FY2026: Records on Revenue, the Test Moves to Cost

Thesis status: No condition weakening, and leadership continuity passes its test. For a second quarter, the cost that moved the margin is one no condition scores.

Key Takeaways

1. A record year, capped by supply. Apple ($AAPL) closed FY2026 with revenue of $478.8B, up 15.1%, on iPhone growth of 21.2% and Greater China growth of 28.0%. Free cash flow rose 37.8% to $136.1B, and $8.8B of that rise is a lower cash tax bill against a FY2025 base inflated by the Ireland State Aid settlement. Supply constrained iPhone, Mac and iPad in the fourth quarter. Management guides December revenue growth of 7% to 10%, a range that "reflects our best view of constrained supply".
2. The margin turned underneath the records. Gross margin before tariff refunds has fallen in each quarter since March, from 49.3% to 48.1% to about 47%, and management guides 45% to 46% for December. Management attributes the decline mostly to memory, with some Siri AI ramp cost in December, and the 10-K states that Apple expects memory and storage cost trends to intensify. On Memo No. 1's method, the fall from March to the guide midpoint is worth about $15.0B a year after tax, nine-tenths of the primer's case for losing the Google search payment in full. Manufacturing purchase obligations rose to $89.7B from $57.0B in one quarter.
3. Capital-light, on more than one line. Capex fell to 1.89% of revenue, from 3.06% in FY2025. Licensed intangible assets rose $13.5B gross over the year, 1.5 times capex, and the 10-K does not say what was licensed. Management stated that Siri AI's cloud costs are recognised in services cost of sales as it rolls out, and the December margin guide includes its ramp costs.
4. What the price asks. At $349.51 on 4 November the price implies free cash flow growth of 13.88% a year for a decade, against 11.11% at Memo No. 1. The price accounts for 1.66 points of the rise and the discount rate for 1.17, more than half of that a change in the model's basis. On the four-driver model the price requires a flat EBIT margin of 59.4% from FY2027 to FY2035, against 33.0% in FY2026. Against seven peers, Apple trades above all of them on operating income and below Microsoft and Meta on free cash flow, where capex takes most of their operating cash flow.
5. Seven conditions, none weakening, and two corrections. Two conditions are affirmed, three developing and one dormant, and leadership continuity passes its test on the first results call under John Ternus as chief executive. Two corrections to the published record: Apple dropped net cash neutral as a formal target on 30 April, seven weeks before the primer, and the Google search remedies the primer dated to a 2H 2027 hearing were ordered in September 2025 and are under appeal by both sides.

This is the second update to the June primer, after Memo No. 1 on the June quarter. The primer asks whether Apple's capital-light model can deliver the earnings its valuation requires, over FY2026 to FY2028. This memo covers the fourth quarter and the fiscal year to 26 September 2026, from the 8-K and earnings call of 2 November and the 10-K filed on 3 November, and re-strikes the reverse DCF at $349.51 on 4 November.

FY2026 Results

The September quarter grew more slowly than the three before it, and its operating income grew more slowly than its revenue (Exhibit 2). Revenue rose 11.7% to $114.5B, a September-quarter record, on iPhone up 17.5% and Mac up 23.7%, with iPad and Wearables lower. Management stated that supply constrained iPhone, Mac and iPad. Operating expenses rose 20.6%, with research and development up 32.3%, so operating margin slipped to 31.1% from 31.6% a year earlier.

iPhone supplied $44.5B of the year's $62.7B revenue increase, seven-tenths of it, and the 10-K attributes the growth primarily to Pro models (Exhibit 3). By geography, Greater China added $18.0B, growing 28.0% with a segment operating margin of 42.9% against 41.8%. The 10-K names renminbi strength as a favourable factor without sizing it. Services added $14.1B, but its growth slowed from 16% in the March quarter to 12.1% in June and 9.8% in September.

The year's gross margin gain came from products. Gross margin rose to 48.8% from 46.9%, and the 10-K attributes the products increase to "product mix and tariff refunds, partially offset by higher costs, including memory." Services gross margin rose 0.8 points to 76.2% on a different mix of services. Operating margin rose by about half as much as gross margin, to 33.0%, because operating expenses grew 21.5%.

Free cash flow grew 37.8%, two and a half times as fast as revenue, and Exhibit 4 lists what flatters it and what holds it back. Cash taxes paid fell $8.8B, because FY2025's bill carried the Ireland State Aid settlement. Accounts payable added $12.2B to operating cash flow against $0.9B a year earlier, more than the $8.9B that went into inventory. The inventory build is components, which rose to $10.4B from $2.1B while finished goods rose $0.5B. In the fourth quarter alone free cash flow fell 2.4%, in a quarter when inventories rose a further $3.4B. Tariff refunds flatter both margin and cash, and Apple discloses no full-year total, so the benefit cannot be sized for the year.

Outlook: The December Guide

Management guides the December quarter below September on both growth and margin, and memory sits behind both (Exhibit 5). Revenue is guided up 7% to 10% against a year-earlier quarter that grew 16%, a range that reflects "our best view of constrained supply". The constraint has changed cause. In September advanced nodes limited iPhone, Mac and iPad, and the move to 2-nanometre nodes is easing that. For December the chief operating officer expects constraints on Mac and iPad "primarily driven by memory".

The gross margin guide of 45% to 46% includes "primarily the impact of higher memory costs, as well as some costs related to the ramp of Siri AI". Price increases on Mac and iPad in June and on iPhone in September offset part of the memory cost, with carry-in inventory and lower costs on other components. Management stated that it is too early to read price elasticity. Services growth of 8% to 9% carries about half a point more currency headwind, after about 4 points between March and September, and a third-party licensing headwind management does not size. Operating expenses are guided up 11% to 13%, against 21.5% in FY2026, with AI research still the main driver.

The Memory Pass-Through

The margin before tariff refunds has now fallen for three quarters running, and on management's account memory explains most or all of each step (Exhibit 6). From March to June it fell 120 bps, and the chief financial officer attributed "more than 100%" of the change to memory. From June to September it fell about a point, "driven entirely by memory impacts", with offsets covering part of a larger gross memory cost. The December midpoint is about 140 bps lower again, mostly on memory. September came in at about 47%, half a point above the 46.5% implied by the July guide.

Management's outlook puts no ceiling on the cost. The chief executive described memory pricing as "unprecedented" and expects purchase prices to keep rising beyond December, while the benefit from carry-in inventory decreases over time. The 10-K states that Apple expects supply constraints and rising costs "to intensify for memory and storage components", and that price increases "may not effectively mitigate these negative impacts, and may also reduce demand".

Sized on Memo No. 1's method, the fall from March to the December midpoint is close to the primer's case for losing the Google search payment in full (Exhibit 7). At 380 bps on FY2026 revenue it is $18.2B before tax and $15.0B after, against $16.5B for the Google case and $4.9B for services margin falling to 70%. At Memo No. 1 the same measure was 280 bps and $10.8B. The comparison holds for magnitude only. The primer's cases remove revenue permanently, and memory pricing has been cyclical. None of the seven conditions reads a cost line, so this variable carries no status.

Apple is committing further ahead to secure supply (Exhibit 8). Manufacturing purchase obligations reached $89.7B at the year end, double the $44.6B at the end of March that the primer recorded, and rose $32.7B in the September quarter alone. The chief executive tied the increase to the work "to improve supply". The part payable after twelve months rose to $7.0B from under $1B in June, so some of the commitments now run into FY2028.

Where the Capital Went

By the primer's measure, the capital-light model strengthened in FY2026. Capex fell to $9.0B, 1.89% of revenue, from $12.7B and 3.06% in FY2025 (Exhibit 9). Over the same year, licensed intangible assets rose $13.5B gross to $29.6B, 1.5 times capex, and total intangible assets, net, rose to $25.7B from $13.3B. The 10-K gives the amounts and does not say what was licensed.

The 10-K also does not say how the licences were paid. Investing cash flow shows $2.4B of other payments, so most of the increase did not pass through it. If the payments ran through operating cash flow, free cash flow already bears them and the capex ratio understates the capital the model uses. If they are still owed, the cash leaves in later years. The income statement bears the cost either way. Amortisation of intangible assets rose to $4.9B from $3.7B and is scheduled at $6.6B in FY2027, with most of the remaining balance amortised by FY2030.

The AI build reaches the statements mostly through operating lines (Exhibit 10). Research and development rose 32.5% to $45.8B. As Siri AI rolls out, its cloud costs are recognised in services cost of sales, and from the December quarter part of device revenue is deferred for Siri AI and recognised over the device's life. Management stated that both will affect gross margin and sized neither. The 10-K adds that Apple relies on third-party cloud providers alongside its own data centres, in a market with "constrained supply, extended lead times, and increasing costs", and that it has made commitments "to secure data center capacity, components, equipment and energy resources". Capex is the measure the primer used for the capital-light claim, and in FY2026 Apple added more in licensed assets than it spent on capex.

Capital Return

The buyback ran above the $22.5B quarterly pace a $90B year needs in three of four quarters, and the year still fell $3.6B short (Exhibit 11). Repurchases were $86.4B on the cash flow basis, and the March quarter's $12.3B, $10.2B below pace, accounts for more than the whole shortfall. With $15.7B of dividends, Apple returned $102.1B, 75% of free cash flow, against 107% in FY2025. The September quarter alone returned 109%. The primer's $90B test applies to FY2027 and FY2028, and $114.0B of authorisation remains. Apple repurchased 296.7M shares, against 401.7M in FY2025, and shares outstanding fell 1.6%.

Net cash fell to $56.8B in September from $62.2B in June, its first fall this year. Across the year it rose $23.1B, mostly because debt fell $18.5B: Apple repaid $12.4B of term debt, cut commercial paper by $5.9B and issued no new term debt. Net cash has no target to be read against. On 30 April Apple stopped providing net cash neutral as a formal target and stated that it would evaluate cash and debt independently (Exhibit 16). On 2 November, the first call under the new chief executive, the chief financial officer confirmed it and restated the allocation philosophy, to "invest everything we need to invest in our business and return excess cash to shareholders".

What the Price Implies

This section characterises the financial conditions the current enterprise value implies, and offers no price target, fair-value estimate or view on whether the shares are mispriced.

At $349.51 on 4 November, an enterprise value of $5,043B implies free cash flow growth of 13.88% a year for ten years, above each of the three earlier stamps (Exhibit 12). That rate would take free cash flow to $499B in year ten, against $136.1B over the last twelve months and average growth of 5.0% a year over FY2016-25. The shares fell 1.0% on 3 November, the first session after the results, and stood 4.4% higher on 4 November.

The price added 1.66 points of the 2.8-point rise since Memo No. 1 and the discount rate 1.17, more than half of that a change in the model's basis (Exhibit 13). The bridge changes one input at a time, in the order shown, which is what lets it attribute the rise. Moving the model to its current conventions for beta, the equity risk premium and the vintage of the risk-free rate raised the discount rate to 9.11% from 8.85% and added 0.64 points. A higher US 10-year yield, 4.91% against 4.69%, took it to 9.33% and added 0.53. The price, $349.51 against $308.91 after the July print, added 1.66. The lower free cash flow base, net cash and the share count together moved it by less than 0.1 point.

The bridge reproduces Memo No. 1's start at 11.12%, a 1bp rounding in that memo's published inputs. The split between the first two steps is approximate, because the model's notes give 9.11% at its 3 September mark. Their 1.17-point total is exact. The share basis moves to 14,590.6M shares outstanding plus dilutive securities, from 14,747M weighted.

The four-driver model asks the same question in operating terms (Exhibit 14). Holding the model's base path for revenue, cash conversion and capex, the price requires an EBIT margin of 59.4% in every year from FY2027 to FY2035, against 33.0% delivered in FY2026. Holding the margin path instead, it requires revenue to compound at 14.8% for nine years, against 6.5% a year over FY2016-25. This memo is the first to publish these instruments, so they have no prior stamp.

The capital-light claim carries a price of its own. The required growth rate rises with the capex the model allows: 14.8% at the base path's terminal 3.2% of revenue, 20.2% at 12% and 28.2% at 20%, a capex level still below the 27% to 43% of revenue the four platforms in Exhibit 15 spent over the last twelve months.

Against peers, Apple's multiple depends on the line it is read on (Exhibit 15). On operating income it trades at 31.9 times, above every peer, against medians of 23.5 times for the platforms and 14.8 times for the device makers. On free cash flow it trades at 37.1 times, below Microsoft at 68.5 times and Meta at 56.5 times. The difference is capex: the platforms spent a median 36.6% of revenue on it, against Apple's 1.9%, so far less of their operating cash reached free cash flow.

Watch Condition Assessment

Two statements in the primer and Memo No. 1 were wrong when they were published (Exhibit 16). Apple dropped net cash neutral as a formal target on its 30 April call, seven weeks before the primer, which treated it as a live commitment and built two triggers on it. Memo No. 1 then read the rise in net cash as drift from that framework. The primer also dated the Google condition's test to a remedy hearing expected in 2H 2027. The remedies were ordered on 2 September 2025 and are under appeal by both the DOJ and Google, so no hearing is pending. Memo No. 1's open question on the intangible assets is answered in part: they are licensed intangibles, and the 10-K does not give their purpose.

No condition is weakening, and only leadership continuity changed status, by passing its test (Exhibit 17). Two conditions are affirmed, three developing and one dormant, and capital return improved on Memo No. 1.

C1 stays dormant with no event to test it. No payment is disclosed and no ruling has come. The 10-K adds that the search arrangements "may not be renewed on favorable terms upon their expiration or at all", and that a reversal on appeal could bring remedies "prohibiting Google from offering the Company commercial terms for search distribution". The filing gives the appeal no date.

C3 holds at Developing because its outcome and its mechanism point different ways. iPhone revenue grew 21.2% for the year, well above the 8% the primer sets for FY2027 and FY2028, and the 10-K credits Pro models. Siri AI, the mechanism the condition names, reached users in beta in September, after most of that growth. A stronger headline with its cause still unread leaves the trend unchanged.

C4 improved, as the buyback ran above pace in both quarters since the March shortfall. The year's $86.4B is not a test result, because the primer's pass years are FY2027 and FY2028. Its weakening trigger, two consecutive quarters below a $70B annual pace, did not fire. Memo No. 1's downgrade rested partly on the net cash drift corrected above. Its other point, that Apple returned 72% of free cash flow, stands, and the year closed at 75%.

C7 passes. Its named test was the first results call under the new chief executive, and that call kept the allocation philosophy and opened no strategic review, after a quarter that returned 109% of free cash flow. Tim Cook remains executive chair, and the chief operating officer has joined the calls. The condition's net cash neutral leg cannot be scored, because Apple retired the target before the primer, so the pass rests on the legs that remain.

What to Watch

The December quarter is the next scheduled read, and it tests the margin guide on actuals (Exhibit 18). It is the first quarter whose margin guide includes Siri AI ramp costs and the first with part of device revenue deferred for Siri AI, so services gross margin carries a new cost line against its 70% floor. It also opens FY2027, the first test year for the $90B buyback and for the iPhone growth the primer requires of the upgrade cycle. The Google appeal has no date, and C1 stays dormant until it is decided. Revenue set the records in FY2026, and the December quarter is the first in which the cost lines carry the test.

Sources: Apple Form 8-K and earnings release for the fourth quarter of FY2026, and the earnings call, 2 November 2026. Apple Form 10-K for the fiscal year ended 26 September 2026, filed 3 November 2026. Apple earnings call for the second quarter of FY2026, 30 April 2026. The Form 10-Q and earnings call for the third quarter of FY2026, as carried in Memo No. 1. The June 2026 primer and Memo No. 1, 3 August 2026. Share prices of $296.42 at the primer's 15 June close, $338.45 and $308.91 at the 30 and 31 July closes, $334.89 at the 3 November close and $349.51 on 4 November. AAPL reverse DCF model, FY2026 file, re-solved by Marvin Labs at $349.51, with a WACC of 9.33% on a US 10-year yield of 4.91% and terminal growth of 3.0%. Peer figures from company filings, over the last twelve months to each company's latest reported quarter, with enterprise value as market capitalisation less cash and investments plus debt, excluding operating leases, at 4 November closes and 3 November exchange rates. Apple's fiscal year ends on the last Saturday of September.