Phase Space AI

03 Valuation and Factor Scorecard

Meta Platforms [META]

Meta Platforms, Inc. [META] — Valuation, Consensus Bridge & Factor Scorecard

Task 3 · as of 2026-07-27 · spot $593.87 · market cap $1,507bn · EV ~$1,513bn (incl. leases)

Model: META_Underwriting_Model_2026-07-27.xlsx — live formulas, verified by AppleScript readback with all workbooks closed first.


1. Model output (verified readback)

$m unless stated FY2025A FY2026E FY2027E FY2028E FY2029E FY2030E
Revenue $200,966 250,987 296,215 340,952 382,361 421,311
growth 24.9% 18.0% 15.1% 12.1% 10.2%
EBITDA (pre-D&A) $101,278 128,003 148,108 167,066 185,445 202,229
D&A (vintage engine) $18,000 24,900 39,201 55,907 74,100 92,769
Operating income $83,278 103,103 108,907 111,159 111,345 109,460
operating margin 41.4% 41.1% 36.8% 32.6% 29.1% 26.0%
Diluted EPS $27.43 $33.56 $35.46 $36.23 $36.50 $36.06
EPS at a 4-year server life $25.63 $31.02 $31.38 $30.40 $28.74 $26.30
Economic FCF (after SBC, after capex) 18,509 (25,023) (29,696) (25,841) (14,778) (1,924)

FY2025A ties to the 10-K exactly: revenue $200,966m ✓, PP&E depreciation $18,000m ✓, operating income $83,278m vs $83,276m reported (a $2m rounding residual).

The central line. EBITDA margin holds between 51.0% and 48.0% — the advertising business is not deteriorating. Operating margin falls from 41.4% to 26.0%. The entire 1,540bp of compression is depreciation, which rises from 9.0% to 22.0% of revenue. This is not a demand forecast. It is the arithmetic of $827bn of cumulative capex meeting a 5.5-year useful life.

Note also the shape of EPS: it rises to $36.50 in FY2029 and then turns down in FY2030. That inflection is the depreciation wall arriving faster than revenue growth can offset it.


2. Valuation

2.A DCF — and its honest information content

DCF bridge Value
Sum of PV of FY2026–30 unlevered FCF negative in every year
Fade period FY2031–35 (capex → maintenance) turns positive FY2032
Enterprise value $203,581m
DCF value per share $78.01
Terminal value as % of EV 177%

As with Amazon, terminal value exceeds enterprise value, so the DCF has low information content and is labelled as such rather than presented as a target. Assumptions: WACC 10.8% (CoE 11.09% from a measured beta of 1.52, rf 4.25%, ERP 4.5%; after-tax cost of debt 4.2%; D/(D+E) 3.8%), terminal growth 3.5%, maintenance capex 10.0% of revenue, SBC treated as a real cost (10.2% of revenue — materially higher than Amazon's 2.7%).

2.B Reverse DCF — the informative version

Solving for the terminal operating margin that makes the DCF equal $593.87:

WACC Required terminal operating margin vs FY2025A actual of 41.4%
6.5% 33.5% BELOW current — achievable
7.5% 41.3% ≈ exactly current
8.5% 50.1% +21% — never achieved
9.5% 59.9% +45% — implausible
10.8% ~72%

At a 7.5% discount rate, Meta at $593.87 requires a terminal operating margin of 41.3% — almost exactly the 41.4% it earned in FY2025. Today's price is therefore priced for "margins never decline from the 2025 peak, ever." At a 6.5% discount rate the requirement falls to 33.5%, which is genuinely achievable. At 8.5% it rises to 50.1%, which Meta has never approached.

Contrast with Amazon, which is the most useful comparison in this pair:

At an 8.5% WACC Required terminal operating margin All-time-high actual Multiple of the high
Meta 50.1% 41.4% 1.21x
Amazon 26.6% 11.2% 2.37x

Meta's price requires a 21% improvement on its best-ever margin. Amazon's requires a 137% improvement on its. On the reverse-DCF test — which is the closest thing to an apples-to-apples valuation comparison available — Meta is materially less demanding than Amazon, despite the depreciation wall being larger in percentage terms. That is a genuinely counter-intuitive result and it is the reason the two memos reach the same decision for different reasons.

2.C Scenarios (verified from the Scenarios tab)

Bear Base Bull
FY2030E revenue $362bn $421bn $489bn
Terminal operating margin 19.0% 26.8% 33.0%
Exit EV/NOPAT 15.0x 19.0x 24.0x
Value per share, discounted to today $212.68 $433.39 $775.62
Return vs spot −64.2% −27.0% +30.6%
Probability 35% 45% 20%

Probability-weighted value per share: $424.59. Scenario-weighted E[R]: −28.5%. Excess over the 4.7% cash hurdle: −33.2%.

Base-rate check per scenario: - Bear (35%) — the depreciation wall lands ($18bn → $74bn by FY2029) while ad growth normalises to high single digits and AI produces no separate revenue line. Margin compresses 41% → 19%. Precedent: Meta's own 2022, when operating margin fell from 40% to 25% in four quarters on a smaller spending programme. This is a reference-class-supported scenario, not a worst case invented for symmetry. - Base (45%) — revenue compounds ~15% and AI ranking gains carry roughly half the incremental depreciation. Margin settles at 26.8%. Note that even the base case implies ~1,460bp of margin decline from FY2025. This is not a flat-margin base case, and it should not be read as one. - Bull (20%) — ranking gains lift price-per-ad structurally and margin holds at 33%. Requires the above-base-rate assumption that a $180bn-a-year capital programme earns its cost of capital within five years. No public company in the reference class has done this. Named as an override and weighted at 20%.

2.D Consensus — sourced

Metric Consensus House Gap
Q2-2026 revenue ~$60.2bn (45 analysts) ~$60.0bn −0.3%
Q2-2026 EPS $7.13 – $7.23 ~$7.20 ~0%
Q2-2026 revenue growth +26.6% YoY +26% ~0%
FY2026 revenue $251.0bn (+24.9%) Consistent with the Q2 run-rate
FY2027 EPS ~$33 (implied from a reported ~18x on a $593.87 price) $35.46 House is +7% ABOVE consensus
Consensus price target ~$827 (63 analysts) Street is +39% above spot
Rating distribution 60 buy / 6 hold / 2 sell Near-unanimous bullish
Institutional ownership 79.33%

The decisive line: on FY2027 EPS the house number is HIGHER than consensus, not lower. A bear thesis requires the house to be below the Street on the metric the mechanism moves. It is above. This is a Gate 2A failure in the most direct possible way — and it is a failure that could only be discovered by actually sourcing the consensus number rather than assuming the Street must be too optimistic.

Caveat, stated: the FY2027 consensus EPS of ~$33 is inferred from a secondary source's "18x 2027 estimates" rather than pulled from a primary consensus feed (Alpha Vantage exhausted). It is directionally reliable but not precise, and the conclusion is stated with that qualification.

2.E Consensus bridge — numbers versus multiple decomposition (required)

House probability-weighted value $424.59 vs Street target ~$827 — a gap of −49%. Decomposed:

Component Contribution Explanation
Numbers — near-term revenue and EPS ~0% Q2-2026 revenue within 0.3%; EPS within ~0%.
Numbers — FY2027+ EPS ≈ −7%, in the WRONG DIRECTION for a bear House FY2027 EPS $35.46 vs consensus ~$33. The house is more optimistic.
Terminal margin / depreciation trajectory ~35% of the gap House models operating margin to 26.0% by FY2030; the Street's target implies materially higher.
Multiple / discount rate ~65% of the gap House exit 19x NOPAT (≈15.4x EBIT) at a 10.8% WACC. Meta currently trades at ~17.8x trailing EBIT.

Two-thirds of the disagreement is the discount rate and terminal multiple; one-third is the terminal margin; and on near-term numbers there is no disagreement at all — indeed on FY2027 the house is above the Street.

Per references/trade-construction.md, this is a valuation-and-terminal-assumption disagreement, not an earnings disagreement. It cannot carry a directional position.


3. Factor & Anomaly Scorecard

Alpaca SIP daily bars (251 trailing returns to 2026-07-27); SEC EDGAR XBRL and R-files.

Factor Value Empirical reading For a LONG For a SHORT
12-1 momentum −22.5% Clearly negative — bottom third but not bottom decile Negative Positive — a short would be WITH the tape
% off 52-week high −24.6% ($593.87 vs $787.42) Broken uptrend Negative Positive
vs 200-day MA −6.7% Below trend Negative Positive
YTD 2026 −9.9% Underperforming SPY (+9.0%) by ~19pp Negative Positive
Realised volatility (1yr) 38.8% annualised Moderate tier (25–45%), at the top of it
Beta vs SPY (measured, 251d) 1.52 High
Gross profitability (GP/A) ~0.45 (FY2025) Well above median — strongly favours long Positive Negative
Accruals (Sloan) (60,461 − 115,800)/366,021 = −0.151 Extremely negative — cash flow massively exceeds earnings. Among the cleanest possible readings Positive Strongly negative — argues hard against a short
Accruals — adjusted for the CAMT charge Normalising for the $15.93bn non-cash tax charge: −0.107 Still very clean Positive Negative
Piotroski F-score 7 / 9 (positive NI ✓, positive OCF ✓, OCF>NI ✓, ROA down ✗, leverage up ✗, current ratio up ✓, no dilution ✓, gross margin up ✓, asset turnover up ✓) Strong Positive Negative
Asset growth +18.2% (assets $309.5bn → $366.0bn) Elevated but well below Amazon's +27.5% Negative Mildly positive
Short interest ~1.64% of float, ~2.0 days to cover (dated May-2025, stale — flagged) No crowding Neutral

3.1 Reading the scorecard honestly — the tension is the point

Meta presents the single most contradictory scorecard in this coverage, and the contradiction is worth stating plainly rather than resolving artificially:

The reference file is explicit about how to handle this: "A Short whose scorecard shows clean earnings quality and strong profitability needs correspondingly stronger narrative evidence to pass this gate, and the memo must say so explicitly rather than ignoring the disagreement."

The narrative evidence here is genuinely strong — the useful-life extension worth 21% of FY2025 operating-income growth, the $28bn RVG, $45.99bn of unconsolidated exposure — but it is all PUBLISHED and already reflected in a 25% drawdown. Strong evidence that everyone already has is not strong evidence for the purposes of Gate 2.

Gate 1 therefore scores PASS on mechanism (the mechanism is named, evidenced and quantified — better than Amazon's) but the accruals/F-score corroboration standard is failed, and that is disclosed rather than ignored.


4. Comparable companies

Trailing reported figures from EDGAR XBRL. NTM consensus multiples are not shown — Alpha Vantage was exhausted and unsourced forward multiples will not be presented as sourced.

Company Price Mkt cap Latest FY Revenue Operating income Op margin EV / trailing EBIT
Meta (META) $593.87 $1,507bn FY2025 (Dec-25) $201.0bn $83.3bn 41.4% ~17.8x
Alphabet (GOOGL) $326.56 $4,065bn FY2024 (Dec-24)¹ $350.0bn $112.4bn 32.1% ~36.2x
Microsoft (MSFT) $389.10 $2,904bn FY2025 (Jun-25) $281.7bn $128.5bn 45.6% ~22.6x
Amazon (AMZN) $231.39 $2,488bn FY2025 (Dec-25) $716.9bn $80.0bn 11.2% ~31.1x
Netflix (NFLX) not pulled FY2025 (Dec-25) $45.2bn $13.3bn 29.5% — (context only)

¹ Alphabet FY2025 not retrievable under the standard XBRL tag; FY2024 shown and labelled.

Reading: Meta carries the LOWEST EV/EBIT in the mega-cap set (~17.8x) alongside the second-highest operating margin (41.4%). On a trailing basis Meta is the cheapest large-cap AI spender. That is a meaningful fact and it argues against a short far more than the depreciation arithmetic argues for one.

The counter, stated for balance: the 41.4% margin is the number the model says will compress to 26.0%, so trailing EV/EBIT of 17.8x becomes ~22x on FY2030 EBIT with no multiple change. Cheap on trailing, less so on terminal.


5. Football field

Method Low Base High
DCF (flagged: low information content, 177% TV) $78 $252 $352
Scenario probability-weighted $212.68 $424.59 $775.62
EV/EBITDA 9–13x FY2027E $512 $614 $716
P/E 17–23x FY2027E $489 $578 $667
Street consensus target range $680 ~$827 $950
Spot $593.87

The multiple-based methods bracket the spot price almost exactly ($578 and $614 against $593.87). The cash-flow methods sit far below. Meta is fairly valued on multiples and expensive on cash flow — the same structure as Amazon, but much less extreme.

House price target for the trade ticket: $578 — the FY2027E P/E midpoint — implying ~3% downside. A 3% gap is not a thesis in either direction, and the memo says so.