Preparing your space
Your library and workspace are being arranged now.
Preparing your space
Your library and workspace are being arranged now.
Loading the business overview…
Explore the research to see how business evidence informs assumptions and valuation.
Start with the business evidence, then follow the assumptions through to scenarios and valuation.
Valuation and assumptionsWritten at USD 571.7000; now USD 648.03000000 (+13%). Upside to the weighted case has moved from -3% to -14%.
Meta Platforms is a wide-moat consumer internet platform undergoing a transition from asset-light advertising to capital-intensive AI infrastructure, whose economics are primarily driven by ad impression volume, average price per ad, and the level and duration of capital expenditure. Its principal competitive advantage is the combination of network effects across 3.6 billion daily users and a data-and-ranking advantage currently producing simultaneous 14% impression growth and 12% price growth, while its most important vulnerabilities are dependence on mobile operating systems it does not control for ad signal, a depreciation schedule from $400bn+ of 2026-2028 capital spending that will compress margin before revenue absorbs it, and a governance structure offering shareholders no recourse if the capital plan proves wrong. Under the base case, 12% revenue CAGR to $440bn by 2031 with terminal operating margin of 37% and capex plateauing near $145bn implies intrinsic value of approximately $520 per share versus a market price of $571.70. The investment therefore depends primarily on whether AI infrastructure spending generates revenue faster than it generates depreciation, a question no disclosed figure can currently answer and none will before 2028. The most important evidence to monitor is FY2027 capex guidance, operating margin excluding one-off charges, D&A as a percentage of revenue, and the quarter in which free cash flow turns decisively positive.
| Field | Value |
|---|---|
| Company | Meta Platforms, Inc. |
| Ticker / exchange | META / NASDAQ |
| Current share price | $571.70 (intraday, 2026-08-31; last confirmed close $578.02 on 2026-08-28) [S10] |
| Market capitalization | ~$1.47 trillion [Est., 2.565bn diluted shares × $571.70] |
| Enterprise value | ~$1.46 trillion [Est., net cash ~$6.6bn] |
| Industry | Digital advertising / consumer internet platforms |
| Primary geographies | US & Canada, Europe, Asia-Pacific, Rest of World |
| Investment classification | Platform |
| Business quality | High |
| Competitive advantage | Strong |
| Financial quality | Moderate (was Strong; degraded by the capex cycle) |
| Growth outlook | Strong |
| Balance-sheet risk | Low, trending Medium |
| Valuation attractiveness | Fair |
| Overall risk | Medium |
| Base-case intrinsic value | $520 |
| Bull-case intrinsic value | $800 |
| Bear-case intrinsic value | $330 |
| Probability-weighted value | $556.50 [Est.] |
| Upside/downside vs price | −2.7% |
| Expected annualized base-case return (3–5yr) | 7–10% [Est.] |
| Investment stance | Hold |
Why it may be attractive. The core advertising business is accelerating, not decelerating: revenue grew 33% in Q1 2026 and 28% in Q2 2026 [S6][S1], the fastest sustained growth since 2021, driven by 14% impression growth and a 12% increase in average price per ad [S2]. The stock is 29% below its August 2025 high [S11] while the operating business is larger and growing faster than it was at that high.
What the market may be misunderstanding. The market is pricing the capex cycle as a permanent impairment of cash economics. FCF fell 91% year-over-year to $784 million in Q2 [S3] and the stock made a 52-week low. But GAAP operating income excluding the $2.4bn legal charge and $1.18bn severance would have risen 9% [S2]. The question is whether the capex is a temporary depression of cash flow (bull) or a structural change in the capital intensity of the business (bear). I lean to a middle position, and that middle position is roughly what the price already reflects.
The most important value driver. Not revenue growth — the trajectory of capital expenditure from 2027 onward. Meta has guided to $130–145bn of 2026 capex [S1] against FY2025's $69.7bn. Whether that peaks in 2027 or compounds is worth more per share than any plausible variation in ad growth.
The largest risk. Depreciation. $130–145bn of 2026 spend plus a likely similar or larger 2027 number lands in the P&L over the following five years. If revenue growth decelerates to the low teens before that depreciation is absorbed, operating margin compresses structurally rather than cyclically, and today's 21× earnings multiple is on a peak-margin number.
What must be true. That AI infrastructure spending translates into revenue at a rate that outpaces its own depreciation schedule, and that the August 2026 settlement's mandated teen design changes [S7][S8] prove to be a 1%-of-revenue cost rather than the template for a global engagement-limiting regime.
But narrowing in one specific and important way, discussed at the end.
Meta's consumer brand equity is poor and worsening; the settlement coverage compares the litigation to the tobacco cases [S7]. Instagram and WhatsApp carry the brand value; the corporate parent carries the liability.
Regulation is a barrier to Meta, not for it. Interestingly, the settlement may invert this: commentary notes it positions Meta to box in TikTok and YouTube by setting a compliance standard others must now meet [Fortune, via S9 family]. A company with $250bn of revenue absorbs a compliance cost that a smaller competitor cannot. This is a real, if uncomfortable, moat-widening effect.
| Test | Result |
|---|---|
| ROIC above WACC | Yes historically, converging fast — see §11 |
| Stable/rising share | Probable [S2, management claim] |
| Pricing power | Yes — +12% price per ad [S2]. This is the strongest single piece of moat evidence in the report. |
| Gross margin | ~82% [Unverified] |
| Incremental margin | Negative in Q2 — expenses grew 55% on 28% revenue growth [S1] |
That last row deserves emphasis. Incremental margin was negative. For a company whose entire investment case has rested on operating leverage, an incremental margin below zero — even for identifiable one-off reasons — is the number that changed the stock's multiple, and correctly so.
| Risk | Type | Prob. | Financial impact | Horizon | Leading indicator | Mitigation | Priced in? |
|---|---|---|---|---|---|---|---|
| AI capex fails to earn its cost of capital | Business | 35% | −$150 to −$240/share | 3–6yr | Incremental ROIC; FCF recovery timing | Capex is discretionary and can be cut | Partially — this is most of the 29% drawdown |
| Depreciation compresses margin structurally | Business | 60% | −$40 to −$80/share | 2–4yr | D&A % of revenue | Revenue growth outrunning it | Poorly — consensus still models 38%+ margins [Unverified] |
| Advertising recession | Business | 25% in any 3yr window | −$100/share peak-to-trough | Any | SMB ad surveys; consumer confidence | Cost flexibility; no customer concentration | No |
| Further regulatory action on engagement | Business | 55% | −$20 to −$60/share | 1–5yr | New AG suits; EU rulings | Settlement sets a template others must meet | Partially — the $17bn is now known [S7] |
| Mobile OS signal degradation (ATT 2.0) | Business | 30% | −$60 to −$120/share | 1–4yr | Apple/Google privacy announcements | On-device modelling; first-party data | No — this is the most underpriced risk |
| Interface shift from feed to AI assistant | Business | 30% | −$150/share | 5–10yr | Time-spent trends; competitor assistant adoption | Meta's own AI investment | No |
| Supplier pricing power | Business | 70% | −$20 to −$50/share | 1–3yr | Escrow disclosures [S11]; component pricing | Custom silicon; JV structures | Partially |
| Rising leverage constrains flexibility | Financial | 30% | Option value loss | 2–4yr | Debt/EBITDA; issuance size | Net cash today | No |
| Multiple compression on capital intensity | Valuation | 45% | −$80/share | 1–3yr | Peer multiples; FCF yield | FCF recovery | Partially |
| Key AI talent departure | Business | 25% | Hard to quantify | 1–3yr | Public departures | Compensation | No |
| Reality Labs continues indefinitely | Capital allocation | 60% | −$50/share NPV | Ongoing | Segment guidance | Zuckerberg could stop it tomorrow | Yes |
| Governance: no shareholder recourse | Governance | 100% | Embedded | Permanent | — | None | Yes |
| Assumption | Value | Basis |
|---|---|---|
| Risk-free rate | 4.3% | [Unverified — I could not confirm the current 10yr Treasury yield] |
| Beta | 1.20 | [Est.] |
| Equity risk premium | 5.0% | [Est.] |
| Cost of equity | 10.3% | Derived |
| Cost of debt (pre-tax) | 5.4% | [Est.] |
| Tax rate | 17% | Management guides 15–17% for remaining 2026 [S2] |
| After-tax cost of debt | 4.5% | Derived |
| Debt weight | 5.4% | $83.7bn debt / ($83.7bn + $1,466bn) |
| WACC | 10.0% | Derived, rounded |
| Terminal growth | 3.0% | Slightly above long-run nominal GDP; justified by advertising's continued share gain |
| Diluted shares | 2.565bn | [Est.] derived from $15.85bn net income / $6.18 diluted EPS [S3] |
| Share count assumption | Flat — buybacks offset SBC dilution | [Est.] |
| Stock-based compensation | Fully expensed — included in the operating expense base, not added back | Deliberate |
| Reality Labs | Included as an ongoing cost at ~$18bn/yr declining to $10bn by 2031 | [Est.] |
SBC treatment matters and I want it stated plainly: Meta's SBC is likely $18–24bn/year [Unverified]. Adding it back, as many sell-side DCFs do, would raise my base case by roughly $70/share on its own. I do not add it back. Anyone comparing this valuation to another should check that assumption first.
All figures USD billions. FCFF = NOPAT + D&A − capex − ΔWC.
| Year | Revenue | EBIT | Margin | NOPAT | D&A | Capex | FCFF | PV @10% |
|---|---|---|---|---|---|---|---|---|
| 2026E | 250 | 83 | 33% | 68.9 | 32 | 137 | −36.1 | −32.8 |
| 2027E | 290 | 95 | 33% | 78.9 | 55 | 150 | −16.1 | −13.3 |
| 2028E | 330 | 115 | 35% | 95.5 | 78 | 150 | 23.5 | 17.7 |
| 2029E | 370 | 135 | 36% | 112.1 | 95 | 145 | 62.1 | 42.4 |
| 2030E | 405 | 150 | 37% | 124.5 | 105 | 140 | 89.5 | 55.6 |
| 2031E | 440 | 163 | 37% | 135.3 | 110 | 135 | 110.3 | 62.3 |
| 2032E | 472 | 175 | 37% | 145.3 | 115 | 135 | 125.3 | 64.3 |
| 2033E | 500 | 185 | 37% | 153.6 | 118 | 132 | 139.6 | 65.1 |
| 2034E | 525 | 194 | 37% | 161.0 | 120 | 130 | 151.0 | 64.0 |
| 2035E | 546 | 202 | 37% | 167.7 | 122 | 128 | 161.7 | 62.3 |
Rounding for false precision, base case ≈ $520.
Terminal value is 70% of total value. That is high but unavoidable for a company whose free cash flow is negative for two of the next three years. It also means the DCF is really a statement about 2035, not about 2027, and should be weighted accordingly against the multiples work below.
Holding WACC at 10.0% and terminal growth at 3.0%, the current price of $571.70 requires:
Any one of these closes the gap. That is worth noting: the market's price is not demanding heroics. It is demanding a slightly better version of what I already assume.
| Multiple | Meta @ $571.70 | 5yr own history [Unverified] | Peer group [Unverified] |
|---|---|---|---|
| P/E on FY2026E EPS ~$27 ex-tax-benefit [Est.] | 21.2× | 22–27× | Alphabet ~22×, S&P ~21× |
| P/E on FY2027E EPS ~$29 [Est.] | 19.7× | — | — |
| EV/EBITDA on FY2026E ~$115bn [Est.] | 12.7× | 12–16× | — |
| EV/Sales on FY2026E $250bn | 5.8× | 6–9× | — |
| Price/FCF on FY2026E ~$0 | n/m | 20–30× | — |
| FCF yield | ~0% | 3.5–5% | — |
| Price/FCF on FY2029E ~$62bn [Est.] | 23.6× | — | — |
The multiples tell a different story than the DCF, and the tension is instructive.
On earnings, Meta is cheap — 21× for a business growing revenue 28% with a wide moat is unremarkable, arguably attractive, and below its own history. On cash flow, it is unvaluable — there is no free cash flow to put a multiple on.
The gap between those two statements is the entire debate about this stock. Whether Meta at 21× earnings is cheap depends entirely on whether those earnings are real owner earnings (§10 suggests they may overstate by 25–30%) and whether the margin behind them is sustainable (§3 suggests not quite).
Adjusting: on my normalized owner earnings estimate of ~$19.50/share [Est., low confidence], Meta trades at 29× owner earnings. That is not cheap. It is also not expensive for this quality of asset. It is fair.
Should Meta trade at a premium or discount? A premium to the market on quality, moat and growth; a discount to its own history on capital intensity, regulatory overhang, governance, and the loss of the buyback. Net: roughly market multiple, which is roughly where it is.
Sum-of-the-parts [Est., illustrative only]:
| Part | Method | Value |
|---|---|---|
| Family of Apps | 22× ~$100bn normalized after-tax FoA earnings | ~$2,200bn |
| Reality Labs | Capitalized loss stream at −$14bn/yr for 5 years, discounted | −$50bn |
| Excess AI infrastructure | Undetermined — this is precisely what SOTP cannot value | ? |
| Net cash | +$7bn | |
| Implied | ~$2,157bn = $840/share |
I do not believe this number, and I include it to show why SOTP is the wrong tool here: it values Family of Apps at a normalized earnings level that assumes the capex problem away. The gap between the $840 SOTP and the $520 DCF is the capex question, priced. The market at $571.70 sits much closer to the DCF than the SOTP, which suggests the market is taking the capital intensity seriously rather than ignoring it.
Earnings power value (no growth, current owner earnings capitalized at 10%): $50bn / 0.10 = $500bn = $195/share. So roughly two-thirds of the current price is growth. For a company of this quality that is not alarming, but it is worth knowing.
What goes wrong: The AI capex is maintenance, not growth. Component pricing continues rising; 2027 capex guides to $170bn+. Depreciation compounds faster than revenue. An advertising recession arrives in 2027–28 while capex commitments cannot be cut fast enough. The teen-safety settlement proves to be the first of several, and the EU restricts behavioural targeting. Growth decelerates to high single digits by 2029.
Most likely trajectory: Revenue growth decelerates from 28% to the mid-teens by 2028 and to high single digits by 2031. AI ad ranking continues delivering price-per-ad gains for another 2–3 years, then normalizes. Capex peaks around $150bn in 2027–28 and plateaus. Depreciation compresses operating margin from ~37% underlying to a trough of ~33% in 2027, then recovers to 37% as revenue outgrows it. FCF turns positive in 2028 and reaches ~$110bn by 2031. The settlement costs roughly 1% of revenue as commentary suggests [S9] and does not repeat at scale. Reality Labs continues to lose money but shrinks. Buybacks resume in 2028.
What must go right: The AI investment produces a genuine second revenue leg. Agentic commerce and business messaging together reach $60bn+ by 2031 at high margin. AI creative tools materially expand the advertiser base rather than just the ad load, adding demand rather than supply to the auction. Capex peaks in 2027 at $155bn and declines thereafter as custom silicon and the JV structure reduce unit costs. Operating margin recovers above 40% by 2030 because the infrastructure is shared across advertising, messaging, and assistant products. The settlement's compliance burden disproportionately damages TikTok and YouTube [per S9 commentary], and Meta gains teen share.
| Scenario | Value | Probability | Weighted |
|---|---|---|---|
| Bear | $330 | 0.25 | $82.50 |
| Base | $520 | 0.45 | $234.00 |
| Bull | $800 | 0.30 | $240.00 |
| Total | 1.00 | $556.50 |
Versus $571.70: −2.7%.
Note the distribution's shape. The weighted value is above the base case because the bull case is further from base than the bear is. This is a genuinely convex situation — the downside is a good business at a lower multiple, the upside is a new category. That convexity is worth something that a single point estimate does not capture, and it is why my stance is Hold rather than Avoid despite a slightly negative expected value.
3.6bn daily users, demonstrated pricing power, no customer concentration, ~37% underlying operating margin. Marked down from 9 because the marginal-cost structure is changing as AI inference enters the cost of an impression.
Wide moat from network effects, scale and data. Marked down because the most durable component (the social graph) is becoming the least relevant as recommendation quality displaces social connection, and because the data advantage is contingent on regulatory and OS permission rather than ownership.
Exceptional product judgement, poor financial judgement. Buybacks at highs and none at lows; $90bn+ into Reality Labs; three consecutive upward capex revisions. No shareholder recourse.
Net cash, investment grade, and roughly 3–5 years of severe-stress runway. Trending down: debt tripled in 18 months, FCF is zero, and $10.8bn of cash is escrowed to suppliers until 2028–2030.
28% revenue growth with real evidence of an AI-driven mechanism behind it. Constrained by a ~30% share of global digital advertising and by a Q3 guide implying sharp deceleration.
21× earnings is fair, not cheap, once earnings quality is adjusted. Base case sits 9% below price; weighted value sits 3% below.
Genuinely convex — bull $800, bear $330 — but with a slightly negative expected value at today's price.
1. Is this a good business? Yes, and by most measures an outstanding one. It is becoming a different business: a capital-intensive one with the multiple of an asset-light one.
2. Stronger or weaker? Operationally stronger — revenue is accelerating and pricing power is demonstrably improving. Financially weaker — FCF has gone to zero, debt has tripled, and returns on incremental capital are currently approximately zero. Both statements are true and the tension between them is the investment case.
3. Durable competitive advantage? Yes. Wide moat. But the durable part (the graph) is losing relevance and the relevant part (the data and ranking advantage) is less durable because it depends on permissions Meta does not control.
4. Three most important value drivers? (i) The 2027–2029 capital expenditure plateau level; (ii) long-run operating margin after depreciation absorption; (iii) the discount rate.
5. Three largest risks? (i) AI capex failing to earn its cost of capital; (ii) structural margin compression from depreciation; (iii) mobile OS signal degradation — the underpriced one.
6. Most important dependency? Apple and Google's control of the mobile operating systems and the ad signal layer. Meta says so itself [S2]. It is the only dependency that has previously removed ~$10bn of revenue overnight.
7. What is the market pricing in? Revenue roughly doubling to $470bn by 2031, operating margin stabilizing at ~36% rather than continuing to fall, and capex ceasing to grow by 2028. All three, and not one of them is unreasonable.
8. Where could the market be wrong? Most likely on long-run margin, by underestimating the depreciation from $400bn+ of 2026–28 capex. Less likely but more consequentially, on the probability of a new OS-level signal restriction.
9. Fair intrinsic value? $520 base case; $556 probability-weighted; a defensible range of $450–$620.
10. Margin of safety today? None. Price is above base-case value. On the probability-weighted number the margin is −3%, which is inside the error bars of the analysis and should be read as "fairly valued," not as "overvalued."
11. Reasonable annualized return over 3–5 years? ~9–10% probability-weighted, of which approximately 13% comes from fundamentals and −3 to −6 points from assumed multiple compression. Bear −6%, bull +20%.
12. Single development that would most strengthen the thesis? FY2027 capex guidance at or below $145bn. It would convert the story from an open-ended capital commitment into a defined cycle with a visible end, and it is scheduled for January 2027.
13. Single development that would most invalidate it? Operating margin ex-charges below 30% for four consecutive quarters while capex remains above $120bn — the signature of capital intensity having become permanent rather than cyclical.
14. At what price does it become clearly attractive? $430 — a 17% discount to base case and roughly 22% below the weighted value, which is the margin of safety appropriate to a business with this much unresolved capital-allocation uncertainty.
15. What price implies excessive optimism? $780 — approaching the prior high and implying the bull case is the base case.
Meta Platforms is a wide-moat consumer internet platform undergoing a transition from asset-light advertising to capital-intensive AI infrastructure, whose economics are primarily driven by ad impression volume, average price per ad, and — now decisively — the level and duration of capital expenditure. Its principal competitive advantage is the combination of network effects across 3.6 billion daily users and a data-and-ranking advantage that is currently producing simultaneous 14% impression growth and 12% price growth, while its most important vulnerabilities are its dependence on mobile operating systems it does not control for ad signal, a depreciation schedule from $400bn+ of 2026–2028 capital spending that will compress margin before revenue absorbs it, and a governance structure that offers shareholders no recourse if the capital plan proves wrong. Under the base case, 12% revenue CAGR to $440bn by 2031 with terminal operating margin of 37% and capital expenditure plateauing near $145bn implies intrinsic value of approximately $520 per share versus a market price of $571.70. The investment therefore depends primarily on whether AI infrastructure spending generates revenue faster than it generates depreciation — a question that no disclosed figure can currently answer and none will before 2028. The most important evidence to monitor is FY2027 capital expenditure guidance, operating margin excluding one-off charges, D&A as a percentage of revenue, and the quarter in which free cash flow turns decisively positive.
All of the above are readable in the Q2 2026 10-Q, the FY2025 10-K, and the 2026 proxy. This report should be re-run against those documents before any capital is committed.
The split between growth and maintenance capital expenditure is the weakest link and is unsupported by evidence. Also unresolved: server depreciation lives and whether they have been extended; stock-based compensation magnitude; Q2 2026 geographic and segment revenue mix; accounting treatment and timing of the $17-18bn settlement accrual; magnitude of off-balance-sheet infrastructure purchase commitments behind the $10.8bn escrow; competitor financials on a comparable basis; management compensation structure. All are readable in the Q2 2026 10-Q, FY2025 10-K and 2026 proxy, none of which were available for this analysis.
These model scenarios depend on their inputs. Review the assumptions and supporting evidence alongside the implied values.
earnings multiple v2 · as of 2026-09-12
Sign in for the price chart, where every input traces back to a filing, an adjustable scenario sandbox, and companies that move with META.
Create a free account →