Full Report

Know the Business

Meta is an attention-to-advertising monopoly: one business unit (Family of Apps) reaches 3.58B daily users and contributes $102.5B of operating profit at 52% margins, funding a second unit (Reality Labs) that loses $19B a year. The lever that actually matters right now is whether 2026's $115–135B capex translates into durable AI-driven ad-pricing gains before incremental depreciation lands on the P&L. The market is probably pricing the AI ad uplift correctly, underestimating how much Reality Labs is pivoting from VR burn to wearable optionality, and underestimating the 2026 margin compression as server depreciation ramps.

1. How This Business Actually Works

Meta is not a "social network" for the purposes of valuation. It is a two-sided auction: the supply side is 3.58B daily users generating ~12% more ad impressions each year; the demand side is ~10 million advertisers bidding for those impressions, with AI-driven targeting pushing the price per ad up ~9% annually. Revenue grows at roughly impression growth × price growth — FY2025 delivered exactly that (+12% × +9% ≈ +22%).

The two levers move independently. Impressions grew even in 2022 when price/ad collapsed 16% — that year's recession was a pure pricing story, driven by Apple's ATT framework obliterating ad measurement and advertisers pulling budgets. Post-2023, AI ranking (Advantage+, Lattice) restored price/ad growth; Reels monetization catch-up is the tail wind still pending.

FoA Revenue ($M)

$198,763

FoA Operating Income ($M)

$102,471

Reality Labs Op Loss ($M)

-$19,188

Family DAP (Billion)

3.58

No Results

The real economic engine is monopolistic: FoA owns ~23% of global digital ad share, competitor ad networks cannot replicate 3.58B cross-product identity graphs, and the marginal cost of serving an extra ad is approximately zero once the data centers exist. R&D spend is the bottleneck — $57.4B in FY2025, +31% YoY — because AI model training requires ever-more expensive compute. Bargaining power is split: Apple (iOS ATT), EU regulators (DMA consent mandates), and Alphabet (Android signals) each squeeze a slice of the targeting stack. The "moat" is not technology; it is two decades of social-graph data that only WhatsApp, Instagram, Facebook, Messenger, and Threads see. Threads alone (~400M MAUs) would be a top-5 social network.

2. The Playing Field

Meta is the fastest-growing of the mega-cap cohort and trades at a meaningful multiple discount despite the second-highest operating margin.

No Results

Only Microsoft earns a higher operating margin, and only at half Meta's growth rate. Alphabet is the only true comp — both are ad-duopoly, both serve global reach — but Alphabet's margin structure is ~10 points lower (YouTube content costs, Cloud at sub-scale margins, Other Bets losses). Amazon's ad business would fit in this peer set, but its consolidated financials drown in commerce mix. SNAP and PINS are tiny reference points — both struggle to monetize their audiences anywhere near Meta's ~$57 ARPP. The peer set reveals one thing clearly: the premium businesses in digital ads are the ones with a closed identity graph. Pinterest and Snap do not have one; that is the full explanation for their margin profiles.

The "advantage" Meta shows over Alphabet is cost discipline and auction density — it earns better incremental margins on each ad dollar because its feed format is pure direct-response (vs YouTube's mix of brand and DR). The "weakness" vs Alphabet is regulatory exposure: Meta has no search fallback revenue and 98% of its top line depends on a targeting stack that EU DMA rulings could meaningfully reshape.

3. Is This Business Cyclical?

Digital advertising is cyclical on price, not volume. Impression supply grows almost mechanically with engagement; ad price is the cycle variable.

FY2022 is the template for the cycle: price/ad fell 16% while impressions rose 18%, producing a -1% revenue year. The drivers were unique-but-instructive — Apple ATT cut signal quality, macro inflation squeezed DR advertisers, and TikTok short-form video split attention and priced its inventory to win share. Revenue recovered fast once AI-driven measurement tools (Advantage+, conversions API, Lattice) filled the ATT gap. The operating line is more volatile than revenue: operating income fell from $46.8B (FY2021) to $28.9B (FY2022), a 38% drop, because fixed R&D and infrastructure costs don't flex with ad pricing.

The **2026 cycle risk is internal, not external**. 2026 capex guidance of $115–135B — a $45–65B YoY increase — will drive material depreciation expense and SBC growth. Management has already signaled "2026 expense growth above 2025 expense growth." If ad pricing decelerates even slightly while cost growth stays elevated, operating margins compress meaningfully.

Working capital is not a cycle driver here — Meta is deferred-revenue heavy and has near-zero DSO. Capex is the cycle driver now: 2026 capex of up to $135B against FY2025 operating cash flow of $115.8B means free cash flow conversion compresses sharply regardless of top-line trajectory. This has already started — FCF fell from $52.1B in FY2024 to $43.6B in FY2025 despite 22% revenue growth.

4. The Metrics That Actually Matter

Forget EPS headlines and total revenue. Five numbers explain value creation here.

No Results

Price per ad (YoY %) — the cleanest single read on the business. When this number is positive, Meta's targeting stack is working and AI investment is earning its keep. When it turns negative (2022), no amount of impression growth saves the operating line. Better than DAU/engagement ratios because it measures monetization quality, not just reach.

Family DAP — the supply-side ceiling. At 3.58B already, growth must come from sessions per user, ads per session, and format shift (Reels) — not new humans. Watch the YoY rate: if it drops below +4%, incremental ad dollars have to come entirely from pricing, which raises execution risk.

ARPP — the monetization quality score. $57/year worldwide, $16.6/quarter in Q4 2025. Compare to Snap (~$11), Pinterest (~$8) to see the moat. Rising ARPP + rising DAP = the only combination that matters.

Reality Labs operating loss — the discretionary burn. FY2025: -$19.2B. Management guided "similar" in 2026. This is the line to watch if AI capex disappoints — RL is where spending gets cut first. Framing is shifting from VR (Quest decline) to wearables (Ray-Ban Meta, Display + Neural Band): ~70% of 2026 RL opex is wearables-directed.

FCF conversion — the capex discipline read. FY2024: $52B FCF on $69B op income = 75% conversion. FY2025: $44B / $83B = 52%. If 2026 conversion drops below 35%, the AI-spend / returns debate goes from theoretical to binding.

Superficial ratios to ignore: P/E (noisy because of OBBBA tax-rate volatility — FY2025 ETR was 30% headline, 13% ex one-time charge), and gross margin (meaningful only as R&D-ratio, since "cost of revenue" excludes the real cost base).

5. What I'd Tell a Young Analyst

Three things to watch, in priority order.

First, watch price-per-ad every quarter, not headline revenue. Meta's moat is pricing power on an ad auction, not user count. If the sequence breaks — impressions up, price down — the AI story is failing faster than the capex is slowing.

Second, assume Reality Labs is $20B/year of permanent opex, not an investment. Do not capitalize it in your model. The market keeps pretending wearables will inflect; seven years of evidence say otherwise. If RL ever shrinks materially, that's upside — don't underwrite it.

Third, the 2026 capex cliff is the biggest risk and the biggest fight. $115–135B in 2026 vs $72B in 2025 means roughly $10–15B of incremental depreciation flows through in 2026–2027, and more after. If Advantage+/Lattice AI pricing gains don't outrun that, margins compress even with 20%+ revenue growth. This is the one number that would change a thesis.

What the market may be underestimating: WhatsApp monetization (paid messaging + business API), which grew FoA "Other" revenue +50% in FY2025 off a small base and has a genuinely different margin profile than advertising. Thread ad rollout (barely monetized) and smart-glasses optionality are lottery tickets — worth tracking but not underwriting.

What the market may be overestimating: the EU DMA/DSA ring-fence. Meta itself warned the Less Personalized Ads mandate could have "significant negative impact on European revenue"; Europe is ~24% of revenue. Also overestimated: how much direct AI revenue exists — there is none. Meta AI has ~600M MAUs and zero dollars of direct revenue attached to it.

Thesis-changer tests: (a) price-per-ad goes negative for two consecutive quarters, (b) DAP growth falls below 4% YoY, (c) Reality Labs loss widens past $22B/year, (d) EU LPA enforcement takes effect unmodified. Short of those, this remains a business where every year spent worrying about TikTok or the metaverse has been wrong, and the simplest model — impressions × price × DAP × ARPP — has been the right one.

The Numbers

Meta is a $1.3 trillion advertising machine whose economics remain extraordinary: 41% operating margins, 30% net margins, and an 82% gross margin. But the single metric that will decide the next rerate is capex. FY2025 capital spend roughly doubled to $69.7B as Meta builds out AI compute — a step-change that has compressed free-cash-flow margin from 33% to 23% in one year. Revenue growth has accelerated (+22% YoY) and profitability is still peer-leading, yet the stock trades at only 26x trailing earnings and 19x forward — the cheapest P/E in Big Tech. The market is paying for the cash today but discounting the bill that arrives tomorrow.

Where Meta stands right now

Share Price (Apr 8, 2026)

$612.42

Market Cap ($B)

1,339

Revenue TTM ($B)

201.0

Operating Margin

41.3%

The stock is trading at $612, roughly 23% below its 52-week high of $794 and below both its 50-day ($638) and 200-day ($684) moving averages. Consensus analyst target sits near $860 — a disconnect the rest of this page tries to explain.

Quality scorecard — is this a well-run business built to last?

No Results

Meta sits in the top decile of US large-caps on balance-sheet safety (Altman Z of 7.4), interest coverage (90x), and profitability (41% op margin, 30% ROE). The one watch item is Piotroski F at 5/9 — dragged down by rising leverage, falling asset turnover, and a compressed current ratio, all direct consequences of the FY2025 AI capex surge, not operational decay.

Revenue and earnings power — the 20-year view

The economic shape: a 40-something percent operating margin was the long-run norm from 2016 through 2021, then collapsed in 2022 under the weight of Reality Labs losses and Apple's ATT-driven ad headwind. Margin recovery since has been dramatic — the FY2023-24 "Year of Efficiency" restored 42% op margins despite ~30% ad-revenue growth. FY2025's 41% op margin on $201B of revenue means this business is earning more profit each year than all but three or four companies on earth.

Quarterly revenue — no slowdown

Q4 2025 revenue of $59.9B was the largest quarter in Meta's history and grew 24% YoY. Every quarter since Q3 2023 has shown double-digit YoY growth, with acceleration through 2025 — the opposite pattern of a mature business.

Earnings surprise history — consistent beats

Twelve straight quarterly beats, averaging +13% above consensus. Management has been consistently sandbagging guidance post-2022 reset.

Cash generation — are the earnings real?

Operating cash flow has exceeded net income every year for a decade — the classic signature of a capital-light, depreciation-heavy advertising model where SBC adds back the non-cash wedge. In FY2025, OCF of $115.8B was 1.9x reported net income, padded partly by a $19B Q3 2025 tax charge (a one-time accrual under new US tax legislation) that depressed GAAP earnings without affecting cash.

Capex and free cash flow — the single chart that explains the derate

Capex jumped from $37B to $70B in a single year — the steepest increase in the company's history. FCF margin fell from 33% in FY2024 to 23% in FY2025 despite record revenue and operating profit. If capex continues rising in FY2026 (analyst expectations are $80–$100B), FCF could flatline or decline even as earnings grow. This is the asymmetry the 19x forward P/E is pricing.

Capital allocation — buybacks, dividends, debt

Meta has returned $31.6B to shareholders in FY2025 (buybacks + dividends), its second-largest return year after FY2021's $44.5B. Stock-based compensation of $20.4B is rising sharply — roughly 34% of the return of capital is being offset by dilution through SBC. Net buyback (repurchases minus SBC) is closer to $6B.

Balance sheet — flexibility is intact, but leverage is rising

Gross debt has climbed from near-zero in 2018 to $83.9B today, including $25B of capital lease obligations (mostly data-center commitments). Cash plus short/long-term investments stand at $81.6B, putting net debt at effectively zero. Debt/EBITDA of 0.8x remains conservative versus peers at 0.5–1.2x. The FY2025 bond issuance funded the AI build without touching cash reserves — a conscious decision to preserve dry powder.

Diluted share count has fallen from 2,956M in 2017 to 2,574M in 2025 — a 13% reduction over 8 years, notable for a high-SBC company. FY2025 GAAP EPS of $23.46 is flat YoY only because of the $19B one-time tax charge in Q3 2025; adjusted annual EPS of $29.70 implied +24% growth.

Valuation — now vs its own history

P/E Trailing

26.1

P/E Forward (2026E)

19.3

EV / EBITDA

13.8

Trailing P/E of 26.1 sits almost exactly on the 10-year median of ~24. Forward P/E of 19.3 is well below that median — implying the market expects 2026 earnings to grow while the multiple stays flat. EV/EBITDA of 13.8 is actually below the historical 5-year average of ~16, making Meta the cheapest mega-cap tech name on that measure.

Peer comparison — the cheapest profile in Big Tech

No Results

Against the three mega-cap peers, Meta has the lowest forward P/E (19.3 vs 19.7–25.8), the fastest revenue growth (22% vs 14–18%), and an operating margin only Microsoft surpasses. The only metric where Meta lags is P/S (7.7), and that reflects Amazon's genuinely different business (low-margin retail) and Alphabet/Microsoft's structurally different growth runways (search monopoly, enterprise software). On quality-adjusted value — margin × growth ÷ multiple — Meta screens as the cheapest name in the cohort.

Fair value scenarios

No Results

At the current $612, Meta is priced almost exactly at the base case. Consensus analyst target of $860 implies a 40% upside but assumes the bull scenario of margin expansion AND capex moderation. The risk/reward is asymmetric only if you believe capex peaks in FY2026–27 and delivers a revenue pickup after.

What the numbers say

The data confirm that Meta's ad business is as powerful as it has ever been: 41% operating margin on accelerating revenue, 30% ROE, near-zero net debt, and a decade of OCF exceeding net income. The popular narrative that Meta is "uninvestable because of Reality Labs" is contradicted outright — the core ad engine produced $83B of operating income in FY2025, dwarfing any metaverse loss. What's genuinely changed is the capex line: free cash flow margin has nearly halved in one year because Meta is spending $70B building AI compute that, today, generates no direct revenue. The metric to watch next quarter is FY2026 capex guidance — if it lands at $80B, the current forward P/E of 19x looks like the low; if it lands above $100B without an accompanying revenue-acceleration signal, the market will likely derate the stock further despite rising GAAP earnings.

Governance Grade: C+

Meta is a founder-autocracy wrapped in a 15-person independent board. Mark Zuckerberg owns roughly 13% of the equity but controls 61% of the vote, and used that control in 2025 to reject — again — a Class A shareholder proposal for one-share-one-vote that had drawn 84% support. The operating executives are deeply embedded and highly paid; the board is freshly stacked with Trump-aligned political allies (Dana White, Dina Powell McCormick) and a commercial counterparty (Broadcom's Hock Tan, who is now exiting after Meta paid his company $987M in 2024). Insiders sold; none bought. Skin-in-the-game is real but concentrated in one man who cannot be outvoted.

The People Running This Company

No Results

The operating bench is long-tenured and internal. Li, Cox, Olivan, and Bosworth each joined Meta between 2005 and 2008; their equity grants now vest in their early forties. That continuity is the strongest argument against succession risk — and the strongest argument that the inner circle is culturally loyal to Zuckerberg rather than independent of him.

Two personnel moves in the last twelve months matter more than anything on the org chart. In January 2026, Meta elevated Dina Powell McCormick — Trump's former deputy national security advisor, ex-Goldman partner, wife of Republican Sen. Dave McCormick — from director to President and Vice Chairman, a newly-created role directly under Zuckerberg. Her appointment followed Meta's January 2025 shift away from third-party fact-checking and preceded its hiring of Curtis Mahoney (Trump's former deputy USTR) as Chief Legal Officer. The governance signal is unambiguous: political access has become a C-suite competency at Meta.

**Succession risk is binary.** Zuckerberg controls 61% of votes and there is no public succession plan. The 14 independent directors cannot remove him. If he were incapacitated, his Class B shares would transfer through the Mark Zuckerberg Trust (2006) and CZI entities — all of which he is the sole trustee of.

What They Get Paid

Zuckerberg's pay is an accounting artifact. He takes $1 in salary and no equity — because he already owns the company. His $27.2M "compensation" is actually a corporate security budget that includes $10.4M for personal security at his residences and during travel, a $14M pre-tax allowance for additional family security, and $2.6M for personal use of a private aircraft that he himself owns and that Meta charters from him. In other words, Meta is renting its founder's plane to protect him from the consequences of being Meta's founder. The audit committee reviews it annually; shareholders have no real recourse.

The four other named officers each collect ~$23–26M, at the 75th percentile of a peer set that includes Alphabet, Amazon, Apple, Microsoft, and Nvidia. In February 2025 — the same week Meta laid off roughly 5% of its workforce for "low performance" — the compensation committee voted to raise executive target cash bonuses from 75% to 200% of base salary, lifting target cash pay from the 15th percentile to the 50th of the peer group. The optics were ugly; the economics are defensible because total pay was already back-loaded into equity.

The committee set the 2024 **company performance percentage at 125%**, paying bonuses above target. The "metric" is explicitly subjective — there are no formulaic revenue, margin, or ROIC gates. Compensation is whatever the committee thinks it should be, within peer norms.

Are They Aligned?

Ownership and control

The 48-percentage-point gap between Zuckerberg's 13.6% economic stake and 61% voting power is the single most important governance fact at Meta. It is why the board can reject a one-share-one-vote proposal that drew 84% support from Class A shareholders in 2024 — and why the 2025 proposal (Proposal Six) was again recommended against. It is why separate Class-A vote disclosure (Proposal Seven) is resisted despite multiple peer companies already providing it. The CEO cannot be fired. The dual-class structure has no sunset, and a voluntary sunset would require Zuckerberg's own vote.

Insider buying vs. selling — the signal is unambiguous

Across Q1 2026 Form 4 filings, every reported insider transaction by a Meta executive or director has been a sale. Susan Li alone disposed of roughly 131,000 shares worth $84M between Feb 20 and Feb 27, 2026. Olivan sells on a metronomic weekly schedule under a 10b5-1 plan (~1,555 shares per week at ~$630). Bosworth sold $5.1M in a single day. None of this is illegal and most is pre-programmed, but the unidirectional flow is a fact the market should see.

Zuckerberg himself has not sold shares recently, but has pledged 12 million Class B shares — about 3.5% of his holdings, 2.1% of total company voting power — as loan collateral. The compensation and audit committees have blessed this arrangement. The collateral level is modest; the precedent is not.

Dilution and equity-plan cost

Meta is proposing a new 2025 Equity Incentive Plan that replenishes its share pool by up to 2.5% of Class A outstanding each January 1 for ten years — a 25% cumulative potential share pool expansion (before buybacks). As of Dec 31, 2024, 122.6M RSU shares were outstanding across the company with 482.6M remaining available for grant. In 2024, the company granted RSUs that consumed roughly 2.1% of shares based on 64.4M shares acquired on vesting; Meta's buyback program has historically more than offset this, but the headline dilution absent repurchase is meaningful.

No Results

The $987M paid to Broadcom in 2024, while its CEO Hock Tan sat on Meta's board, is the largest related-party transaction in recent Meta history. It was signed off under the policy, but it is a conflict that should never have been allowed to reach that size. Meta disclosed on April 14, 2026 that Tan will not stand for re-election, coincident with a multi-gigawatt chip deployment agreement extending the commercial relationship through 2029. The conflict is being resolved by removing the director, not by unwinding the contract. The $23.7M paid to the CFO's spouse as Chief Revenue Officer is legal and disclosed but is exactly the kind of arrangement outside CFOs would be expected to recuse from discussing internally.

Skin-in-the-game score

Skin-in-the-Game Score (out of 10)

A 6. Zuckerberg has the largest possible personal stake — his entire net worth is Meta stock — which pulls the score up. Li, Cox, Olivan and Bosworth each hold shares well above the $4M executive minimum. But the scheduled monthly sales across the entire operating team, the fact that non-CEO officers now receive ~$22M a year in RSUs regardless of Meta's stock performance, and the absence of a single open-market insider purchase cap the score firmly in the middle.

Board Quality

Composition

Meta's board is 15 seats, 14 independent + Zuckerberg. Five of those independents (Collison, Elkann, Powell McCormick, Songhurst, White) joined since early 2024, so one-third of the board has less than two years' tenure. The remaining third (Andreessen since 2008, Alford since 2019, Kimmitt/Killefer/Houston/Travis since 2020) provides institutional memory.

Expertise & independence scorecard

No Results

**Three directors are formally independent but materially conflicted or ideologically aligned:**

Hock Tan — his company received $987M from Meta in 2024 while he sat on the board. Exiting April 2026, which is the right outcome three years late.

Peggy Alford — chairs the compensation committee; served as CFO of the Chan Zuckerberg Initiative 2017-2019, a personal philanthropic vehicle of the CEO and his spouse. Technically independent under Nasdaq rules, but not independent in the informal sense that matters for compensation challenge.

Dana White + Dina Powell McCormick — both publicly aligned with President Trump; appointed in the context of Meta's January 2025 policy pivot away from third-party fact-checking. Whether that is governance strength (political skill) or weakness (ideological capture) depends on the reader's priors — but it is not a coincidence.

Committees

The Audit & Risk Oversight Committee (Travis chair, + Alford, Arnold, Killefer, Tan) is strong on finance expertise. The Compensation, Nominating & Governance Committee (Alford chair, + Andreessen, Houston, Xu) is staffed by peer tech CEOs who understand the equity-heavy playbook — that is both a strength (peer literacy) and a weakness (reciprocal generosity norms). The Privacy & Product Compliance Committee (Killefer chair, + Alford, Kimmitt) oversees the most existential risks Meta faces — child safety litigation, FTC antitrust, European regulation — and is arguably the most important committee on any tech board today.

Governance lapses and pressure points

No Results

The Verdict

Skin-in-Game (of 10)

Independent Directors / 15

CEO Voting Control

Broadcom RPT 2024 ($M)

$987

Governance Grade: C+

The letter grade is C+. Meta deserves credit for a deep, operationally coherent executive bench with long tenure; for a compensation structure that is genuinely equity-heavy and therefore aligned with the share price; for a board that has been actively refreshed; for a committee structure that matches the company's actual risks; and for a controlling founder whose personal net worth is overwhelmingly in Meta stock.

The strongest positives are (1) Zuckerberg's unrivaled long-term commitment — he has built this company for 22 years and will not leave, (2) the internal succession pipeline behind the CEO, and (3) a compensation committee advised by an independent consultant with robust anti-hedging, anti-pledging (with the Zuckerberg carve-out), and clawback policies.

The real concerns are (1) a dual-class capital structure that the controlling shareholder refuses to sunset against the stated wishes of 84% of the outside equity, (2) a $987M related-party contract with a director's company that took three years to unwind, (3) a board refresh pattern that pulls in politically-connected appointees at the same moment the company is retreating from content moderation — raising the question of whether independence means independent of management, or merely independent of the other directors, (4) uninterrupted insider selling with no open-market buying, and (5) an executive pay increase timed days after a layoff announcement.

What would move the grade up: a credible sunset on Class B super-voting rights; a permanent ban on director-company commercial relationships exceeding $50M; and a single open-market insider purchase by any operating executive.

What would move the grade down: an adverse FTC antitrust ruling that forces Instagram/WhatsApp divestiture without meaningful management accountability; a Zuckerberg health or capacity event in the absence of a disclosed succession plan; a second major related-party transaction pattern; or Powell McCormick's scope as President expanding to unseat the operating COO without an independent search.

The Full Story

Between the 2022 crash and today, Meta has run three different narratives — metaverse-first, efficiency-first, and now superintelligence-first — in three consecutive 10-Ks. The efficiency pivot delivered: margins doubled, earnings beat consensus every quarter for 12 straight periods, and the stock recovered to new highs. The metaverse promise was quietly re-scoped twice, and a new, larger AI capex cycle has replaced it as the load-bearing bet. Management credibility on ads is at a multi-year high; credibility on long-horizon hardware/AI bets is unresolved — the market gave back $307B in four April-2026 trading days questioning exactly that.

1. The Narrative Arc

Three mission statements in three years. Each rewrite absorbed the prior era's unfinished promise and moved the goalposts.

No Results

The three lines tell the arc. Revenue and operating income recovered aggressively from the 2022 trough as advertising normalized after ATT and the "Year of Efficiency" flushed costs. Then — just as the old narrative was vindicated — CapEx lifted off again, this time for AI infrastructure. The question ahead is whether that red line pays back the same way the 2017–2019 data-center buildout did, or the way the Reality Labs buildout has not.

**The anchor pivot:** In February 2023, with the stock still 70% below its 2021 peak, Zuckerberg declared the "Year of Efficiency" — a phrase that did not appear in any prior filing. The 22% headcount reduction, first-ever dividend (Feb 2024), and two consecutive years of 40%+ operating margins all flow from that single rhetorical reset.

2. What Management Emphasized — and Then Stopped Emphasizing

Counting how often each theme shows up across the 10-K business + MD&A + risk sections captures the narrative drift that prose descriptions obscure.

The heatmap is the full story in one visual. Two arrows matter:

Metaverse → AI → Superintelligence. "Metaverse" peaked in the 2023 10-K (47 mentions), was halved in 2024, and is nearly gone in 2025 (7). In its place, "AI" more than quadrupled and a brand-new vocabulary — superintelligence — appeared from zero with 11 mentions. This is the same playbook Meta used when it pivoted from Facebook to "Meta" in 2021: swap the noun, keep the ambition.
Efficiency is no longer the story. "Efficiency" language was load-bearing in the 2023 filing (14 mentions) and is essentially absent in 2025 (3). The spending discipline frame was retired the moment revenue re-accelerated. Capex tells the same story — $28B → $39B → $72B.

**Quietly dropped:** "Building community" was the founding mission ("give people the power to build community and bring the world closer together"). It was removed from the mission in 2024 and does not appear in the 2025 vision at all. The new mission — "build the future of human connection and the technology that makes it possible" — is compatible with an AI assistant instead of a social network.

3. Risk Evolution

Meta's risk language is normally sticky — the same 40-page risk factors section gets lightly remixed each year. Three shifts since 2023 are significant enough to read as real strategy changes, not just legal hygiene.

Three observations from the heatmap:

AI risk is now a first-class section. The 2023 10-K had no dedicated AI risk disclosure. The 2025 version contains multi-paragraph treatment of "generative AI and superintelligence" risks including deepfakes, elections, chatbot harm, and the FTC/Congressional/state-AG inquiries into Meta AI chatbots — explicit new language in 2025. Frontier-AI talent competition is also newly called out ("limited and competitive talent market"), consistent with Meta's 2025 hiring of the Superintelligence Lab.
Regulatory scope is widening, not narrowing. The 2025 bullet listing named regulations now includes the UK Online Safety Act, EU AI Act, and UK DMCC — all new in 2025. "Youth" was added as a regulatory category and the risk section newly flags "several bellwether trials in our youth-related litigation matters are scheduled for 2026 and beyond."
Ad-targeting risk language softened. ATT/Apple and GDPR still appear, but with less prominence — evidence that management believes privacy-enhancing ad tech and AI-driven targeting have reduced exposure. The 10% ad-price growth in 2024 and 9% in 2025 corroborate this.

**New in 2025:** "We have been the subject of media and government scrutiny relating to AI and AI chatbots, including inquiries and investigations by the FTC, members of Congress and state attorneys general." Reuters reporting on scam-ad revenues and Senate-led calls for FTC/SEC investigation post-date this filing — the next 10-K is likely to expand this section further.

4. How They Handled Bad News

Meta's pattern on bad news is consistent across cycles: absorb the loss, re-baseline guidance downward, then over-deliver on the new baseline. This shows up most clearly in how the Reality Labs losses have been framed.

No Results

Note the walk-back in 2025: "we expect our 2026 RL operating losses to remain similar to 2025." After three straight years of guiding "losses will increase," management is finally signaling a cap — but only after RL losses compounded from $13.7B to $19.2B, a ~$5.5B increase, which is itself larger than the total P&L of many mid-cap companies. The bet has been re-scoped from "metaverse as next computing platform" to "AI glasses as hardware wedge into an AI-assistant future." The dollar cost of the prior framing is fully booked; the benefit is not yet visible.

The market is already skeptical. Per news coverage, in April 2026 Meta shares fell ~17% in four trading days — the worst run since November 2022 — wiping out $307B in market value on exactly the same concern that triggered the 2022 crash: investors questioning whether the current capex cycle pays back. The 2025 10-K acknowledges the parallel only obliquely: *"costs related to acquisitions... charges associated with impairment or abandonment of any assets."*

5. Guidance Track Record

Looking at major promises that were measurable and mattered to valuation, the record is strongly positive on near-term financial commitments, mixed on segment/cost guidance, and unresolved on long-horizon product claims.

No Results

Credibility Score (1–10)

7.5

Why 7.5 / 10. Meta's near-term credibility is unusually strong for a mega-cap: 12 consecutive EPS beats, delivered on the Year of Efficiency commitment (margins up ~17 points), initiated a dividend, and funded $55B+ of buybacks across 2024–2025. The advertising franchise — the thing that actually prints cash — has been rebuilt more effectively than bears thought possible in late 2022. Offsetting that: the three-year drift from "metaverse" to "AI" to "superintelligence" involves retiring specific prior promises without ever calling them misses, and the new $72B AI capex bet has exactly the shape of the 2022-era metaverse spend that the market previously punished. One strike against honest bad-news handling: RL losses were described as "increasing meaningfully" in 2024 but the 10% actual increase stretches the word "meaningfully." A score above 8 would require that management name the metaverse program as something other than "evolving" and acknowledge re-scoping in plain language.

6. What the Story Is Now

The 2026 Meta is three companies under one P&L: (1) a cash-gushing ad machine generating ~$200B in revenue and 50%+ FoA operating margins, (2) a Reality Labs segment that has burned $66B over 2022–2025 with no visible revenue inflection, and (3) a newly-funded frontier-AI lab competing with OpenAI and Google for both talent and compute. Management's story is that all three are the same company — that ads-AI feedback loops are already paying off in ad performance, that AI glasses are the monetization wedge for the Reality Labs spend, and that superintelligence is the destination that makes the whole thing coherent.

No Results

What has been de-risked: the core ad business is a more durable cash engine than it looked in 2022. ATT is absorbed. Reels is still lower-monetizing but growing impressions. ARPP is up 64% since 2023 ($34.72 → $57.03). Buybacks plus dividend returned ~$31B to shareholders in 2025 alone.

What still looks stretched: the $72B CapEx run-rate now consumes almost all of operating cash flow — and the 2025 net income actually declined year-on-year ($62.4B → $60.5B) because of the One Big Beautiful Bill Act tax charge (17 points of effective rate, mostly one-time). Once ratio analysts look past the tax noise, 2025 underlying earnings were fine; the question is 2026–2028. If AI glasses are a real product category, Meta is positioned. If they are the new Quest Pro, RL losses have room to grow further, and the rhetorical softening from "next computing platform" to "complex, long-term initiative" will need to continue.

What the reader should believe vs. discount:

Believe	Discount
Ad revenue growth, margin expansion, EPS beats	Superintelligence timing claims
Server-life extension as real accounting benefit	"RL losses to remain similar" (same guidance, repriced)
Dividend and buyback cadence as durable	DEI/content-policy pivots as apolitical "legal landscape" responses
Infrastructure capex is real and being deployed	2026 payback from AI infra spend
Zuckerberg controls the company (dual-class, ~55% voting) — that is both a strength and an accountability gap	That the narrative will not shift again if this bet doesn't land

**The question for the next 12 months:** Does the superintelligence pivot produce a monetizable product by the 2026 10-K? If not, expect another narrative reset. Meta has earned the right to try new expensive things because the ad business actually works — but the pattern of renaming the long-term bet every 18 months will eventually reach a limit with investors who have been patient through two cycles already.

What's Next

The next nine days are the whole story. Q1 2026 earnings land on April 29, 2026 — the first print since the April tape wiped $307B off the market cap on capex fears, and the first real-time look at whether the FY26 $115–135B capex guide holds, whether price-per-ad is still accelerating, and whether the Reality Labs loss trajectory finally flattens. Every other catalyst on the calendar is a follow-on to what this single print says.

No Results

The single line the tape will trade on April 29 is price-per-ad YoY. At +9% it is the proof point for the AI-ad-flywheel thesis; below +6% it is the first visible crack. The capex reaffirmation is almost as important but has a range — a guide of "~$120B at the midpoint, trending lower by year-end" is different from "raising to $125–140B on demand". Expect the stock to move on the qualitative framing of capex more than the headline EPS print.

Not catalysts worth waiting on: headline AI model releases (Llama cadence is now decoupled from revenue), Threads MAU updates (zero ad revenue attached), and generic macro prints. This name trades on two numbers — ad pricing and capex — and one binary (FTC remedy).

For / Against / My View

For

**Lowest forward P/E in Big Tech, highest growth in the cohort.** Meta trades at 19.3x forward earnings against 22% YoY revenue growth and 41% operating margin — the combination does not exist anywhere else in the mega-cap tech set. Alphabet is 25.8x at 18%; Microsoft is 19.7x at 17%; Amazon is 25.8x at 14%. On EV/EBITDA, Meta is at 13.8x versus its own 5-year average near 16x.

**AI capex is already monetizing inside the ad auction today.** Price-per-ad is +9% YoY and ARPP has risen 64% since 2023 ($34.72 → $57.03). Advantage+/Lattice/conversions API absorbed Apple's ATT hit and are driving auction clearing prices higher while impressions still grow ~12%. Q4 2025 was the largest revenue quarter in company history at $59.9B, +24% YoY — acceleration, not maturity.

**Fortress cash generation absorbs the buildout with room to spare.** FY2025 operating cash flow hit $115.8B — 1.9x reported net income. Even after the $70B capex year, Meta generated $46B of free cash flow and returned $31.6B to shareholders. Net debt is effectively zero ($83.9B debt vs $81.6B cash/investments), Altman Z 7.4, interest coverage 90x. There is no solvency path for this balance sheet even if 2026 capex lands at $135B.

Bull price target: $820 in 12–18 months — $32 FY26E EPS × 25.5x, anchored to the 10-year trailing median of ~24x. Disconfirming signal: price-per-ad turns negative for two consecutive quarters.

Against

**The capex cliff reprices the equity.** Capex leapt from $37B to $70B in FY2025 and FY26 guidance is $115–135B — a $45–65B YoY incremental spend. FCF already fell from $54B → $46B despite 22% revenue growth; FCF/net-income conversion halved from 75% to 52%. At $120B capex against ~$130B of operating cash flow, 2026 FCF collapses toward $10B, with $10–15B of incremental depreciation landing on the P&L in 2026–27.

**The narrative walk has reached a limit.** Three missions in three 10-Ks (metaverse → AI → superintelligence). Reality Labs has burned $66B cumulatively since 2022, losses keep compounding ($13.7B → $16.1B → $17.7B → $19.2B), and the Llama open-source pledge was softened in 2025. The "superintelligence" vocabulary appeared from zero with 11 mentions in the 2025 10-K, has zero revenue attached, and exists to absorb a capex bet with exactly the shape of the 2022 metaverse spend that crashed the stock to $88.

**Governance is decaying, not compounding.** Zuckerberg controls 61% of votes against 13.6% economics; the board rejected one-share-one-vote despite 84% Class A support. Two Trump-aligned directors elevated (Dana White, Dina Powell McCormick — director to President and Vice Chairman in twelve months). Meta paid $987M to Broadcom in 2024 while Hock Tan sat on the board. Every reported insider transaction in Q1 2026 was a sale — Susan Li disposed of $84M in seven days. An active FTC trial targeting Instagram/WhatsApp divestiture is running.

Bear downside target: $450 in 12 months (26% below current $612) — $28 FY26 EPS × 16x. Primary trigger: FY26 capex reaffirmed ≥$115B at Q2 2026 without price-per-ad ≥+10% YoY.

The Tensions

1. The $70B FY25 capex — productive buildout or the start of a permanent margin hit?

Bull reads the $70B capex as already monetizing — price-per-ad is +9% YoY and FCF stayed positive at $46B, so the flywheel is working. Bear reads the same $70B (and the $115–135B FY26 guide) as the trajectory where FCF conversion already halved from 75% to 52% before the 2026 depreciation even hits the P&L. Both cite the same numbers: the $37B → $70B step-up and the $115–135B FY26 range. Resolves on: the April 29 Q1 2026 print — specifically whether price-per-ad prints ≥+7% and the FY26 capex band is reaffirmed without drifting toward $135B. Either condition missed flips the read.

2. The 19.3x forward multiple — cheapest in Big Tech or the market correctly pricing binary outcomes?

Bull reads 19.3x forward against 22% growth and 41% margins as the asymmetric discount in the mega-cap cohort — peers are 20–28x and Meta's own 10-year median is 24x. Bear reads the same 19.3x as the market doing its job, pricing capex ramp plus the FTC trial plus EU DMA plus state-AG youth-safety exposure; the April $307B market-cap wipe in four trading days is the signal that the multiple is thin for a reason. Both cite the same 19.3x forward P/E. Resolves on: the FTC antitrust remedy ruling (expected mid-to-late 2026) — a no-divestiture outcome plus a stable capex guide closes the discount; a structural remedy re-bases the thesis.

3. Management credibility — earned the right to spend, or the same drift pattern as 2022?

Bull reads the 12-quarter EPS beat streak, the 17-point operating-margin expansion from the 2022 trough, and the clean Year of Efficiency delivery as proof this team has earned the right to run a second long-horizon capex cycle. Bear reads the same management team writing three different missions in three consecutive 10-Ks, softening the Llama open-source pledge, guiding Reality Labs losses upward through four successive revisions, and inventing "superintelligence" from zero — the pattern that preceded the 2022 metaverse reset. Both cite the same team and the same communication history. Resolves on: the Q2 2026 print (late July) — whether FY26 capex holds at ≤$115B or creeps toward the top of the range. A firm midpoint is a credibility vote; a creep is the drift pattern repeating.

My View

Close call, slight edge to waiting. The Against side's capex concern is real but the Bull's fortress balance sheet makes it survivable rather than existential — this is not a solvency question, it is a multiple question. What tips me toward patience is Tension #1: the biggest single catalyst, the April 29 print, is nine days away and directly resolves the capex-versus-ad-pricing question that both sides are fighting over. Buying ahead of it is paying for optionality on a print that will either validate the +9% price-per-ad thesis or crack it open, and the tape has already shown (via the April $307B wipe) that the asymmetry is not obviously favorable. I'd wait. The one condition that would flip me to lean constructive: Q1 price-per-ad prints ≥+10% YoY and the FY26 capex band is reaffirmed without lifting the midpoint — at that point Tension #1 resolves decisively for the bull, and the 19.3x discount becomes genuinely attractive rather than deserved.

The Bottom Line from the Web

Between the filings and today, three things moved that change the thesis. First, Meta won the FTC antitrust trial on November 18, 2025 — the single largest structural governance risk (forced Instagram/WhatsApp divestiture) is effectively off the board. Second, management guided 2026 capex to $115–$135B, nearly doubling the $72.2B spent in 2025, triggering a multi-week sell-off that erased over $300B in market cap between March and April 2026. Third, the $987M related-party risk around director Hock Tan (CEO of Broadcom) was resolved not by recusal or competitive bidding but by Tan stepping off the Meta board as Meta expanded the Broadcom chip contract to 1 GW through 2029 — a governance outcome that is cleaner than it looked in the proxy but cements Broadcom as a strategic dependency.

What Matters Most

The ten findings below are ranked by how much each one would move an investor's view of META. Filing-based work by the specialists already covers the in-10-K reality; this section is strictly what the web adds on top.

**1. FTC antitrust case — Meta wins outright (Nov 18, 2025).** U.S. District Judge James Boasberg ruled after a six-week bench trial that the FTC failed to prove Meta holds a monopoly in personal social networking, rejecting the attempt to unwind the 2012 Instagram and 2014 WhatsApp acquisitions. Meta will **not** be forced to divest either asset. Sources: [Reuters](https://www.reuters.com/sustainability/boards-policy-regulation/meta-defeats-us-antitrust-case-over-instagram-whatsapp-2025-11-18/), [Sullivan & Cromwell memo](https://www.sullcrom.com/insights/memo/2025/December/Meta-Prevails-FTC-Monopolization-Case), [NPR](https://www.npr.org/2025/11/18/nx-s1-5495626/meta-ftc-instagram-whatsapp-antitrust-ruling). **Why it matters:** the ruling was described by Reuters as the first decisive court win for big tech in the current antitrust wave — removing the one scenario that could have forced a breakup of ~40% of the ad-revenue base.

**2. 2026 capex guide of $115–135B — nearly 2× 2025's $72.2B.** Guidance was set on the Q4 2025 call (January 28, 2026). Management said "the majority of expense growth will be driven by infrastructure costs… higher depreciation, and higher infrastructure operating expenses." Midpoint $125B exceeds the GDP of over 100 countries. Sources: [Data Center Dynamics](https://www.datacenterdynamics.com/en/news/meta-estimates-2026-capex-to-be-between-115-135bn/), [Global Data Center Hub](https://www.globaldatacenterhub.com/p/metas-135b-capex-signal-ads-are-becoming), [Futurum](https://futurumgroup.com/insights/ai-capex-2026-the-690b-infrastructure-sprint/). **Why it matters:** a step-change in depreciation is about to run through the P&L, and every analyst model now turns on whether AI-driven ad pricing offsets it. The March-April 2026 drawdown of ~13% YTD is the market's first vote on that question.

**3. Hock Tan / Broadcom related-party transaction resolved by board exit (April 14, 2026).** Meta committed to deploying **1 gigawatt of custom MTIA chips** co-designed with Broadcom through 2029. In conjunction, Broadcom CEO Hock Tan will **not stand for reelection** to Meta's board (he joined in 2024). Sources: [CNBC](https://www.cnbc.com/2026/04/14/meta-commits-to-one-gigawatt-of-custom-chips-with-broadcom-as-hock-tan-agrees-to-leave-board.html), [Techbuzz.AI](https://www.techbuzz.ai/articles/meta-locks-in-1gw-chip-deal-with-broadcom-as-ceo-exits-board). **Why it matters:** the largest RPT in company history (the $987M starting figure is now dwarfed by a 1 GW multi-year commitment) is being de-risked the correct way — removing the director — but Broadcom is now a structural strategic dependency, not just a vendor.

**4. Scam-ad allegations — Senate calls for FTC/SEC probe (Nov 2025).** A [Reuters investigation](https://www.reuters.com/investigations/meta-is-earning-fortune-deluge-fraudulent-ads-documents-show-2025-11-06/) reported internal documents showing Meta earned a substantial share of revenue from fraudulent/banned ads, with a 2025 document worrying "abrupt reductions of scam advertising revenue could affect its business projections." U.S. senators subsequently demanded FTC and SEC action. [CNET](https://www.cnet.com/tech/services-and-software/meta-allegedly-profited-by-16b-from-scam-ads-us-senators-demand-ftc-sec-probe/) reports alleged scam-ad revenue around **$16B**. **Why it matters:** this is a new regulatory vector and a potential disclosure expansion in the next 10-K. Unlike the FTC monopoly case it is early-stage, but the financial impact is concretely quantified.

**5. Dina Powell McCormick hired as President and Vice Chairman (Jan 12, 2026).** Former Trump administration adviser and Goldman Sachs senior executive; reports directly to Zuckerberg. Meta stated she will serve on the management team and help "guide strategy and execution," with an explicit lobbying/external-relations tilt. Sources: [Reuters](https://www.reuters.com/business/dina-powell-mccormick-joins-meta-president-vice-chairman-2026-01-12/), [Meta press release](https://about.fb.com/news/2026/01/dina-powell-mccormick-joins-meta-as-president-and-vice-chairman/), [Axios](https://www.axios.com/2026/01/12/meta-dina-powell-mccormick-president-vice-chairman). **Why it matters:** this is **not** a COO-replacement — Javier Olivan remains in place — so it is best read as an external-affairs principal for Washington and international regulators, not an operational succession signal.

**6. Q3 2025 GAAP profit down 83% — a $15.93B one-time tax charge.** Meta disclosed a non-cash income tax charge of $15.93B in Q3 2025 triggered by a provision of the "One Big Beautiful Bill Act" (OBBBA) interacting with Corporate Alternative Minimum Tax. Sources: [Variety](https://variety.com/2025/digital/news/meta-q3-2025-earnings-tax-charge-trump-one-big-beautiful-bill-1236564932/), [Deep Quarry](https://deepquarry.substack.com/p/metas-159-billion-tax-shock-when). **Why it matters:** trailing P/E and GAAP EPS are distorted by this charge. Investors comparing headline multiples without adjusting will misread the 2025 earnings trend.

**7. Reality Labs — cumulative burn ~$73B; 2025 operating loss widened to $19.19B.** RL revenue was $2.21B in 2025 (+2.8% YoY) — essentially flat against losses that continue to widen. Ray-Ban Meta smart glasses did hit the 2M sales milestone (March 2026) with ~$1.1B of quarterly RL hardware revenue, but that is a rounding error against the loss trajectory. Sources: [Auganix](https://www.auganix.org/xr-news-meta-reality-labs-2025-financial-report/), [Yahoo Finance](https://finance.yahoo.com/news/meta-platforms-lost-73-billion-165823364.html), [Next Reality](https://virtual.reality.news/news/meta-ray-ban-smart-glasses-hit-2m-sales-milestone/). **Why it matters:** the 2021 metaverse bet has not paid back; the pivot story is now AI/superintelligence, but RL remains a cash hole the market has learned to tolerate.

**8. New executive pay plan — up to $2.7B potential payouts (March 2026).** Incentive stock options for CFO Susan Li, CTO Andrew Bosworth, CPO Chris Cox, and COO Javier Olivan could pay up to $2.7B apiece at stretch stock-price targets. Concurrently Meta is laying off ~700 staff. Cash costs of employee stock-based comp reached ~$42B in 2025 — about 96% of $43.6B FCF. Sources: [CNBC](https://www.cnbc.com/2026/03/24/meta-offers-stock-awards-options-for-executives-aggressive-timing.html), [Business Insider](https://www.businessinsider.com/meta-boost-executive-pay-compensation-targets-stock-price-market-cap-2026-3), [TNW](https://thenextweb.com/news/meta-9-trillion-valuation-executive-stock-options-layoffs). **Why it matters:** the SBC overhang is becoming visible at the buyback-effectiveness level. Targets are set to near-doubling of the stock, which resets management's bar and dilutes on a scale the ad-business cash machine has to absorb.

**9. EU DMA — Meta to offer less-personalized ads from January 2026 (agreed Dec 8, 2025).** Meta's revised "pay or consent" proposal won EU approval and avoids daily DMA fines. Users get a real choice between full personalization and a less-personalized experience. Sources: [European Commission DMA page](https://digital-markets-act.ec.europa.eu/meta-commits-give-eu-users-choice-personalised-ads-under-dma-2025-12-08_en), [Reuters](https://www.reuters.com/business/meta-offer-choices-personal-facebook-instagram-ads-eu-says-2025-12-08/). **Why it matters:** Europe is ~24% of revenue. The binding question — how many users opt into "less personalized" — is still unresolved, but the regulatory ceiling (daily fines) has been lifted.

**10. Superintelligence Labs — hiring pause and departures (Aug–Sept 2025).** A memo reported by The Verge said Meta is "temporarily pausing hiring across all MSL teams, with the exception of business critical roles." Observer and others chronicled a wave of high-profile departures after the initial bonus-driven recruiting campaign. Sources: [The Verge](https://www.theverge.com/ai-artificial-intelligence/767746/meta-ai-superintelligence-lab-departures-scale-zuckerberg-memo), [Observer](https://observer.com/2025/09/meta-superintelligence-team-shedding-staffers/), [Built In](https://builtin.com/artificial-intelligence/meta-superintelligence-labs). **Why it matters:** the "AI pivot" narrative that is supposed to justify $115–135B of 2026 capex is showing organizational strain six months after launch. Talent retention is the leading indicator; revenue is the lagging one.

Recent News Timeline

No Results

What the Specialists Asked

Insider Spotlight

Three insider themes emerged from the web research beyond what the filings show.

Zuckerberg's compensation structure is unique among large-cap CEOs. 2024 pay ratio of 65:1 with the median Meta employee. Salary of $1. Total 2024 "other compensation" disclosed at $24.4M+ in prior-year filings. Perquisite components: residential/travel security ~$10.43M, annual pre-tax security allowance $14M, personal use of private aircraft ~$2.5M. Voting-power-to-economic-ownership ratio (~61% vote vs ~13% equity) remains the single largest governance feature of the company.

New stretch-option plan for the operating leadership (March 2026). CFO Susan Li, CTO Andrew Bosworth, CPO Chris Cox, and COO Javier Olivan are now on incentive stock options that could pay up to $2.7B per executive at the highest stock-price hurdles. This was announced concurrently with ~700 layoffs — an optics point that drew pushback. Source: Business Insider.

All 2025–2026 insider sales are under 10b5-1 plans. COO Olivan's plan was adopted November 17, 2025; Director Kimmitt sold 580 shares at $632.02 in March 2026 under the same pre-planned framework. No discretionary insider buying was found in the web research covering the last 12 months.

No Results

Industry Context

Meta's $115–135B 2026 capex is part of a sector-wide spending cycle. Per Futurum and Tech Insider, the big four hyperscalers are projected to spend roughly $690B combined in 2026. Meta at the $125B midpoint sits below Amazon (~$200B) and Alphabet ($175–185B) but above Microsoft (~$120B tracking). That positions Meta as the fastest-growing capex line relative to its 2025 base (1.73× vs Alphabet ~1.4× and Microsoft ~1.4×).

Sources: Futurum, Tech Insider, Yahoo Finance. Figures are reported/guided ranges rounded to reader-friendly integers. Interpretation: Meta is the only one of the four whose 2026 capex is projected to exceed its own 2025 revenue growth — that is why the March–April sell-off landed here first and hardest.