Apple AAPL

Reference table

1. The business — half a phone company, a quarter a services annuity

Apple designs and sells consumer hardware and monetises the resulting installed base through services. The profile lists 166,000 full-time equivalent employees (confirmed verbatim in the 10-K: "approximately 166,000 full-time equivalent employees" as of 2025-09-27), chief executive Timothy D. Cook — until 2026-09-01 — headquarters in Cupertino, and a listing on NASDAQ since 1980-12-12. The 10-K adds a detail the vendor profile does not: 40% of FY2025 net sales went through direct channels and 60% through indirect channels (carriers, resellers, retail partners).

Revenue by product category (fiscal years, $B):

Revenue by geographic segment (FY2025, $B):

Both vendor segment tables reconcile to reported revenue to the dollar ($416.161B against $416.161B), with real category labels and a populated seg_geo. That is not the norm and it is worth stating: on the two fields that most often fail, this file is clean.

Three structural observations.

First, the mix shift is the quality story. Services has gone from 19.6% of revenue in FY2020 to 26.2% in FY2025, compounding 15.2% a year while iPhone compounded 8.8%. Services carries a materially higher gross margin than products — Apple discloses the two separately, which is the single most important margin disclosure on this name, and it is discussed and quantified as far as the data allows in Section 8.

Second, Wearables has been shrinking for two years — −7.1% then −3.6%. Every bullish new-category claim in the knowledge base (glasses, camera-equipped audio wearables, a camera pendant) lands in this line. It is the line to watch to test whether the new product cycle is real, and it is currently going backwards.

Third, Greater China reversed in Q3 FY2026 and the filings say so in successive documents. The 10-K: "Greater China net sales decreased during 2025 compared to 2024 primarily due to lower net sales of iPhone." The 10-Q: "Greater China net sales increased during the third quarter and first nine months of 2026 compared to the same periods in 2025 primarily due to higher net sales of iPhone. The strength in the renminbi relative to the U.S. dollar had a favorable year-over-year impact."

What the 10-Q adds that the vendor payload cannot. The vendor supplies no quarterly segment or category data at all — seg_prod and seg_geo are annual only, ending FY2025. The 10-Q gives the direction of every line for Q3 FY2026: iPhone up on Pro models; Mac up on laptops; iPad DOWN on lower iPad mini and iPad Air sales; Wearables up on Accessories and Wearables; Services up on advertising and cloud services; and all five geographies up. The numeric tables themselves are absent from our extracted filing text — see Section 8, defect 1 — so this is directional, and it is labelled as such throughout.

2. The re-acceleration, and exactly how much of it is durable

This is the analytical crux of the name. Everything else — the multiple, the fair value, the disagreement with our own knowledge base — resolves into one question: is fiscal 2026 a new growth rate or a good cycle with a non-recurring margin kicker?

The trajectory, quarterly, from the vendor income statement (corroborated by the 10-Q's narrative and by earn_cal):

Nine-month FY2026 revenue is $364.357B against $313.695B — up 16.2%. Annual gross margin ran 44.13% (FY2023), 46.21% (FY2024), 46.91% (FY2025). Q3 FY2026's 50.06% is a 357 basis point year-over-year expansion on a business with $109 billion of quarterly revenue.

Now the honest part. Three separate deductions have to be made.

Deduction one — the tariff refunds are in the margin and they do not recur. The 10-Q, verbatim: "Products gross margin and gross margin percentage increased during the third quarter and first nine months of 2026 compared to the same periods in 2025 primarily due to a different mix of products and tariff refunds, partially offset by higher costs, including memory." The mechanism is disclosed too: "on February 20, 2026, the U.S. Supreme Court issued a ruling striking down certain tariffs previously imposed under the International Emergency Economic Powers Act of 1977. The Company has applied for a refund of tariffs paid… and has recognized any refunds received as a reduction of products cost of sales." The size is not disclosed. Compare the FY2025 10-K, where tariffs ran the other way: "Products gross margin percentage decreased during 2025 compared to 2024 primarily due to a different mix of products and tariff costs." So the year-over-year margin comparison is flattered at both ends — a tariff cost in the base, a tariff refund in the current period.

Deduction two — the company is warning about the cost line, in the strongest language it uses. From the 10-Q's macroeconomic discussion: "The Company is experiencing a period of supply constraints and increasing costs for components driven by factors such as industry supply-demand imbalances for components, including advanced semiconductors, storage (NAND) and memory (DRAM). The Company expects these trends to intensify." And on margins generally: "the Company believes, in general, gross margins will be subject to volatility and downward pressure." This is not boilerplate risk language; it is a forward statement placed in the macroeconomic section of the MD&A, and it is corroborated independently by the knowledge base (odd_lots, 2026-02-16, bearish, conviction 70: "Memory shortage is destroying downstream demand — smartphone/PC cuts, delayed premium launches; meaningful margin impact shows up in second half of 2026").

Deduction three — the FY2025 earnings comparison everyone quotes is wrong. Reported net income grew from $93.736B (FY2024) to $112.010B (FY2025), which reads as +19.5%, and diluted EPS from $6.08 to $7.46, or +22.7%. Both are artefacts. The 10-K discloses that in Q4 FY2024 Apple "recorded a one-time income tax charge of $10.2 billion, net" following the European Court of Justice's confirmation of the Irish State Aid Decision — "$15.8 billion payable to Ireland via release of amounts held in escrow, partially offset by a U.S. foreign tax credit of $4.8 billion and a decrease in unrecognized tax benefits of $823 million" — and that FY2025's effective tax rate was lower than FY2024's "due to a $10.7 billion year-over-year decrease in the provision for income taxes related to the State Aid Decision." Normalising FY2024 net income to $103.936B, FY2025 growth was +7.8%, and normalised diluted EPS growth was +10.6%, not +22.7%. This is visible in the quarterly data too: Q4 FY2024 shows net income of $14.736B on $29.591B of operating income, with a $14.874B tax provision.

What survives all three deductions. Revenue growth of 15.7/16.6/16.4% is a volume-and-mix fact, not an accounting one. Operating margin at 32.62% against 29.99% is a 263 basis point expansion that the tariff refund alone is unlikely to explain in full. Services growth is led by advertising and cloud, which are structural rather than cyclical. And R&D rose to fund artificial-intelligence infrastructure — 7.80% of revenue in FY2023, 8.02% in FY2024, 8.30% in FY2025 and 9.19% on a trailing basis — meaning the margin expansion happened despite a rising investment rate, not because of a harvest.

Our reading: the re-acceleration is real, roughly 12-14% of underlying revenue growth rather than 16%, with a gross margin that is more likely to settle in the 48-49% range than to hold 50%. That is still far better than the "mature cash cow" framing that dominated the lane through 2025, and it is meaningfully worse than what a 35x multiple assumes.

3. Growth — what consensus actually models

The FY2029 and FY2030 rows are rejected and italicised. FY2029 revenue of $483.092B is below FY2028's $557.766B, and FY2029 net income of $177.092B exceeds FY2029 EBIT of $161.800B — an impossibility. Coverage collapses to 13 and 9 analysts. No conclusion in this dive uses FY2029 or FY2030. The FY2026-FY2028 rows, on 29, 31 and 19 analysts, are the usable set and are deep by any standard.

A consistency check on FY2026. Nine-month actual revenue is $364.357B; consensus FY2026 of $477.480B implies a Q4 of $113.123B, against Q4 FY2025's $102.466B — +10.4%, a deceleration from 16.4% that the street has already written down. earn_cal independently carries a Q4 revenue estimate of $113.380B, within 0.2%. Nine-month actual diluted EPS is $6.88; FY2026 consensus of $8.80478 implies a Q4 of $1.92, against earn_cal's $2.00 for the 2026-10-29 print — a 4% inconsistency between the two vendor estimate sources, noted in Section 8.

The shape that matters: +14.7%, then +9.1%, then +7.1%. Consensus does not model a new growth company. It models one very good year followed by reversion to high-single-digit growth. EPS compounds 10.1% a year from FY2026 to FY2028. That is the number the 35x multiple has to be judged against.

4. Balance sheet — and the sign error in the vendor's net debt

As at 2025-09-27 (FY2025, 10-K filed 2025-10-31), $B:

• Cash and cash equivalents $35.934 · short-term marketable securities $18.763 · long-term marketable securities $77.723
• Total cash, cash equivalents and marketable securities: $132.420B — and the 10-K states the figure itself: "balances of cash, cash equivalents and marketable securities, which totaled $132.4 billion as of September 27, 2025." The vendor's three components reconcile to the filing exactly.
• Fixed-rate notes: $91.3B aggregate principal ($90.7B carrying), with $12.4B payable within 12 months; future interest payments $37.0B. Commercial paper $8.0B at a 4.19% weighted-average rate. Total notes and commercial paper: $98.7B.
• CORRECTED NET CASH: $132.420B − $98.7B = +$33.8B.
• Total assets $359.241B · total liabilities $285.508B · total equity $73.733B
• Operating lease liabilities $13.720B (weighted-average term 9.8 years, discount rate 3.4%); the 10-K's contractual-obligations table lists $16.8B of fixed lease payment obligations on an undiscounted basis

As at 2026-06-27 (Q3 FY2026, 10-Q filed 2026-07-31) — what the filing discloses:

• Fixed-rate notes carrying amount $82.3B, down from $90.7B. Fair value $71.2B — the notes trade $11.1B below carrying value, a legacy of low-coupon issuance.
• Commercial paper $2.0B, down from $8.0B.
• Gross debt: $84.3B, a $14.4B reduction in nine months.
• Total deferred revenue $14.9B, up from $13.7B — 64% to be recognised within a year, 23% in one-to-two years, 11% in two-to-three, 2% beyond.
• Manufacturing purchase obligations $57.0B ($56.2B within 12 months); other purchase obligations $29.3B, up sharply from $14.8B at FY2025 year-end.
• The remaining $8.8B deemed repatriation tax was paid during the nine months.
• Unrecognised RSU compensation cost $26.7B over a weighted-average 2.7 years, up from $21.8B over 2.5 years.

The vendor's enterprise value carries a sign error. km_ttm.enterpriseValueTTM is $4,585.506B against a vendor market cap of $4,540.743B — an implied net DEBT of $44.763B. That figure nets total debt against cash and cash equivalents ONLY, ignoring the marketable securities entirely. Apple has $96.486B of short- and long-term marketable securities. It is net CASH.

Corrected enterprise value = market cap $4,511.79B (14,594,180,000 shares at $309.15) + notes and commercial paper $98.7B − cash and marketable securities $132.420B = $4,478.03B.

The vendor overstates enterprise value by $107.48B — the $78.5B net-cash-versus-net-debt swing plus $28.8B from a stale share count (Section 8, defect 2). In percentage terms that is 2.4%, which on a $4.5 trillion company is small enough that it barely moves the multiple (corrected EV/FY2026E revenue is 9.38x against the vendor's 9.82x). We report it anyway, and prominently, because the sign is wrong: every screen shows Apple as a net borrower, and it is not. $107 billion of misclassification is larger than the market capitalisation of most companies in the index.

One nuance in the other direction, stated for completeness. If the $13.720B of lease liabilities are treated as debt — the 10-K describes them as operating leases, and the vendor maps them into capitalLeaseObligations and includes them in totalDebt — net cash falls to $20.0B. Still net cash. And gross debt at 2026-06-27 is $14.4B lower than at FY2025 year-end, so the direction of travel is toward more net cash, not less. We use $33.8B and disclose the alternative.

What the balance sheet does and does not do. It removes solvency from the risk list entirely: interest coverage is not meaningfully computable because the vendor reports interestCoverageRatioTTM of 0 (rejected, Section 8), but debt-to-market-cap is 1.86% and free cash flow covers gross debt in about a year. It does not support the equity multiple. Price-to-book is 42.2x because equity is only $73.733B — Apple has distributed almost everything it ever earned; retained earnings are negative $14.264B. That is by design and it is not a warning sign, but it means book value is not a valuation anchor here and none is attempted.

5. Cash flow — the machine that funds the buyback

Free cash flow has exceeded $98B in each of the last four years and has not grown. FY2022's $111.443B is the peak; FY2025's $98.767B is 11.4% below it. Against a $4,511.8B market capitalisation that is a 2.19% free-cash-flow yield and a 45.7x price-to-free-cash-flow. Capital intensity remains trivial — capex is 3.06% of revenue in FY2025 and 2.21% of operating cash flow — which is the concrete basis for the knowledge-base claim that Apple "avoided the capex bonfire" (macrovoices, 2026-04-02, bullish, conviction 68). It is the single most attractive structural feature of the financial profile relative to peers.

The capital return, from the filings rather than the vendor. The 10-K: "During 2025, the Company repurchased 402 million shares of its common stock for $89.3 billion" — an average of $222.14 a share. The 10-Q: "During the nine months ended June 27, 2026, the Company repurchased 215 million shares of its common stock for $61.8 billion" — an average of $287.44 a share. Q3 FY2026 alone was $25.8B of repurchases and $4.0B of dividends and dividend equivalents. In May 2026 Apple entered accelerated share repurchase agreements with $10.0 billion of up-front payments, settling in Q4 FY2026.

Authorisation remaining: $138.0 billion — $38.0B left on the May 2025 $100B programme plus a fresh $100B programme announced 2026-04-30.

Annualised, the nine-month buyback pace is $82.4B, a 1.83% yield. The dividend is $0.27 a quarter (raised from $0.26 in FY2025), $1.08 annualised, a 0.35% yield. Total shareholder yield is roughly 2.2%.

Two honest qualifications. First, the buyback is being executed 29% higher than a year ago ($287.44 against $222.14). Repurchasing at 35x forward earnings destroys per-share value if the multiple is wrong — the arithmetic that made the FY2025 buyback excellent makes the FY2026 buyback merely adequate. Second, the share count is falling slowly: 14,776,353 thousand shares at 2025-10-17 (10-K cover) against 14,594,180 thousand at 2026-07-17 (10-Q cover) — a 1.23% reduction in nine months despite $61.8B of repurchases, because RSU issuance offsets roughly a third of it. Unrecognised RSU cost is $26.7B and stock-based compensation runs 2.94% of revenue.

6. Valuation — priced in or room?

At $309.15, market cap $4,511.8B on the filing's 14,594,180,000 shares, corrected enterprise value $4,478.0B:

Peer context is unusable and must be said. The vendor peer set contains four legitimate mega-cap comparables — Alphabet ($4,572.8B), Microsoft ($3,657.5B), Nvidia ($5,140.2B), Meta ($1,498.0B) — plus Taiwan Semiconductor ($2,160.4B) and Sony ($131.6B) as supply-chain and consumer-electronics adjacencies. It also contains Algorhythm Holdings (RIME) at a $1.0 million market capitalisation, Turtle Beach ($276M) and Nextpower ($14.9B). A peer list spanning $1 million to $5.1 trillion is not a comparison universe, and no peer multiples are supplied in the file, so no peer-multiple comparison is drawn anywhere in this dive.

The vendor's own scoring model rates Apple B with an overall score of 3, with a return-on-equity sub-score of 5 and price-to-earnings and price-to-book sub-scores of 2 and 1 — the model's own quality/valuation split says the same thing this dive does.

6a. What today's price assumes (the inversion)

At $309.15 — 35.1x FY2026 consensus EPS and 32.2x FY2027 — the price embeds roughly these falsifiable claims:

• Earnings compound at least 10% a year through FY2028 AND the multiple does not fall below ~30x. (Consensus-derived: $8.80478 → $9.59026 → $10.6768, a 10.1% FY2026-FY2028 CAGR. Our arithmetic: at a held 29.0x FY2028 multiple the price is $309.6 — exactly spot. The price therefore assumes a 29-35x multiple persists for two more years.) This is the most fragile assumption in the price and it is named as such.
• Gross margin holds near 50% despite the company's own warning. (Filing-derived: Q3 FY2026 gross margin 50.06%, up 357bp year over year, helped by tariff refunds; the 10-Q states memory and semiconductor cost trends are "expected to intensify" and that gross margins will be "subject to volatility and downward pressure." Our estimate: a 48-49% settling range.) Every 100bp of gross margin is roughly $4.8B of annual gross profit at FY2026E revenue, or about $0.25 of EPS pre-tax.
• Greater China's Q3 FY2026 reversal is a recovery, not a quarter. (Filing-derived: the 10-K says FY2025 Greater China fell to $64.377B on lower iPhone sales; the 10-Q says Q3 FY2026 rose on higher iPhone sales, aided by the renminbi. 15.5% of revenue.)
• No adverse regulatory outcome reaches the commission structure or the search-licensing revenue. (Filing-derived: the 10-Q discloses a €500M fine already levied under DMA Article 5(4) and under appeal; preliminary findings in Article 6(4) where fines can reach "10% of the Company's annual worldwide net sales" — over $41B; the App Store commission case at the Supreme Court, which granted certiorari 2026-06-30; and the search-licensing arrangements exposed to remedies on appeal in a separate antitrust case.) The price appears to assume all of this resolves benignly or slowly.
• The 2026-09-01 leadership transition is a non-event. (Filing-derived, 8-K 2026-04-20.) The market has had four months to price it, and the stock is up over that period, so this is the market's stated view rather than an oversight.
• Where our knowledge-base signal reads on this: the recent lane is more bullish than consensus, not less — several sources carry an explicit $400 target (+29.4%). We are below both the street and our own recent knowledge-base lane on price, and that is deliberate.

6b. The return bridge (why the multiple moves)

Expected return over 12-24 months decomposes as: EPS growth (+10.1% a year, consensus) + shareholder yield (+2.2%) + multiple drift.

Our base case assumes modest multiple COMPRESSION — 35.1x FY2026 to 30.5x FY2028 — and that assumption is doing most of the work. The mechanics are worth stating precisely, because the guardrail against confusing rolldown with re-rating applies directly here: at a constant $309.15, the forward multiple falls mechanically from 35.1x to 29.0x simply by rolling from FY2026 to FY2028 estimates. That is not compression. Compression is the market paying a lower multiple at each date. Our base of $326 pays 30.5x FY2028 earnings in roughly two years' time against 35.1x FY2026 earnings today — genuine, modest de-rating of about 13%, which offsets most of the 21% cumulative EPS growth and leaves +5.5%.

Why compression rather than a hold? Three reasons, all forward. Growth maturation: consensus itself takes revenue growth from 14.7% to 7.1% over the same window, and multiples follow growth rates. Margin normalisation: a 50% gross margin achieved partly through non-recurring tariff refunds, with the company warning about component costs, is unlikely to be the run-rate. And risk normalisation in the wrong direction: the regulatory tail is getting nearer, not further away, with the Supreme Court and the European Commission both live inside the window.

If instead the multiple simply HOLDS at 35.1x, applied to FY2028E $10.6768, the price is $374.8 — +21.2%. That is close to the bull case and it is the entire bull case. Note what this means: essentially all of the return above the mid-single digits is a multiple assumption, not an earnings assumption. That is the definition of a fragile leg, and it is exactly the leg the most rigorous bear in the knowledge base attacks — not the earnings, but the mechanism setting the multiple.

6c. Variant perception (where we differ, what would surprise)

• We differ from the 2025 bear consensus, and from a large part of our own knowledge base, on the growth. Forty-six bearish claims since 2025 rest on a premise the filings have now falsified: "revenue growth has slowed to almost nothing, and it becomes a mature cash cow" (invest_like_the_best, 2025-06-05, conviction 65) and "Apple hasn't grown income in 3-4 years… no real growth drivers" (mike_green, 2026-02-14, conviction 78). Revenue grew 15.7%, 16.6% and 16.4% in the three quarters of FY2026 reported so far and gross margin crossed 50%. We treat those claims as dated and say so. Watchable number: Q4 FY2026 revenue on 2026-10-29 against the $113.380B consensus, which requires only 10.4% growth.
• We differ from the recent bulls, and from the street, on the price. Our base of $326 is 4.4% below the street consensus of $341.11 and 18.5% below the $400 target that appears twice in the recent lane at conviction 85. The disagreement is not about the business; it is entirely about what 35x assumes. We think the crowd is extrapolating a cyclical upgrade year into a durable growth rate, at a multiple that already discounted the good news before the good news arrived — the stock is up 53.3% over twelve months against SPY's 24.3%, most of which preceded the FY2026 prints.
• We differ on the quality of the margin. The market appears to be treating 50.06% as the new baseline. The 10-Q says products margin rose partly on tariff refunds and that memory cost pressure is expected to intensify. We model 48-49%. Watchable number: Q4 FY2026 consolidated gross margin against Q3's 50.06% and Q4 FY2025's 47.18%.
• We differ on the balance sheet, and this one is not a judgement call. Every screen shows Apple carrying $44.8B of net debt. Counting the $96.5B of marketable securities the vendor omits, it holds $33.8B of net cash, and gross debt fell a further $14.4B in the nine months to 2026-06-27. A $78.5B sign error on the most-screened stock in the world.
• We do NOT differ on the structural bear case, and we cannot refute it. The passive-flow argument — that inflows disproportionately buy the largest and least-liquid names, mechanically supporting the multiple (mike_green, 2026-02-11, conviction 85; brent_johnson, 2025-08-17, conviction 80) — is not falsified by good earnings and is not falsifiable by them. It is a reason the multiple could compress far more than our base assumes, and it is why the bear case is $230 rather than $270.
• Positive surprise that would force a re-rate: FY2027 consensus EPS revised UP from $9.59026 — one source predicts precisely this on an iPhone replacement cycle — combined with a gross margin that holds above 49% without tariff refunds. Secondarily, a disclosed artificial-intelligence monetisation arrangement of the $20-30B scale the knowledge base identifies as the only genuine re-rating trigger, or Wearables returning to growth as the new categories ship. Watchable numbers: FY2027 consensus EPS; Q4 gross margin; the Wearables line.
• Negative surprise that would break the thesis: an adverse ruling in the EU Article 6(4) investigation (fines to 10% of worldwide net sales), a Supreme Court outcome that eliminates the commission on external purchases, or remedies that end the search-licensing payments. Any of these lands directly in Services, which is 26.2% of revenue at the highest margin in the company, and none is priced. Secondarily: a Q4 gross margin below 48%, which would confirm the tariff-refund read.

Synthos fair values

All three anchors are multiples of FY2028 consensus EPS of $10.6768 (19 analysts), cross-checked against FY2027 consensus EPS of $9.59026 (31 analysts).

• Bear ~$230 — 21.5x FY2028E. Cross-check: 24.0x FY2027E. The scenario: the FY2026 upgrade cycle pulls demand forward and FY2027 flattens; tariff refunds stop and memory costs bite, taking gross margin back toward 47%; an adverse regulatory outcome reaches the commission structure or the search-licensing revenue; and the multiple compresses to what a high-single-digit-growth consumer hardware business with a regulatory overhang is normally paid. This does not require the passive-flow thesis to be right — if it is, the bear is lower. −25.6%.
• Base ~$326 — 30.5x FY2028E. Cross-check: 34.0x FY2027E. The scenario: revenue grows 14.7% in FY2026 and roughly 9% and 7% thereafter as consensus models; gross margin settles at 48-49%; the buyback continues at roughly $80B a year; the regulatory matters grind on without a decisive adverse ruling inside the window; and the multiple de-rates modestly with the growth rate. +5.5%, plus a ~2.2% shareholder yield — a total return in the high single digits.
• Bull ~$390 — 36.5x FY2028E. Cross-check: 40.7x FY2027E. The scenario: the September product cycle and the new categories deliver, FY2027 estimates are revised up, an artificial-intelligence monetisation arrangement materialises at the scale the knowledge base sizes, gross margin holds above 50%, and the market keeps paying today's multiple on a higher base. +26.1%. Note this sits just below the $400 target carried twice in the recent knowledge-base lane at conviction 85, and it requires the multiple hold that Section 6b flags as the fragile leg.

Base is 5.5% above spot; asymmetry is roughly 1:1 (25.6% down, 26.1% up). A symmetric payoff, a base case below the street, a leadership change 28 days out, and a price in the 75th percentile of its annual range after a 53% run. That is a Hold, and it would be a Hold even if the growth story is entirely right — because the growth story is what you are already paying for.

7. Knowledge base — 218 claims, 44 sources, and a lane that inverts by date

Raw hits naming Apple or AAPL: 229. Quarantined and EXCLUDED: 11 (quarantine_misattributed). Usable: 218, across 44 distinct sources, spanning 2015-12-09 to 2026-07-28.

Two structural facts about this lane before any claim is quoted.

First, it is genuinely independent. The largest sources are compound_and_friends (29 claims), invest_like_the_best (15), mike_green (14), jordi_visser (14), forward_guidance (10), raoul_pal (10), odd_lots (9), all_in (9), we_study_billionaires (9) and darius_dale (9). There is not one claim from Apple management anywhere in the lane. Management-class voices total five claims from three other companies' executives (jensen_huang 3, eli_lilly_ceo 1, mark_zuckerberg_mgmt 1) — half-weighted, and none is about Apple's own prospects. Promoter-class voices total three claims (anthony_pompliano 2, anthony_pompliano_show 1) — low-weighted. That leaves 39 fully-weighted independent sources, which makes this one of the most credible lanes in the coverage universe, and materially more credible than the typical mega-cap lane dominated by the company's own executives.

Second, and this is the finding: the lane inverts by date.

The bear case is overwhelmingly 2025-vintage and it is built on premises the FY2026 filings contradict.

The bear lane — quoted, weighted and dated:

> 2025-06-05 · invest_like_the_best · bearish · conviction 65

> "Apple is probably going to be disrupted by AI; it missed autonomous vehicles through repeated management turnover, revenue growth has slowed to almost nothing, and it becomes a mature cash cow."

> 2026-02-14 · mike_green · bearish · conviction 78

> "Apple hasn't grown income in 3-4 years (last from COVID stimulus)… no real growth drivers, yet trades at record-high multiple." (Quoted with a third-party company name removed per house style.)

> 2026-05-07 · mike_green · bearish · conviction 62

> "The market pays ~30x for a mature company like Apple, whereas he estimates Apple is worth probably a single-digit PE."

> 2026-02-11 · mike_green · bearish · conviction 85 — the highest-conviction bearish claim in the lane

> "Passive creates a price-weight spiral: inflows disproportionately buy the largest, least-liquid names like Apple, reinforcing momentum mechanically and entrenching concentration."

> 2026-02-16 · odd_lots · bearish · conviction 70

> "Memory shortage is destroying downstream demand — smartphone/PC cuts, delayed premium launches; meaningful margin impact shows up in second half of 2026."

> 2026-05-20 · jim_bianco · bearish · conviction 30

> "Rhetorically questions Apple's AI strategy — they don't seem to be doing anything; could become the next Cisco, which took 25 years to reclaim its 2000 peak."

> 2026-07-08 · mike_green · bearish · conviction 45 — the most recent bearish claim in the lane

> "Apple more than doubled its memory-related price increases to protect margins, pushing a growing fraction of users to defer replacing their iPhones."

The bull lane — the recent, specific, price-targeted claims:

> 2026-07-28 · Josh Brown · bullish · conviction 85 — the most recent claim in the lane

> "Absolutely convinced Apple is the one to own for the second half; capital-light on-device AI makes it a toll booth for consumer AI without trillion-dollar capex — going to 400."

> 2026-07-07 · compound_and_friends · bullish · conviction 85

> "Apple is the mega-cap to watch in H2 and can reach $400 from ~$311 (a 33% gain); it's breaking out now — been pounding the table on it since the year began."

> 2026-07-14 · compound_and_friends · bullish · conviction 72

> "Apple's 2.5B installed base means it owns the consumer AI relationship — whatever LLM you use, you access it through iOS and Apple gets paid regardless of who wins."

> 2026-07-07 · compound_and_friends · bullish · conviction 65 — directly testable against our data

> "Consensus has Apple EPS up 17.4% in 2026 and ~9.8% in 2027; the 2027 number gets revised higher on the iPhone comeback cycle — this is no longer a non-growth story."

> 2026-07-23 · odd_lots · bullish · conviction 68

> "The next three years will be the biggest three product years in Apple's history after five years of quietly building its most packed product overhaul ever — a lot of stuff is coming."

> 2026-07-21 · compound_and_friends · bullish · conviction 68

> "Staying long Apple — China approved on-device AI for iPhones, pricing power lets them hold prices as memory costs fall (earnings expansion), with more catalysts in Sept/Oct not yet on people's radars."

> 2026-04-02 · macrovoices · bullish · conviction 68

> "Apple avoided the capex bonfire; sitting on cash it can buy distressed AI infra as second/third owner and capture edge/on-device inference driven by privacy and latency."

The most useful claim in the entire lane is neutral:

> 2026-07-07 · compound_and_friends · neutral · conviction 55

> "Only AI monetization — $20-30B contracts with frontier models — can re-rate Apple's multiple, not the foldable; skeptical Siri suddenly works agentically."

What this lane establishes, and how we weight it.

One: the bull case is now better supported by the filings than the bear case. The compound_and_friends claim that "consensus has Apple EPS up 17.4% in 2026 and ~9.8% in 2027" is checkable against our own data: consensus is +18.0% and +8.9%. That source is reading the same numbers correctly and drawing the conclusion that this is no longer a non-growth story. We agree with that conclusion. The "mature cash cow with near-zero revenue growth" framing from mid-2025 is simply wrong now.

Two: the strongest bearish argument is untouched by any of it. mike_green's passive-flow thesis at conviction 85 attacks the multiple mechanism, not the earnings. A 50% gross margin and 16% revenue growth do not answer it. It is unfalsifiable in the short run and it is the reason our bear case is $230 rather than $270, and the reason the verdict is Hold rather than Buy despite agreeing with the bulls on the business.

Three: where we differ from the recent lane is entirely on price. Two sources carry a $400 target at conviction 85. Our bull is $390 and our base is $326. The knowledge base and this dive agree on the direction of the business and disagree on what it is worth, which is a cleaner and more useful disagreement than the alternative.

Why Moderate rather than High conviction. Breadth is exceptional (44 sources) and independence is exceptional (zero company-management claims). But the full-history net stance is near parity, the inversion is recent enough that it has not been tested through a down quarter, and no claim in the lane engages seriously with the 35x multiple except to argue it should be lower.

11 claims discarded and why: all eleven carry the quarantine_misattributed source tag, meaning the underlying utterance could not be reliably attributed to a named speaker under the attribution policy. Several are near-duplicates of claims that DO appear under a properly attributed slug — for example, two quarantined 2026-06 claims about the index not being safe past 2030 mirror jordi_visser claims of 2026-06-06 and 2026-06-17 almost verbatim. They are excluded from every count, every stance tally and every quotation in this dive.

8. Data integrity — what we rejected from the vendor file and why

Nine material findings. Every filing-versus-vendor conflict is named here.

1. The numeric tables are MISSING from our extracted filing text — the largest gap in this dive, and it hits the single most important disclosure on the name. Apple reports products and services gross margin separately, in dollars and percentages, in both the 10-K and the 10-Q. Our extracted text contains the surrounding narrative — "Products gross margin and gross margin percentage increased during the third quarter and first nine months of 2026… primarily due to a different mix of products and tariff refunds" — but every numeric cell in every table was dropped in extraction, including the products/services gross margin table, the quarterly segment table, the quarterly category table, the balance sheet, and the income statement. The same is true of the 10-K. Consequence: this dive reports the DIRECTION of products and services gross margin verbatim from the filing but cannot report the percentages, and consolidated gross margin (50.06% in Q3 FY2026) is computed from the vendor income statement instead. Inline narrative figures survived and are used extensively (cash of $132.4B, notes of $91.3B/$82.3B, commercial paper of $8.0B/$2.0B, deferred revenue of $14.9B, buybacks of $89.3B and $61.8B, share counts, purchase obligations). This is stated rather than papered over.

2. km_ttm.enterpriseValueTTM treats Apple as a net BORROWER. It is net CASH — a $78.5B sign error. The vendor reports enterprise value of $4,585.506B against a market cap of $4,540.743B, implying net debt of $44.763B. That nets debt against cash and cash equivalents only, ignoring $96.486B of short- and long-term marketable securities — which the 10-K itself aggregates: "cash, cash equivalents and marketable securities, which totaled $132.4 billion as of September 27, 2025." Corrected: net CASH of $33.8B; corrected enterprise value $4,478.03B. In percentage terms the enterprise-value error is 2.4%, small enough not to distort the multiple, but the sign is wrong on the most widely screened equity in the market. evToSalesTTM (9.823x), evToEBITDATTM (27.216x), netDebtToEBITDATTM (0.266x) and enterpriseValueMultipleTTM (27.216x) are all recomputed or rejected accordingly.

3. The vendor market capitalisation uses a stale share count — the 10-Q wins. The vendor reports $4,540.596B, implying 14,687.4 million shares at $309.15. The 10-Q cover page states 14,594,180,000 shares issued and outstanding as of 2026-07-17. Correct market capitalisation: $4,511.79B — the vendor overstates by $28.81B, or 0.64%. All per-share and market-cap-derived figures in this dive use the filing's count.

4. FILING VERSUS VENDOR on the FY2025 buyback: the vendor reports $90.711B; the 10-K says $89.3 billion; we use $89.3 billion. cf_a.commonStockRepurchased for FY2025 is −$90.711B, while the 10-K states "During 2025, the Company repurchased 402 million shares of its common stock for $89.3 billion." A $1.4B difference, most likely settlement timing or net-share-settlement treatment. The filing wins. The dividend figure agrees — vendor $15.421B against the 10-K's $15.4B — as does FY2024 (vendor $94.949B against the 10-K's $95.0B), which suggests the FY2025 discrepancy is specific rather than systematic.

5. FILING VERSUS VENDOR on deferred revenue: the vendor reports $9.055B; the 10-Q reports $14.9 billion total; we use the filing. bal_a.deferredRevenue for FY2025 is $9.055B and deferredRevenueNonCurrent is $0 — but the 10-K and 10-Q both disclose a total deferred revenue balance, $13.7B at 2025-09-27 rising to $14.9B at 2026-06-27, of which only 64% is current. The vendor is capturing the current portion and zeroing the non-current portion, understating total deferred revenue by roughly $4.6B. This matters because deferred revenue is the cleanest available proxy for services-attach momentum, and it grew 8.8% in nine months.

6. est FY2029 and FY2030 rows are internally impossible — REJECTED. FY2029 revenue of $483.092B is below FY2028's $557.766B, an implied 13.4% decline that no analyst is forecasting; FY2029 net income of $177.092B exceeds FY2029 EBIT of $161.800B, which cannot happen; and the implied share count from FY2029 net income and EPS is ~15.0 billion, above today's 14.59 billion despite an $80B annual buyback. Coverage falls to 13 and 9 analysts. No conclusion uses FY2029 or FY2030. By contrast the FY2026-FY2028 rows (29/31/19 analysts on EPS) are internally consistent and are used throughout. Consistent with the standing warning, forward valuation here runs on epsAvg, not on the EBITDA/EBIT block.

7. The two vendor estimate sources disagree on the next quarter by 4% — flagged, both reported. est FY2026 EPS of $8.80478 against nine-month actual diluted EPS of $6.88 implies a Q4 FY2026 of $1.92; earn_cal carries $2.00 for the 2026-10-29 print. Revenue agrees far better ($113.123B implied against $113.380B). We quote $2.00 for the print because earn_cal is the calendar-specific field, and note the annual figure implies less.

8. ratios_ttm.interestCoverageRatioTTM reports 0 — REJECTED as meaningless. The vendor also reports debtServiceCoverageRatioTTM of 10.91x and operatingCashFlowCoverageRatioTTM of 1.74x, which are usable. Interest coverage of zero is a null, not a reading; Apple's interest burden ratio is reported as 1.0037, implying net interest income rather than expense. No interest-coverage figure appears in this dive.

9. The tech block disagrees with the quote block on the price and the 52-week range — quote is used. tech.last is $310.15 against quote.price of $309.15; tech.hi52 is $340.08 against quote.yearHigh of $344.57; tech.lo52 is $202.92 against quote.yearLow of $202.16; tech.sma50 is $309.63 against priceAvg50 of $309.50; tech.sma200 is $278.54 against priceAvg200 of $277.96. All price and range figures in this dive come from quote. The relative-return series (ret_12m, spy_12m, qqq_12m) has no counterpart in quote and is used as supplied, directionally. Related oddity: rsi14 of 39.5 sits awkwardly against a stock that closed +1.89% on the day, level with its 50-day average and 11.2% above its 200-day. We report RSI but place no weight on it.

Also flagged, not defects: the vendor peer set spans $1.0 million (Algorhythm Holdings) to $5.14 trillion (Nvidia) and is not a usable comparison universe — no peer multiples are drawn anywhere in this dive. bal_a.goodwill and intangibleAssets of $0 are correct: Apple does not separately present them. capitalLeaseObligations of $13.720B is a mapping artefact — the 10-K describes these as operating leases with a 9.8-year weighted-average term and a 3.4% discount rate — and their inclusion in totalDebt of $112.377B is disclosed and handled explicitly in Section 4. retainedEarnings of −$14.264B is real and reflects decades of distributions exceeding retained profit; it is why price-to-book of 42.2x is not used as a lens.

Not available on this plan, and stated as a gap: earnings-call transcripts. Apple's Q3 FY2026 call took place on 2026-07-30. Management's commentary on the tariff refund's size, the memory cost outlook, the September product cycle and the leadership transition is absent from this analysis, and the 8-K filed 2026-07-30 carries the earnings release, not the call. No management quotation appears anywhere in this dive, and none has been invented.

9. Technicals — extended, not exhausted

• Price $309.15, up +1.89% on the day on 45.8 million shares against a 54.4 million average — a rising day on below-average volume, in a session that also lifted the semiconductor complex hard.
• −10.3% from the 52-week high of $344.57; +52.9% above the 52-week low of $202.16. The stock sits at the 75th percentile of its annual range.
• Level with the 50-DMA ($309.50, −0.11%) and +11.2% above the 200-DMA ($277.96). The 200-day is well below and rising. This is an uptrend, not a base.
• Relative performance: 3-month +12.0% against SPY +7.6% and QQQ +7.7%; 6-month +14.9% against SPY +11.1% and QQQ +15.7%; 12-month +53.3% against SPY +24.3% and QQQ +30.8% — a 29-point outperformance against the index over a year.
• RSI 39.5 · MACD +3.17 — the RSI is inconsistent with the rest of the price structure and is flagged in Section 8; the MACD is mildly positive.
• Practical read on entry. This is the opposite of the MKL setup. There is no sentiment trough here: 71 of 111 analyst ratings are buy or strong buy, the price target dispersion runs $245 to $400 with a $355 median, and the stock has beaten the index by 29 points in a year. Buying at $309.15 is chasing a name that has already re-rated on the news this dive is analysing. The 200-DMA at $277.96 is 10.1% below and would be a materially different entry; so would the $245 low end of the target range. Anyone establishing a position here is paying for the re-acceleration in full, four weeks before a leadership change.

10. Insiders — one real sale, from the departing chair

The reading, stated carefully. Six of the eight transactions in this file are mechanical: restricted-stock-unit settlements at a zero price and the share withholding that pays the resulting tax bill. They carry no signal. The 116-share sale by the principal accounting officer is trivial.

One transaction is real: Arthur Levinson sold 50,000 shares at $311.02 on 2026-05-27, a $15.6 million open-market disposition, retaining 3,764,576 shares. The context matters and it cuts both ways. Levinson is the Board Chair who, per the 8-K of 2026-04-20, becomes Lead Independent Director on 2026-09-01 as part of the leadership transition — so a director restructuring a very large, very long-held position around a governance change is unremarkable, and he retained 98.7% of it. But it is the only discretionary sale in the file, it was executed within 0.6% of the price this dive is struck at, and there is not one open-market purchase by any insider anywhere in the file.

The 10-Q adds two disclosures the vendor insider feed does not carry, and they are material. Under Item 5, Insider Trading Arrangements: Tim Cook entered a Rule 10b5-1 trading plan on 2026-05-28 — the day after Levinson's sale, and five weeks after the succession announcement — providing for the sale of up to 16,250 shares plus shares vesting during the plan, and the gifting of up to 37,104 shares, expiring 2026-10-30. Jennifer Newstead entered a plan on 2026-05-05 covering up to 24,912 shares plus 90% of shares vesting through 2027-03-15.

Net: no insider is buying, the departing chair sold at essentially today's price, and the departing chief executive has a selling plan running through October. This is a mildly negative signal, not a damning one — post-transition share disposition by executives changing roles is routine — but in a dive weighing whether to pay 35x, it belongs on the same side of the ledger as the price.

11. Verdict, kill-criteria and flip conditions

Hold.

What is genuinely good here, stated plainly, because the house bias runs against mega-caps and it should not run past the evidence. Revenue grew 15.7%, 16.6% and 16.4% year over year in the three FY2026 quarters reported, after three years at −2.8%, +2.0% and +6.4%. Gross margin has expanded every single quarter for five quarters and crossed 50% for the first time in the quarter reported five days before this dive. Operating margin reached 32.62%. Greater China reversed a full-year decline. Services is 26.2% of revenue, compounding 13.5%, now led by advertising and cloud rather than App Store commissions. Free cash flow has exceeded $98 billion in each of the last four years on capital expenditure of 3.06% of revenue. The balance sheet holds $33.8 billion of net CASH — a fact no screen in the market displays correctly — and gross debt fell $14.4 billion in nine months. And the knowledge-base lane, 44 independent sources deep with zero company-management claims, has flipped to 26 bullish against 7 bearish since 2026-04-01. The "mature cash cow" thesis that dominated 2025 is dead, and this dive says so against the grain of its own historical claim base.

Why that is still not a Buy. At $309.15 you pay 35.1x FY2026 and 32.2x FY2027 consensus earnings for a company whose own consensus decelerates revenue growth from 14.7% to 9.1% to 7.1% and compounds EPS 10.1% a year through FY2028. Our base fair value of $326 is +5.5%, 4.4% below the street, and the bear-to-bull span is symmetric at −25.6% / +26.1%. Three concrete cautions the filings supply: part of the record margin is tariff refunds that cannot recur; the company says component cost pressure will "intensify" and that gross margins face "volatility and downward pressure"; and the EU can, per the 10-Q, fine up to 10% of annual worldwide net sales — over $41 billion — in a live investigation, while the App Store commission sits at the Supreme Court and the search-licensing revenue is exposed to remedies on appeal. All of that regulatory exposure lands in Services, the highest-margin quarter of the business. Add a chief executive change 28 days away, a stock up 53.3% in twelve months against the index's 24.3%, no insider buying and one real insider sale at $311.02, and the picture is a good business at a full price with a symmetric payoff. That is the definition of Hold.

Pre-registered KILL criteria — what would move this to Avoid:

• A final adverse determination in the EU Digital Markets Act Article 6(4) investigation, or any fine sized against worldwide net sales rather than a fixed amount.
• A Supreme Court outcome, or a remanded District Court injunction, that eliminates the commission on purchases made outside an app. The 10-Q's own risk-factor language: "If the Company is ultimately unsuccessful in defending its commission structure… the Company's business, results of operations, and financial condition could be materially and adversely affected."
• Remedies that end the search-licensing arrangements, which the 10-Q flags explicitly as a revenue stream at risk on appeal.
• Consolidated gross margin below 48% for two consecutive quarters, which would confirm that the 50.06% print was tariff refunds plus mix rather than a new baseline, and that the memory cost warning has landed.
• FY2027 consensus EPS revised DOWN below ~$9.00, breaking the earnings leg the entire valuation rests on.
• Greater China resuming its decline after the Q3 FY2026 reversal.
• A visible execution stumble in the leadership transition — departures of senior product or operations leadership in the two quarters after 2026-09-01.

Pre-registered FLIP TO BUY (upsize) conditions:

• A price below ~$278 (the 200-day average) with the FY2027 earnings path intact — that is 30.5x FY2026E and 29.0x FY2027E, where the base case works without any multiple assumption.
• FY2027 consensus EPS revised UP through $10.00 while gross margin holds above 49% without tariff refunds. This is the single cleanest positive test and one knowledge-base source has pre-registered it.
• A disclosed artificial-intelligence monetisation arrangement at the $20-30B scale the knowledge base identifies as the only genuine re-rating trigger.
• Wearables returning to growth after −7.1% and −3.6% years, which would evidence that the new product categories are reaching material revenue rather than merely being announced.
• Resolution of the EU and App Store matters on terms that preserve the commission structure, removing the largest unpriced tail.
• Any open-market insider purchase, of which there is currently none.

Where AAPL fits in the Synthos Framework Portfolio. The mega-cap quality / low-volatility sleeve, as a maintained position at index-like weight rather than a new commitment. Suggested weight for a new position: 0%, or 1% only as a starter with the explicit intent of adding at or below $278. The reasoning is the same as the verdict: the analytical work here changes what you believe about the business, not about the price. If a portfolio already holds Apple, the re-acceleration and the corrected net-cash position are reasons to keep holding through the transition rather than to trim into strength. If it does not, there is no edge in paying 35.1x four weeks before a leadership change, when the same earnings path at the 200-day average would be a genuinely attractive entry. Logged as a tracked Synthos call (Hold) as of 2026-08-04 at $309.15, with the pre-registered flip level at $278 and the pre-registered positive test at FY2027 consensus EPS above $10.00.

Single biggest risk: a regulatory decision that reaches the commission structure or the search-licensing revenue. These are not diffuse "regulatory overhang" concerns; they are dated, live matters with named forums. The 10-Q discloses that the European Commission has already fined Apple €500 million under Article 5(4), that preliminary findings exist in Article 6(4) where fines can reach 10% of annual worldwide net sales — over $41 billion on FY2025's $416.161B — that the Supreme Court granted review of the App Store contempt standard on 2026-06-30, and that the search-licensing arrangements are exposed to remedies under appeal in a separate case. All of it concentrates in Services, which is 26.2% of revenue at the highest margin in the company and the entire basis of the quality argument for the multiple. Nothing in the price appears to reflect an adverse outcome in any of them.

Most fragile assumption in the price: that the market keeps paying 30-35x forward earnings for a business consensus expects to grow earnings ~10% a year. Strip that assumption out and the arithmetic is stark — at a held 29.0x FY2028 multiple, the price today is exactly fair value with zero return. Every dollar of expected gain above the shareholder yield is a bet on the multiple, not on the company. And the most rigorous voice in our own knowledge base argues, at conviction 85, that the multiple is a mechanical artefact of passive flows into the largest and least-liquid names rather than a fundamental judgement. We cannot falsify that argument with earnings, and neither can anyone else. That is why the verdict is Hold and not Buy, on a name whose business results this dive otherwise defends.

Provenance & disclosures

• Traceability: 229 raw knowledge-base claims name Apple or AAPL across 51,928 records; 11 are quarantine_misattributed and are EXCLUDED from every count, tally and quotation, leaving 218 usable claims across 44 distinct sources (breadth 44, net conviction mixed) spanning 2015-12-09 to 2026-07-28. Largest sources: compound_and_friends (29), invest_like_the_best (15), mike_green (14), jordi_visser (14), forward_guidance (10), raoul_pal (10), odd_lots (9), all_in (9), we_study_billionaires (9), darius_dale (9). Zero claims originate from Apple management — an unusually independent lane for a mega-cap. Management-class voices (half weight, none about Apple's own prospects): jensen_huang (3), eli_lilly_ceo (1), mark_zuckerberg_mgmt (1). Promoter-class voices (low weight): anthony_pompliano (2), anthony_pompliano_show (1). That leaves 39 fully-weighted independent sources. Full-history stance 100 bullish / 76 bearish / 42 neutral; since 2026-04-01, 26 bullish / 7 bearish / 4 neutral. All quotations are verbatim from stored claim text except where a third-party company name was removed for house style, which is marked at the point of quotation.
• Data as-of: fundamentals — income statement 2026-06-27 (Q3 FY2026, reported 2026-07-30, 10-Q filed 2026-07-31, five days before this dive); balance sheet and cash flow 2025-09-27 (FY2025, 10-K filed 2025-10-31 — the vendor payload contains no quarterly balance sheet or cash flow, so all balance-sheet figures are FY2025 year-end, updated where the 10-Q discloses a narrative figure). Estimates 2026-08-04 · prices 2026-08-04 ($309.15, +1.89%; 50-DMA $309.50; 200-DMA $277.96) · knowledge-base claims 2026-08-04. Filings used: 10-Q 2026-07-31 (Q3 FY2026), 10-Q 2026-05-01 (Q2 FY2026), 10-K 2025-10-31 (FY2025), 8-K 2026-07-30, 8-K 2026-04-30, 8-K 2026-04-20. No figure comes from memory, recall or external retrieval.
• The three 8-Ks, identified: 2026-07-30 — Item 2.02, the Q3 FY2026 earnings release (Exhibit 99.1; the release itself, not the call). 2026-04-30 — Item 2.02, the Q2 FY2026 earnings release. 2026-04-20 — the leadership succession: "Tim Cook will transition from his role as Chief Executive Officer to Executive Chair of Apple's Board of Directors, effective September 1, 2026… the Board appointed John Ternus, Apple's Senior Vice President of Hardware Engineering, as Chief Executive Officer and a member of the Board… Art Levinson, current Chair of the Board, will become Lead Independent Director." Ternus, 50, joined Apple in 2001. This is the most consequential of the three and it is an input to the verdict, the risk score and the short-horizon stance.
• Data rejections (detailed in Section 8): km_ttm.enterpriseValueTTM corrected from $4,585.506B to $4,478.03B — the vendor's implied net DEBT of $44.763B ignores $96.486B of marketable securities and the true position is net CASH of $33.8B, a $78.5B sign error; vendor market cap corrected from $4,540.596B to $4,511.79B on the 10-Q's share count of 14,594,180,000; cf_a.commonStockRepurchased FY2025 of $90.711B rejected in favour of the 10-K's $89.3 billion; bal_a.deferredRevenue of $9.055B rejected in favour of the 10-Q's total of $14.9 billion; est FY2029 and FY2030 rows rejected as internally impossible (revenue below FY2028; net income above EBIT); interestCoverageRatioTTM of 0 rejected as a null; the tech block's price and 52-week range rejected in favour of quote; rsi14 reported but unweighted; the peer set rejected as unusable (spans $1.0M to $5.14T); priceToBookRatioTTM of 42.2x reported but rejected as a lens given negative retained earnings of −$14.264B; capitalLeaseObligations of $13.720B disclosed as an operating-lease mapping and handled explicitly; the est versus earn_cal Q4 FY2026 EPS disagreement ($1.92 implied against $2.00) flagged with both figures reported.
• The single largest data gap: every numeric table in both the 10-K and the 10-Q was dropped in text extraction, including the products-versus-services gross margin table — the most important margin disclosure on this name — and the quarterly segment and category tables. This dive therefore reports the DIRECTION of products and services gross margin verbatim from the filing narrative and computes consolidated gross margin (50.06% in Q3 FY2026) from the vendor income statement. Inline narrative figures survived extraction and are used throughout. No products or services gross margin percentage is quoted anywhere in this dive, because none is available.
• Basis note — the FY2025 earnings comparison: reported FY2025 net income growth of +19.5% and diluted EPS growth of +22.7% are artefacts of a one-time charge in the base year. The 10-K discloses a $10.2 billion net one-time income tax charge recorded in Q4 FY2024 following the European Court of Justice's confirmation of the Irish State Aid Decision. Normalising FY2024 net income to $103.936B, FY2025 net income grew 7.8% and normalised diluted EPS grew 10.6%. This dive uses the normalised figures wherever a growth rate is quoted.
• Basis note — the tariff refund: the Q3 FY2026 products gross margin improvement is attributed by the 10-Q "primarily due to a different mix of products and tariff refunds", following the U.S. Supreme Court's 2026-02-20 ruling striking down certain tariffs imposed under the International Emergency Economic Powers Act. The size of the refund is not disclosed. This dive therefore treats part of the 357 basis point year-over-year gross margin expansion as non-recurring and models a settling range of 48-49% rather than 50%+.
• Estimate coverage: deep on the rows used — 29 analysts on FY2026 EPS, 31 on FY2027, 19 on FY2028; 21, 30 and 21 on the corresponding revenue lines. FY2029 (9 EPS analysts) and FY2030 (15) are rejected for the internal contradictions set out in Section 8. Forward valuation runs on epsAvg; the est EBITDA and EBIT rows are not used.
• Peer note: the vendor peer set contains four legitimate mega-cap comparables (Alphabet $4,572.8B, Nvidia $5,140.2B, Microsoft $3,657.5B, Meta $1,498.0B) plus Taiwan Semiconductor and Sony, alongside Algorhythm Holdings at a $1.0 million market capitalisation, Turtle Beach at $276M and Nextpower at $14.9B. No peer multiples are supplied and no peer-multiple comparison is drawn.
• Fair-value caveat: the $230 / $326 / $390 anchors are multiples of FY2028 consensus EPS of $10.6768 (21.5x / 30.5x / 36.5x), cross-checked against FY2027 consensus EPS of $9.59026 (24.0x / 34.0x / 40.7x). Stated arithmetic, not a discounted cash flow. The base case assumes modest multiple compression from 35.1x FY2026 to 30.5x FY2028 and that assumption drives the result: at a held 35.1x on FY2028 estimates the price would be $374.8 (+21.2%), and at a held 29.0x it would be $309.6 (0.0%). Forward-multiple rolldown from 35.1x FY2026 to 29.0x FY2028 at a constant price is mechanical and is NOT counted as compression.
• Non-equity tripwire: not triggered. AAPL is common equity — beta 1.097, a $202.16-$344.57 annual range, 45.8 million shares traded on the day, a variable and rising dividend. No preferred or baby-bond characteristics.
• Timing: Q3 FY2026 was reported 2026-07-30 and the 10-Q filed 2026-07-31 — five and four days before this dive — so the freshest possible primary source carries the analysis. The next print is 2026-10-29 (consensus EPS $2.00, revenue $113.380B), 86 days away, so no earnings banner applies. The nearer dated event is the chief executive transition on 2026-09-01, 28 days away. The quote is the 2026-08-04 16:00 ET close and the stock rose 1.89% that session; that move is real, not a data error, and its implication for entry is addressed in Section 9.
• Known gap — earnings-call transcripts are NOT available on this plan. Apple's Q3 FY2026 call was held 2026-07-30 and management's commentary on the tariff refund, memory costs, the product cycle and the transition is absent. The 8-K of 2026-07-30 carries the earnings release, not the call. No management quotation appears in this dive and none has been invented.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.