the question raise its weight in our bear case.

Siri AI's Compute Mix and What Faster Adoption Would Cost

The last AI question asked how Siri AI's token consumption splits between the device, Apple's private cloud and third-party cloud, whether usage is accelerating, and whether faster global adoption would change Apple's compute plans. The CFO said the mix is too early to predict, on-device queries cost almost nothing, and server-side inference costs flow to services cost of sales.

Q: "In the early days of launching this, can you give us some sense of how the token consumption associated with Siri AI is being done on-device versus private cloud compute versus GCP? And have you started to see any acceleration of usage? And if it does accelerate really fast globally, would it change your outlook on what you need to spend for compute?"
— Wamsi Mohan, Bank of America

A: "The mix of that is still very early days, and so it's hard to predict exactly what the mix of those costs are going to look like. Obviously, on device -- it's very, very efficient. No real costs associated with that. But on the server side, we would have some inference costs, and that's really the connection of the Siri AI-related costs that I mentioned that will be recognized as services cost of sales."
— Kevan Parekh, CFO

Assessment: The answer named only on-device and private cloud compute and did not address the third-party cloud the question raised, or whether usage is accelerating. A cost that is "hard to predict" and already in the December margin guide is the kind of variable that widens the FY27 estimate range. Our FY27 range now carries a Siri AI cost we cannot size, and the risk to it is one-directional.

What They're NOT Saying

Each item below was checked against the release, the call and the Form 10-K filed on November 3 before inclusion.

1. The price terms behind $33 billion of new commitments: Manufacturing purchase obligations rose to $89.7B from $57.0B in one quarter, and the share payable after twelve months rose to $7.0B from $0.8B. Neither the call nor the 10-K names a supplier, a component, a tenor or a price mechanism, and the filing gives no reason for the increase. Asked twice whether Apple is signing long-term memory agreements, management repeated the figure. Whether these commitments cap Apple's memory cost or simply lock in volume at today's prices is the most important unanswered question for FY27 gross margin.
2. Who licensed Apple $13.5 billion of intangible assets: The 10-K's new intangible-asset note shows gross licensed intangibles rising to $29.6B from $16.1B in fiscal 2026, most of it recorded by March, with other non-current liabilities up $12.8B over the same year, mostly in the December quarter. The filing names no licensor, asset or acquisition, and the call did not mention it. Scheduled amortization rises to $6.6B in fiscal 2027 from $4.9B in fiscal 2026. That is roughly $1.7B of additional annual expense, and the filing does not say where in the income statement it lands. We flagged this balance in July. It is now partly explained (licensed, not acquired) and still unattributed.
3. The cost of Siri AI: Management said cloud inference costs will hit services cost of sales and that a new Siri AI deferral will reduce hardware revenue recognized at sale. Neither was sized on the call. The 10-K adds Siri AI to the bundled-services performance obligation but discloses no deferred amount, and the paid "expanded access" tier has no price or date.
4. Which "third-party licensing" arrangement drives the Services slowdown: The December guide names it as a primary sequential headwind. The 10-K separately describes Apple's search licensing arrangements with Google and others as subject to court-ordered remedies, appeals and renewal risk. Management did not say whether that is the headwind, how large it is, or whether it recurs.
5. The size of the supply constraint: This is the fourth consecutive quarter in which Apple cited supply constraints as a cap on revenue and declined to quantify them. Asked directly to size September and December, management described which products are constrained and why, but not by how much. The Duo's date for supply balance was also left open.
6. Units versus price: Every major hardware line now carries a price increase taken since June. Asked to separate price from quantity in the December guide, the CFO declined. Without that split, the 7% to 10% revenue guide cannot be read as a demand signal.
7. A gross-margin floor beyond December: For the third consecutive call, management bridged the next quarter's margin and declined to bracket the period after it. This quarter's 10-K adds that Apple "expects these trends to intensify for memory and storage components," which says the direction and leaves out the size.
8. Siri AI in the European Union and China Mainland: In July, management called both regions work in progress. This call did not mention either region in connection with Siri AI. The 10-K repeats that EU Digital Markets Act requirements may cause Apple "to not launch or maintain products, services and features, such as Siri AI, in certain jurisdictions." Europe and Greater China produced $49.5B of September-quarter revenue, 43% of the total.
9. Why Europe halved and Japan declined: Europe slowed to 11.5% from 22.4% in June, and Japan fell 1.4% to become the only declining segment. Neither was discussed on the call. The 10-K's full-year discussion cites yen weakness for Japan, but it does not address the September quarter.

Market Reaction

• Pre-print setup: AAPL closed at $338.32 on November 2, 2026 entering the print, inside a 52-week closing range of $246.63 to $348.91. The stock was +24.4% YTD (from $271.86 at 2025 year-end), +25.1% over the trailing twelve months and +1.4% over the trailing 30 days (from $333.69), against an S&P 500 up 10.2% YTD. It had risen 9.5% since the $308.91 close that followed our July downgrade. Options priced a move of roughly 4% to 4.8%.
• After-hours move (November 2): Roughly 3% lower after the release, with one print at $328.85 (−2.8%), after trading as much as about 4% lower during the call.
• Reaction session (November 3): The stock opened at $325.58, a 3.8% gap down, traded a range of $325.33 to $334.50, and closed at $331.90, down 1.9% ($6.42). Volume was 63.2M against a 38.7M 30-day average, 1.6x normal. The S&P 500 rose 0.4%, so AAPL lagged the index by about 2.3 percentage points. Memory names rose the same day: Micron +2.6%, SanDisk +4.8% and Western Digital +2.6%.
• Follow-through session (November 4): AAPL rose 5.3% to $349.51 on 88.8M shares, a new 52-week closing high above the prior $348.91. The S&P 500 rose 1.6%. Memory names reversed hard: Micron −7.0%, SanDisk −7.9% and Western Digital −6.8%. Coverage tied both moves to the midterm election results and the expectation that a slower US data-center build-out eases memory pricing.

The reaction session itself was orderly. The gap at the open matched the after-hours read and sat inside the options-implied range, and the stock recovered half of it by the close as buyers weighed the revenue beat and a record December revenue guide against the margin guide. That shape, a gap that half-fills on 1.6x volume, says the market saw the quarter as mixed rather than broken, which is also our read of the September print itself.

The November 4 session matters more. The stock gained more in one day on a macro read-through to memory costs than it lost on the guide. Coverage framed Apple as a beneficiary of slower AI infrastructure spending: a large memory buyer less exposed to the build-out than its mega-cap peers. Memory stocks fell 7% to 8% on the same logic. In effect, the market priced an earlier memory peak than Apple's own management described one day earlier.

We think that trade runs ahead of the evidence for the period that sets FY27 earnings. Apple's guidance, issued the day before the election, says the prices it pays for memory keep rising after December, and its 10-K, filed on Election Day, says it expects the trend "to intensify." Apple has also committed to $82.7B of manufacturing purchases payable within twelve months, at terms it did not disclose. A slower data-center build-out would help memory pricing with a lag measured in quarters. It does not change the December guide, and it changes FY27 only if contract prices turn early in the fiscal year. It is the main risk to our call, and the first test is the March-quarter gross margin guide in January.

Street Perspective

Debate: Is FY27 an earnings-decline year, or does memory peak in time?

Bull view: The constructive camp treats memory as the most cyclical input in electronics and argues the cycle is already turning: AI infrastructure spending faces political and power constraints, price increases on every hardware line offset part of the cost, and Apple's scale secures supply that smaller rivals cannot. Consensus FY27 EPS near $9.6 reflects that view, implying growth of about 7% on FY26's $8.98 as margins recover through the year.

Bear view: The skeptics point out that management's own December guide pairs 8.5% revenue growth with roughly 2% EPS growth, that management says memory purchase prices keep rising after December, and that the next three quarters lap a 49.3% March margin and $0.16 of tariff refunds. Several desks cut FY27 and FY28 estimates on lower gross margin and higher R&D, and the 10-K itself warns that price increases "may also reduce demand."

Our take: The bears have the arithmetic. A December guide at 45% to 46%, in Apple's largest and highest-leverage quarter, with the cost driver still rising, does not support a margin recovery large enough to deliver $9.6 in FY27. We model $8.20 to $9.10 and expect consensus to move toward us as the March guide arrives.

Debate: Is the iPhone Duo a supercycle or a constrained halo?

Bull view: The bull case sees the Duo as the first new iPhone form factor in years, with first-day preorders slipping into 2027 ship dates and some estimates calling for 10 million units over the next few quarters and 20 to 25 million in 2027 as supply ramps. A foldable at a premium price lifts ASP and mix for several years and gives the installed base a reason to upgrade that is not tied to AI.

Bear view: The bear camp notes that the Duo shipped no units in the September quarter, is constrained with no date for supply balance, and arrives in a lineup that dropped the standard iPhone 18. Even the bullish unit estimates are small against Apple's annual iPhone volume, and a device with a larger display and more memory content is not obviously margin-accretive in a memory shortage.

Our take: The Duo is a genuine revenue and mix tailwind for FY27, and it supports our view that iPhone demand is the healthy part of the thesis. It does not address the cost line, and its constraint makes the December quarter harder to beat by much. We credit it in revenue and not in margin.

Debate: Does "defensive quality" justify the multiple?

Bull view: The optimistic camp argues Apple deserves a premium as the mega-cap least exposed to AI capex risk: fiscal 2026 free cash flow of $136B rose 38% while capex fell, the balance sheet still carries net cash, higher regulatory barriers favor incumbents, and the stock behaves as a defensive asset in a volatile tape. On this reading 35x-plus forward earnings is the price of safety.

Bear view: The bear case is that the multiple is now the highest in recent memory on next-twelve-month estimates, and that the estimates themselves are falling. Some desks call the stock extremely overvalued on slowing revenue growth. A defensive premium is hard to defend for a company guiding gross margin down and Services to high single digits.

Our take: Quality deserves a premium and we pay one: our 30x to 32x is the same premium multiple we applied in July. It does not deserve 40x a year in which earnings likely decline. The free cash flow is real, but it is already being returned in full, so it supports the share count rather than the multiple. At $349.51 the stock prices a memory recovery that management has not described.

Model Update & Valuation Framework

How we get to FY27. The December guide is the anchor. At the midpoints it produces about $2.89 of EPS, roughly 2% above last year's $2.84, on 8.5% revenue growth. The next three quarters lap a different base: March FY26 printed a 49.3% gross margin with no refund, and June and September FY26 carried $0.11 and $0.05 of refunds. If full-year gross margin runs at 44.5% to 46.0%, bracketing the 45% to 46% December guide, revenue grows 6% to 9%, and operating expenses grow 10% to 12%, FY27 EPS lands at $8.20 to $9.10, against $8.98 reported for FY26 and roughly $8.82 excluding refunds. The midpoint, about $8.65, is a decline. At that midpoint the March, June and September quarters together earn about $5.76, below the $6.13 they earned in FY26. At the $9.10 high end they earn about $6.21, slightly above it. Our high end requires gross margin to hold 46%, the top of the December guide, for all four quarters while memory costs rise.

Valuation impact. We anchor to the November 4 close of $349.51, the last completed session before publication and a new 52-week closing high. On a like-for-like basis the multiple expanded again. On FY26 earnings the stock trades at 38.9x the $8.98 reported, against 34.9x our ~$8.86 FY26 estimate at the $308.91 close after July's print. On FY27 it trades at 40.4x our new ~$8.65 midpoint, and at 36.8x even on July's higher $9.50 midpoint, against 32.5x on that same estimate in July. The stock has risen 13.1% since our downgrade while our FY27 estimate fell about 9%. We keep the 30x to 32x multiple we applied in July, but we apply it differently. July applied 30x and 32x to the midpoint of its EPS range. This note applies 30x to the low end of the FY27 range and 32x to the high end, which frames a 12-month value of $245 to $290, 17% to 30% below the current price. On July's method, the range would be about $260 to $277, 21% to 26% below. The wider construction carries the full spread of our FY27 estimate into the value range, and its $290 top is also our price trigger. Even the bull case, which needs memory to peak in the first half and a 36x multiple, lands slightly below today's price. The risk-reward has inverted: the upside case returns roughly nothing and the base case loses about a fifth.

Thesis Scorecard Post-Earnings

July's Checklist, Graded

The Q3 FY26 recap closed with ten commitments to watch in the September print. Management delivered on every number it guided and on none of the questions it declined to answer in July.

July's Re-Engagement Triggers, Graded

July named three developments that would bring us back toward a more constructive rating. We grade them as written.

Standing Thesis

Overall: The thesis weakened further, and this time the weakening is in the forward earnings line rather than the reported quarter. September was a clean print. It beat the top of the revenue guide, landed gross margin at the top of its range, and closed Apple's best fiscal year with revenue of $478.8B and EPS of $8.98. None of that is in dispute. What changed is the shape of FY27. December's guide pairs a record revenue quarter with roughly flat EPS, because gross margin falls about 270bps year over year and operating expenses grow faster than revenue. Management told us memory costs keep rising after December, the inventory buffer fades, Siri AI adds a cost line, and Services slows to 8% to 9%. Carry those inputs forward and FY27 EPS lands at $8.20 to $9.10, a likely decline from FY26. The demand side of the thesis is intact: iPhone, Mac and Greater China all ran above 17% and the Duo is constrained on demand. The earnings side is not, and the price moved the opposite way.

Action: Downgrade to Underperform from Hold. This is a valuation call on a lower earnings path, not a call that the franchise is impaired. At $349.51 the stock trades at about 40x our FY27 midpoint for a year in which EPS likely declines, and our 12-month value range of $245 to $290 sits 17% to 30% below the price. The bull case lands near flat. We would move back to Hold on any of four developments, each outside our current base case: (1) a March-quarter gross margin guide at or above 46.5%; (2) disclosure that Apple's new memory commitments carry fixed or capped pricing; (3) Services guided back to 10% or higher; or (4) a price at or below $290, the top of our value range. If none of the first three occurs, only a price at or below $290 changes the rating. These four replace July's three triggers, and the changes are these. July's memory trigger carries forward as (2). July's Services trigger asked for evidence in the December print; (3) tests the guide instead, so a Services guide of 10% or more in January would count even if the December print falls short. July's third trigger, a December guide showing constrained demand converting, does not carry forward. Triggers (1) and (4) are new. We expect AAPL to trail the S&P 500 over the next twelve months.