Deep Dive: Meta Platforms (META) — Full Dossier

Deep Dive #1 — first full run of the Deep Dive template end-to-end. Built Wed Oct 28, 2026 (pre-market). Prices as of the Tue Oct 27, 2026 close. Every figure carries its source; aggregator figures are labelled SECONDARY; anything not sourced is marked UNVERIFIED and is not load-bearing. Research and education — not personalised buy/sell advice.

1. Verdict

META — $620.19 (Oct 27 close, Nasdaq) · market cap ≈ $1.57T · Label: UNDERVALUED · IT factor: the Reality Labs AI-glasses platform, carried at zero.

Meta just printed the strongest revenue quarter in its history — Q3 2026 revenue $64.60B, +26% YoY, ads +26% on +10% impressions and +14% price per ad, Q4 guided $73–76B (+24% at midpoint) — and the market responded by taking the stock down 20% from its Sep 24 high ($777.59) because the bill arrived in the same print: capex of $34.9B in one quarter exceeded operating cash flow ($32.8B), free cash flow went negative (−$2.12B) for the first time in the AI era, a $9.89B legal charge cut GAAP EPS to $3.83 (ex-charge ≈ $7.40), and the company funded the gap six days later with a $37.5B raise ($25B of Class A at $625 + $12.5B mandatory convertible preferred). The result is the cheapest hyperscaler multiple on the board: ~19–21× forward earnings for the fastest mega-cap growth (+26%) — a ~30% discount to the hyperscaler peer median (27.1×) and a PEG of ~1.0 against peers at 1.3–3.4. On the evidence assembled here the multiple is not a bubble and not even full: it is UNDERVALUED versus its industry, because it already prices a sharp growth fade and a depreciation drag that the guided numbers (+24% Q4) do not yet show. The IT factor — the most exciting thing the market is overlooking — is Reality Labs' AI-glasses platform: a segment the market prices as a permanent ~$17B-a-year loss (Q3: revenue $443M, operating loss −$4.13B) while it quietly holds the only scaled consumer AI-glasses franchise in the world — the one candidate asset that could give Meta a hardware distribution layer of its own in the AI era, carried today at zero or negative value in every target price on the Street (average $848, built entirely on the ads engine). The tape, however, is not confirming the label yet: the last 7 sessions were decisive offloading (price −12.6%, signed volume −141.7M shares, short interest already covered down to 1.29% of float), institutions were accumulating as of the Q2 13F quarter, and the position question and the trend question currently have opposite answers — the dossier states both, dated, and does not merge them.

2. IT factor callout

THE IT FACTOR — Reality Labs' AI glasses: the next-platform option the market prices at zero. Reality Labs produced $443M of revenue against a −$4,129M operating loss in Q3 2026 (Meta Q3 2026 release, Ex-99.1, accession 0001628280-26-066700) — 0.7% of company revenue, and the segment's nine-month loss is −$12.78B. In every analyst model and every October target revision on file here, RL enters as one thing only: a cost line to be subtracted from the Family of Apps engine. Its glasses business — the only consumer AI-glasses franchise shipping at scale — is ascribed no platform value at all. Why it is overlooked: the P&L gives the market nothing else to see. RL revenue is small and was down 6% YoY in Q3 ($443M vs $470M), the loss is the single biggest line in the bear case ("the bill outruns the cash engine"), and Meta discloses no glasses unit numbers — so the debate is entirely about how big the loss gets, never about what the installed base could be worth. The materiality test: Meta's structural weakness since Apple's ATT changes is that it rents distribution from two gatekeepers. A glasses platform at even a fraction of phone scale converts RL from a −$17B/yr drag into Meta's own AI-era distribution layer — and removes the central line of the bear case in the same move. That is a story-changing event, not a quarter-changing one. What proves it: RL revenue inflecting to sustained double-digit growth (multi-$B annualised) with the loss narrowing while units ship — the first print where glasses revenue visibly bends the segment line. What kills it: RL losses widening past ~$5B in a quarter with revenue still under ~$0.5B through 2027 — then the option is a bonfire and the zero value was generous.

3. Valuation block

Label: UNDERVALUED (vs the industry, on earnings).

Peer forward P/Es (stockanalysis.com statistics, Oct 27 closes, SECONDARY). Set justification: GOOGL is the only like-for-like — the other half of the global ads duopoly. AMZN/MSFT/AAPL are the AI-capex hyperscaler set (same spend regime, platform earnings bases). SNAP/PINS/APP are ad-tech comps shown for context but excluded from the median — thin or distorted earnings bases at far smaller scale.

*GOOGL trailing P/E is distorted by equity-gains-inflated TTM EPS; forward is the cleaner comparator (flagged in the data file).

Own-history range: forward P/E by fiscal year (stockanalysis forecast, SECONDARY): FY21 24.69× · FY22 16.39× · FY23 20.78× · FY24 23.96× · FY25 22.59× · FY26E 20.61×. Today's forward multiple sits at the low end of Meta's own five-year range while growth is at the top of it.

Why UNDERVALUED, not FAIRLY PRICED: the multiple is below every hyperscaler peer, ~30% below the peer median, at the low end of its own history, with the fastest growth in the set and sector demand rising into share gains (Meta takes the #1 global digital-ad share in 2026F — §6). A "fair" price on these inputs is the peer-median multiple; the market is charging materially less. Why not a cash-flow label: on FCF the stock is not cheap — Q3 FCF was −$2.12B and FY26 capex is guided $135–145B. The label is an earnings-multiple label, stated as such; it dies if the earnings themselves are cut (falsifier below). Why not BUBBLE: the §5 bubble test fails on every leg — the multiple is below peers, below its own mid-range, and the price does not require demand to accelerate; it requires growth merely to not collapse.

Is the P/E rational? At ~19–21× forward earnings, the market is paying for Meta as if +26% growth roughly halves from here and as if the capex wave's depreciation lands squarely on 2027 EPS — i.e., the multiple already contains the bear case's arithmetic. Against that, the demand evidence is primary-sourced and current: ads +26% (impressions +10%, price +14%), DAP 3.64B, Q4 guided +24% YoY, and Meta overtaking Google for the #1 global ad share. A buyer at $620 is not paying for acceleration, perfection, or the glasses option (priced at zero); they are paying ~19× for the ads engine's decelerated future, while the actual prints keep arriving above the guide (11 of 11 guidance beats in the AI-capex phase — §5). The P/E is therefore rational-to-cheap as an earnings price; the honest caveat is that it is an expensive price as a cash-flow price until capex crests.

Scenarios (12-month, our scenario math on the labelled inputs above — FY27 EPS figures are scenario assumptions, not consensus):

Probability-weighted ≈ $718 (+16%). Swing factor: whether FY27 EPS absorbs the depreciation ramp from $135–145B of capex — the entire debate in one line.

Falsifier for the label: re-rate to FAIRLY PRICED/OVERVALUED if (a) FY27 EPS is cut below ~$30 on depreciation/talent costs, or (b) ad revenue growth falls under +15% for two consecutive quarters, or (c) the price re-rates to the ~27× peer median without earnings delivery. Re-rate the other way (deeper undervaluation confirmed) if Q4 prints at/above the $76B top end with FY27 capex guided flat.

4. Price chart & price structure

Chart (page): true daily line, META vs SPY indexed, 1,546 sessions (Sep 30, 2020 → Oct 27, 2026), with the three leadership-phase bands of §5. Full window: META +136.80% vs SPY +129.82% (price returns, dividends not reinvested — Nasdaq closes, prices_daily.csv). META outperformed across six years that included a −72% drawdown: the entire excess was earned in the reset and the AI-capex phase.

Price structure (as of Oct 27, 2026 close, computed from the verified daily series — tape_trend.json):

No 200DMA-rule backtest is run here: with one −72% regime break inside the window, a mechanical rule's output would describe 2022, not the current setup (method note, stated rather than buried).

5. Leadership eras — one founder-CEO, three phases

META is a single-CEO company: Mark Zuckerberg, founder-CEO throughout the window (and, per §7, controlling shareholder — there is no succession layer in this dossier). Segmentation is therefore by leadership phase, each boundary proven from a filing (full evidence in leadership.json). The one C-suite change: CFO David Wehner → Susan Li, effective Nov 1, 2022 (8-K Item 5.02, accession 0001326801-22-000085).

*P1's first four quarters carried qualitative-only revenue outlooks; dollar guidance ranges begin with the Q3 2021 release. UNVERIFIED by construction, never scored as beats (method: guidance_record.json). Overall scored record: 18 BEAT / 1 MET / 1 MISS of 20 — and the single MISS (Q2 2022, −0.61% vs midpoint) still landed inside the guided range.

Phase reads (judgement layer, every claim tied to the table): - P1 spent the pandemic ads windfall on the metaverse pivot and still lagged the index by 13pp; it ended on Feb 2, 2022, the night of the Q4 2021 release — the next session was −26.39%, then the worst drawdown in company history. Gross margin held ~81% throughout: the phase's problem was never the core engine, it was capital allocation and the ATT shock. Falsifier for a harsher verdict: none needed — the market's verdict is in the drawdown. - P2 is the discipline phase: revenue fell to $27.91B, EPS bottomed at $1.64 (Q3 2022), ~21,000 roles were cut across Nov 2022–2023, the "Year of Efficiency" was declared in the Q4 2022 release (accession 0001326801-23-000008, with a $40B buyback increase the same night), and margin expanded through the reset (GM 78.48% → 80.82%) while FCF rose to $62.89B. A +9.59% phase return that contains a −72.47% drawdown and a full recovery is two different companies: pre-Nov 2022 and post. - P3 is the spend phase: $190.0B of PP&E capex in 11 quarters (plus $7.26B finance-lease principal), revenue +77% first-to-last quarter, dividends initiated ($14.43B paid), guidance beaten 11 of 11 — and the phase still ends with the stock 20% off its high and FCF negative in its final quarter, because the market stopped paying for the revenue and started auditing the bill. Zuckerberg's earnings-call pattern in this phase: guide revenue conservatively, raise capex guidance almost every quarter (FY26 capex guide path in §6), and let the print beat the revenue guide — the stock now reacts to the capex line, not the revenue line (Oct 22, 2026: revenue beat, stock −6.98%).

Financial history (25 quarters, primary — full table in financials_quarterly.csv): revenue $21.47B (Q3 2020) → $64.60B (Q3 2026), a 3.0× in six years; GAAP gross margin 79–82% in every phase (Q3 2026: 79.31%); GAAP diluted EPS trajectory by phase is in era_scorecards.json (note the one-off quarters: Q3 2025 $1.05 — tax-charge quarter; Q1 2026 $10.44; Q3 2026 $3.83 — legal-charge quarter; trade the ex-charge line). Buybacks were $0 in Q3 2026 (10-Q tags $0 YTD for H2 — verified as real, not a data gap): the buyback has been stood down to fund capex and the raise.

6. Sector & bottleneck position

Wave position: Meta is not a layer in the AI stack — it is the demand at the end of it. Its ads sector (the money engine) is a mature, late-cycle duopoly; its AI position is that of the cycle's largest single buyer, sitting downstream of Wave 1 (compute) and pulling Waves 2–6 behind its capex. Sector stage: ads — late/mature; AI infrastructure demand — mid-cycle and still accelerating on Meta's own guides.

Market structure — oligopoly tightening into a duopoly-plus-one: global digital advertising, 2026 forecast (EMARKETER, Apr 2026, SECONDARY): Meta 26.8% (#1 — surpassing Google for the first time), Google 26.4%, Amazon 9.0%; the trio takes 62.3% of global digital ad spend. Meta's share is still rising (ad revenue +26% vs the market growing high-single-digits), powered by AI ranking/recommendation — impressions +10% and price-per-ad +14% in the same quarter is pricing power demonstrated, not asserted. Barriers: 3.64B daily users (DAP, Q3 2026 release), the advertiser tooling stack, and now the AI compute base itself.

Bottleneck test (on Meta's sector): digital ads are not a bottleneck — supply (attention inventory) is not capacity-constrained and the top-3 concentration gate is about sellers of attention, not a scarce input. Meta itself fails the bottleneck test and passes the bottleneck-customer test: its FY2026 capex guide of $135–145B (path during 2026: Jan $115–135B → Apr $125–145B → Jul $130–145B → Oct narrowed $135–145B; 2024 actual $37.26B, 2025 $72.22B — releases, primary) is one of the largest single demand signals feeding the actual bottlenecks.

Correlation to Bottleneck Intelligence (candidates.json, checked across all 13 candidates): META is not a candidate and this dossier does not propose adding it — it owns no scarce layer. The correlation runs the other way and is material to the module: Meta's capex is a demand pillar input for Wave 3 (packaging/substrates — CoWoS, ABF/T-glass), Wave 2 (memory/HDD), Wave 4 (optical), Wave 5 (power — transformers, turbines) and Wave 6 (cooling); only SRM-energetics and HALEU have no direct link. This Deep Dive therefore strengthens the demand pillar of those candidates (a named, primary-sourced $135–145B buyer still raising through Oct 2026) and adds a monitoring note: a Meta capex-guide cut would be an early demand warning for the whole bottleneck book — the falsifier travels in both directions.

7. Ownership — who owns it, who is doing what

The founder class is real here. Zuckerberg holds 639,347 Class A + 341,823,978 Class B shares = 99.8% of Class B (DEF 14A, 2026 proxy). Class B carries 10 votes per share, so the dossier's central ownership fact is a wedge: economic stake 13.49% — voting control 60.84%. There is no version of META governance in which the founder can be outvoted; every other class below is a passenger on that fact. (Sheryl Sandberg is verified not in the 2026 proxy table.)

The split (total shares 2,538,423,304, proxy Apr 1, 2026; sums to 100.00%):

Quarterly positioning trend (13F — quarterly, lagged up to ~45 days; Q3 2026 filings not due until Nov 14, 2026): institutions were ACCUMULATING in the latest reported quarter — share of Class A 74.63% → 76.96% Q1→Q2 2026, +58.82M shares, adders 2,679 vs reducers 1,896, new positions 290 vs sold-out 182 (Nasdaq). Only two quarters are honestly sourced; earlier quarters are UNVERIFIED, not filled (§12). Per-class indicators: Institutions: Accumulating (Q2 2026). Insiders: Net selling, planned — zero open-market purchases in 12 months across 163 parsed Form 4s; open-market sales ≈ 400,524 shares / ~$259.2M, predominantly 10b5-1 (largest: CFO Susan Li, 167,000 sh / $105.7M). Retail (est.): residual share fell as institutions added — direction: falling share of a rising institutional book.

Top holders — who did what (Q2 2026 13F): BlackRock +4,553,693 (+2.70%) to 173.40M — largest holder, adding. FMR/Fidelity −20,719,009 (−17.77%) — the big top-10 cut. Capital Research Global Investors +8,709,803 (+28.18%). Invesco +22.77M — the largest add overall. (Data-hygiene note in ownership.json: Nasdaq's aggregate double-counts Vanguard via a stale group row; combined Vanguard entities = 181.83M. No 13D/activist stake exists; passive 13Gs only.)

Last 7 days — tape signal (separate horizon, lower confidence; Oct 19–27, 2026, computed from the verified daily series): OFFLOADING. Price −12.57%; up-volume 29.2M vs down-volume 170.9M (signed balance −141.7M); A/D delta −70.2M, OBV delta −141.7M, both down. One Form 4 in the window: COO Javier Olivan sold 1,575 sh @ $707.20 on Oct 19 (filed Oct 21, 10b5-1 flagged); Form 144 proposed-sale notices (Alford, Olivan) Oct 26. Options flow in the window is two-sided noise around the print (§8). Quarter tape trend (13 weekly bars, tape_trend.json): heavy offloading in the Q2-reaction week (Jul 30: 42.3M shares, close −8% on the day), genuine accumulation Aug 24–Sep 11 (signed +55.0M, +54.7M in consecutive weeks) carrying price to the Sep 24 high — then the trend rolled over into the earnings collapse. Read: accumulation built the top; the last-7-days offloading is the unwind, and the quarter's net is distribution into strength. This tape trend will be scored against Q3 2026 13F positioning when filings land (~Nov 14).

Post-period capital events (primary): Oct 27, 2026 — 40,000,000 new Class A shares at $625 (FWP, accession 0001193125-26-432549) plus $12.5B of 6.50% mandatory convertible preferred (250M depositary shares @ $50, converting ~Aug 2029) — a $37.5B raise to fund capex. Because the equity leg is voting Class A, Zuckerberg's voting power dilutes slightly pro-forma: 60.84% → 60.41% (60.34% if the 6M-share greenshoe fills). Separately, Class C — a new non-voting class — is authorised via PRE 14C (filed Oct 21): created Jan 19, 2027 by a 1-for-1 stock dividend to A+B holders; the filing states explicitly it is voting-neutral. A planned up-to-$60B Class C ATM is press-sourced only (SECONDARY).

8. Options — simple view

As of the Oct 27, 2026 session (OptiView/OPRA delayed + StockNear chain sums, SECONDARY; full data in options_snapshot.json). Earnings are already out (Oct 21) — this is a post-print book, so positioning runs through year-end and January, not the front week.

• Calls vs puts, plainly: volume 573,859 contracts — calls 359,492 / puts 214,367 (P/C volume 0.60); premium P/C 0.83. Open interest is call-heavy: call OI 1,677,895 / put OI 803,843 (P/C OI 0.48). Max pain $645.
• Expiry concentration — the money is in January, not this week: the nearest two expiries (Oct 28 + Oct 30) hold only 6.7% of OI. Jan 15, 2027 alone holds 37.1%; Nov 20 16.8%, Dec 18 15.7%; expiries from Jan 2027 out (LEAPS) = 56.0% of all OI. Whatever the options market thinks about META, it has positioned it as a Q4-print/January story, not a this-week trade.
• Walls (scope-stated — two honest answers): front-expiry basis (SquawkFlow, Oct 27 close): call wall $650 / put wall $620, net GEX −$123.8M per 1%, gamma flip $632.50. Whole-chain basis: call wall $700 (41,444 OI at the strike), put wall $600. The stock closed at $620.19 — sitting on the front put wall, below the flip: near-term dealer hedging is a headwind until $632.50 is reclaimed.
• IV / expected move: ATM IV30 39.2%, IV rank 21 — low (post-earnings crush; the pre-print market priced ±6.5% and got −6.98%). Nearest dated straddle read (Oct 23 basis): Oct 30 expiry ±$35.15 / ±5.47%. No next-earnings expected move exists to quote — Q4's date is UNVERIFIED (§12); one aggregator's ±11.4% figure carries no stated expiry and is flagged UNVERIFIED, not used.
• Dated unusual flow (individually checked): Oct 27 — $630 calls (Oct 28 expiry) volume 22,358 (4.0× normal), $620 puts 12,995 (11.5×), $610 puts 9,198 (3.5×): short-dated two-sided hedging into the raise, no clean directional tell. Oct 22 — a $28.31M bearish-leaning 3-leg Jan-27 $750/$830 structure (33,873 contracts) the day after earnings: the one genuinely directional institutional-sized item of the week, and it is upside-capped, not outright short. No other dated TheFly/TipRanks/Barchart unusual item was found for Oct 21–28.

9. Dark pool

Real FINRA-reported off-exchange prints only (ChartExchange, SECONDARY), darkpool_snapshot.json. Modelled/simulated "dark pool" products were excluded by rule; one source encountered (an options-premium tracker) is called out in the data file as not equity dark-pool data.

• Level: Oct 27 off-exchange volume 24,918,805 shares = 48.19% of total volume (lit 51.81%) — below the 30-day average of 49.49%. Roughly half of META's volume never touches a lit exchange; this week, slightly less than usual did.
• Last 7 sessions: off-exchange share ranged 43.03%–52.01% with no monotonic trend — earnings day (Oct 21) printed the lowest dark share (43.03%, the print pulled flow onto lit venues) and the reaction day (Oct 22) the largest dark volume (~19.19M shares) only because total volume spiked to 40.57M. Derived weekly dark share: 49.9% → 52.0% → 49.7% → 47.5% → 46.8% (partial week) — a gentle drift down, not a distribution spike.
• Venue breakdown / largest prints: unavailable on free sources — the prints/levels tables returned empty and the ATS-vs-wholesaler split sits behind FINRA's lagged weekly files; nothing is invented to fill this (§12). Sole block evidence: MobyTick (≥50K-share prints, week ending Oct 23) shows $3.76B of META block value at an average $686.88 — size crossed above the current price during earnings week, consistent with institutional repositioning into the drop, not proof of it.
• Read, with its limits stated: prints alone don't prove direction. Combined with §7's tape (signed lit volume decisively negative) the honest sentence is: off-exchange activity was orderly and slightly light while lit-market flow sold — institutions crossed blocks near $687 average during the fall; retail-lit flow did the visible damage. Accumulation claims beyond that would exceed the data.

10. Short interest & borrow

short_interest.json (MarketBeat/Nasdaq series + ShortSqueeze cross-check, SECONDARY).

• Short interest: 28,417,942 shares (Oct 15, 2026 settlement) = 1.29% of float, 1.8 days to cover — down 15.35% from Sep 30 (33,572,792). Shorts covered into the earnings and the decline; this is not a crowded short and not a squeeze setup. Whatever is pressuring the stock (§7), it is long-side selling, not short-side attack.
• Borrow fee / shares available: UNVERIFIED. iBorrowDesk's feed is now behind a keyed API gate (403) and no figure is estimated in its place. The squeeze read therefore rests on SI + days-to-cover only: squeeze risk low on the sourced inputs.

11. Catalysts, risks & falsifier

Catalysts (dated): - Q4 2026 print (late Jan 2027, date UNVERIFIED) — the $73–76B guide test, FY27 capex guide, and the first depreciation read. The dossier's label lives or dies here. - Jan 19, 2027 — Class C creation (1-for-1 stock dividend; voting-neutral per PRE 14C) and any launch of the press-reported Class C ATM (SECONDARY). - ~Nov 14, 2026 — Q3 13F filings due: the tape-vs-positioning score for §7 (were institutions buyers of the earnings drop, as they were in Q2?). - Offering settlement/greenshoe (6M shares) and mandatory-convertible overhang into Aug 2029. - Analyst centre of gravity sits far above the price: average target $848.38, median $850 (n=62, Strong Buy consensus) (stockanalysis/S&P Global, SECONDARY, Oct 28) — October actions include Wells Fargo $1,000, Truist $1,000, Canaccord $950 (Oct 27), JPMorgan $940 / Goldman $885 / Barclays $880 / BofA $845 / Piper $850 (all Oct 22, post-print), against Scotiabank Hold $650 (Oct 27).

The live debate, one line each (attributed, dated): - Bull: "Ads compounding on AI: +26% ad revenue, Q4 guided $73–76B — the capex is already feeding the core" (Q3 release, Oct 21; Pivotal, Oct 13). - Bear: "The bill outruns the cash engine: capex > OCF, FCF negative, RL −$4.13B a quarter, buybacks at zero, funded by a $37.5B raise" (post-print recaps, Oct 22–27).

Risks, stated before the upside (per the standard): 1. Depreciation wall: $135–145B of FY26 capex (on top of $72B in 2025) converts into a multi-year depreciation ramp; FY27 EPS is where the bill lands. This is the label's falsifier #1. 2. Cash-flow regime change: FCF has gone from +$114B across Phase 3's first 10 quarters to negative; funding now includes dilutive equity (Class A at $625) and a mandatory convertible — the capital-return era (buybacks $0 in Q3) is paused, and if the raise repeats, per-share math degrades. 3. Concentrated legal/regulatory exposure: a $9.89B charge in one quarter (state AG youth settlement) demonstrates the tail size; further youth-safety/privacy actions are unquantifiable from filings. 4. Founder control is absolute: 60.84% voting on 13.49% economic — minority holders cannot force a capex course-correction; the governance discount question is permanent, and the Class C structure (non-voting shares for future issuance) extends it. 5. RL burn: −$12.78B in nine months with revenue down YoY; if the IT factor's kill-test triggers (§2), the segment is a pure drag with no offsetting option value. 6. Tape risk (near-term): offloading trend + price below both moving averages + dealer gamma flip above at $632.50 — the path of least resistance is down until the tape turns, whatever the label says.

Falsifier (what changes the verdict): the UNDERVALUED label is withdrawn if FY27 EPS is guided/cut below ~$30 at the Q4 print, or ad growth prints under +15% twice running, or FY27 capex is guided above $145B with no revenue-guide raise to match. The IT factor is withdrawn on its own kill-test (§2). Both are pre-registered (§12).

12. Limitations, UNVERIFIED items, sources

Pre-registration (DOSSIER-STANDARD scoring): this dossier's label (UNDERVALUED, at $620.19) and IT factor (Reality Labs AI-glasses platform) are registered in ~/workspace/deep-research/labels-registry.json on publication, review dates +6mo (2027-04-28) and +12mo (2027-10-28). A META capex-guide cut is additionally logged as a demand-warning watch input for the Bottleneck Intelligence book (§6).

UNVERIFIED / partial — nothing below is load-bearing: - Q4 2026 earnings date — late-Jan 2027 by pattern only; not confirmed from IR/EDGAR. - Q3 2026 consensus EPS basis — sources conflict ($6.76 vs $3.56); the revenue consensus ($63.30B, beat by $1.23B) is the one used. Forward P/E rests on aggregator forecast bases (18.83× statistics / 20.61× FY26E), both SECONDARY (S&P Global via stockanalysis). - Borrow fee & shares available — UNVERIFIED (iBorrowDesk keyed-gate 403). Squeeze read rests on SI/DTC only. - Institutional trend beyond two quarters — only Q1/Q2 2026 13F quarters honestly sourced; the institutional share total (~1,697M) is derived from published deltas, and holder counts conflict between sources (5,030 vs 5,056). Cross-check vs a second aggregator (13.37% insiders / 67.66% institutional) is disclosed in ownership.json; the split shown in §7 uses the proxy-primary insider figure. - Pension-fund class = 16 identified direct filers only — a floor, not the pension total. - Largest individual dark-pool prints + ATS venue split — unavailable on free sources; not estimated. - Options: front-expiry absolute OI conflicts across three sources (ratios agree, absolutes don't — ratios used); Citi's October target ($800 vs $895) and Truist's raise date (Oct 13 vs 16) conflict across aggregators; one ±11.4% expected-move figure lacks a stated expiry (excluded). - META own-history P/S range rests partly on a stale-basis aggregator average (~7.4× 5-yr) — used as context only. - Two $1M rounding seams (Q3 2026 FCF −$2,121M computed vs −$2,120M printed; FY25 quarterly revenue sum $200,965M vs 10-K rung $200,966M) — logged in the data files, unadjusted, immaterial. - Q1 2026 GAAP EPS of $10.44 and Q3 2025's $1.05 contain one-off items per the respective releases; TTM trailing P/E inherits that noise — the forward basis is the cleaner read, as flagged in §3.

Sources (as-of Oct 28, 2026): SEC EDGAR — META submissions + company facts (CIK 0001326801); 25 earnings releases (8-K Ex-99.1, incl. Q3 2026 accession 0001628280-26-066700); DEF 14A 2026 proxy; 163 Form 4 filings; FWP accession 0001193125-26-432549; PRE 14C (filed Oct 21, 2026); CFO-change 8-K accession 0001326801-22-000085; phase-boundary releases 0001326801-22-000015 / -23-000008 / -24-000010 / -24-000044. Nasdaq official daily history (META + SPY, 1,546 sessions, volume retained). SECONDARY, labelled wherever used: stockanalysis.com (multiples, forecasts, targets — S&P Global data), OptiView/StockNear/SquawkFlow/Metis (options), ChartExchange + MobyTick (dark pool, FINRA-reported), MarketBeat/ShortSqueeze (short interest), BusinessQuant/Nasdaq (institutional flows), EMARKETER Apr 2026 (ad-market shares). Bottleneck Intelligence candidates: ~/workspace/goals/supply-chain-bottleneck-intelligence-for-long-term-growth-picks/hidden_files/candidates.json (13 candidates, checked in full).

Data files in this folder (every figure reproducible): financials_quarterly.json/.csv (25 quarters) · prices_daily.csv (1,546 sessions, OHLCV) · leadership.json · era_scorecards.json · guidance_record.json (20 scored + Q4'26 pending) · tape_trend.json (daily A/D+OBV 13 wks, weekly bars, last-7 block) · ownership.json · short_interest.json · options_snapshot.json · darkpool_snapshot.json · catalysts_valuation.json · sources.json · build_meta_data.py (pipeline) · raw/ (all pulls).

Appendix (full per-quarter financial detail) ships as the data files above and the collapsed appendix on the dossier page.