Capex Bill Comes Due: Top Line Beats and FoA Rips +33%, but FY26 Capex Raised $10B, Buybacks Zeroed, and DAP Misses — Stock -7% AH
Initial Read: Operationally clean — FoA ad revenue +33% with impressions +19% and price/ad +12% is the strongest single-quarter ad print of the cycle. But the spending bill the Q4 print teed up just got bigger: FY26 capex raised $10B at the midpoint to $125-145B, Q1 buybacks were zero (vs. $12.75B in Q1 2025), and DAP missed at 3.56B vs. ~3.62B Street. Maintain Hold; the bear-side capex pillar from our Q3 2025 downgrade has been reconfirmed.
Key Takeaways
- Revenue $56.31B (+33% YoY, +29% constant currency) beat consensus ~$55.45B by ~1.5%; the FX tailwind contributed ~4 points of growth, but the underlying ex-FX +29% is still the fastest organic growth print since Q2 2025 and the second-fastest of the cycle. Ad impressions +19% YoY with average price per ad +12% YoY — the volume-and-price combination is the cleanest read on AI-driven ad ROAS we've had.
- Reality Labs operating loss narrowed to $(4.03)B from $(4.21)B in Q1 2025 — small magnitude but the directional signal matters because the Q4 2025 framing was that 2025 marked "peak" RL losses. Q1 2026 is the first quarter testing that pledge and it printed inside it. RL revenue $402M (-2% YoY) is on the long pre-Glasses-V2 air pocket we expected.
- FY26 capex raised to $125–145B from $115–135B prior — a $10B step-up at the midpoint. Management attributes to "higher component pricing this year and, to a lesser extent, additional data center costs to support future year capacity." The first half of that explanation is exogenous and broadly believable (the AI-server cost curve is harder than expected); the second is the same framing they used three quarters ago and has now compounded into the largest absolute capex year of any company in our coverage. FY26 opex is unchanged at $162–169B.
- Q2 2026 revenue guide $58–61B (mid $59.5B) — roughly in line with Street ~$59.5B. With a ~2% FX tailwind embedded, the implied ex-FX growth rate decelerates from Q1's +29% to ~+22% at the midpoint. After three straight quarters of accelerating ad revenue, this guide implies the deceleration the bears have been waiting for begins in Q2.
- Q1 buybacks were zero for the second straight quarter (vs. $12.75B in Q1 2025); only the $1.35B dividend was returned. Cash plus marketable securities of $81.18B is roughly flat sequentially because operating cash flow ($32.23B) was nearly entirely consumed by capex ($19.84B) and tax/SBC payments. The implicit message is the same one Q4 sent: capital allocation is now AI-build first, shareholder return second.
- Preliminary Rating: Hold (Maintained). Q3 2025 downgrade thesis is reconfirmed: spending magnitude has decoupled further from ROI evidence on a Q/Q basis even though the ad-side ROI itself remains strong. Q1 OI of $22.87B (+30% YoY) keeps the absolute-dollar OI-growth pledge intact, which is the soft floor that prevents an Underperform call here.
Results vs. Consensus
| Metric | Actual (Q1 2026) | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $56.31B | $55.45B | Beat | +1.6% |
| Diluted EPS (GAAP, reported) | $10.44 | ~$6.79 (adj.) | Beat (mechanical) | $8.03B tax benefit inflates |
| Diluted EPS (ex-tax benefit) | ~$7.31 | ~$6.79 | Beat | +7.7% |
| Operating Income | $22.87B | ~$21.5B (implied) | Beat | +~6% |
| Operating Margin | 41% | ~39% (implied) | Beat | +~200 bps |
| Family DAP | 3.56B | ~3.62B | Miss | -1.7% |
| Free Cash Flow | $12.39B | n/a | In line | +20% YoY |
Quality of Beat
- Revenue: The headline +33% includes a ~4-point FX tailwind. Ex-FX +29% is still strong but is roughly on top of the +29% reported in Q4 2025 — call it a stable acceleration plateau, not an extension. The beat vs. consensus is a thin ~$860M, well within FX noise. Importantly, ad impressions +19% is the highest impression-growth print of the cycle (vs. +14% in Q1 2025, +18% in Q4 2025), and the +12% price/ad pairs with +19% volume — both volume and price contributing roughly equally is the textbook signature of a healthy ad market with AI-driven ROAS gains. This is fundamentally a strong number.
- Operating margin: 41% is flat with Q1 2025's 41%. After the Q3 2025 / Q4 2025 margin-compression episode (Q3 op margin compressed 290 bps YoY), holding flat at 41% on +33% revenue with 35% expense growth is a creditable showing. R&D growth is the cost driver (+46% YoY to $17.7B) — SBC and infrastructure depreciation tied to the AI build. Marketing and G&A are well-controlled (+5% and +15% respectively). Margin is stable, not expanding.
- EPS: The $10.44 GAAP print is mechanically inflated by the $8.03B tax benefit from Treasury Notice 2026-7 (which addresses CAMT treatment of capitalized R&D and partially reverses the $15.93B Q3 2025 OBBBA charge). The "real" EPS — operating result through a normalized tax rate — is ~$7.31, which still beats Street ~$6.79 by ~8%. So the operating EPS beat is real, but a PM looking at headline GAAP and pricing in ~$10 EPS run-rate would be wildly wrong.
- DAP miss: 3.56B vs. ~3.62B Street is a real soft spot, partially explained by Iran internet disruptions and Russia restricting WhatsApp access — both exogenous to the business. Q/Q decline of ~5% on those geographies is plausible but the Street will demand quantification on the call. Engagement (impressions per user) is presumably up given +19% impression growth on +4% DAP growth, which is the more important durable trend.
- FCF: $12.39B FCF on $32.23B operating cash flow with $19.84B of capex. FCF up +20% YoY — solid, but the gap between OCF and FCF widened materially as capex has scaled. FCF conversion has gone from ~43% of OCF in Q1 2025 to ~38% in Q1 2026.
Family of Apps — The Engine
| FoA Metric | Q1 2026 | Q1 2025 | YoY |
|---|---|---|---|
| Advertising Revenue | $55.02B | $41.39B | +33% |
| Other Revenue (incl. WhatsApp Business) | $885M | $510M | +74% |
| Total FoA Revenue | $55.91B | $41.90B | +33% |
| FoA Operating Income | $26.90B | $21.77B | +24% |
| FoA Operating Margin (implied) | ~48% | ~52% | -~400 bps |
| Ad Impressions YoY | +19% | +5% | +14 pp acceleration |
| Average Price per Ad YoY | +12% | +10% | +2 pp acceleration |
| Family DAP (March) | 3.56B | 3.43B | +4% |
Ad Revenue: Volume and Price Both Accelerating
FoA advertising of $55.0B (+33% YoY) is the strongest single-quarter ad-revenue print in dollar terms in Meta's history and the second-fastest growth rate of the cycle behind Q4 2025's +33% (note: rounding hides that Q1 is fractionally stronger ex-FX). The composition is what matters most. Impression growth of +19% is the highest of any quarter we have on file going back to the start of the AI investment cycle in 2024 — this is the GEM/Andromeda recommendation engine plus Threads/Reels inventory plus Meta AI feed integration generating real impression supply. Price-per-ad of +12% maintains the ROAS-driven pricing power that has been the singular bull-case data point since Q1 2025 initiation. The volume-plus-price combination is what a PM wants to see — this is not impression-growth-with-price-collapse (which would be inventory expansion at the cost of unit economics) and it is not price-growth-with-volume-collapse (which would be macro-narrowed scarcity). Both are working.
FoA Margin Compression: Engineering Spend Catches Up
FoA operating margin of ~48% (calculated as $26.9B OI / $55.9B revenue) is down ~400 bps from Q1 2025's ~52%. This is the more concerning side of the FoA print: even with revenue growing 33%, FoA-segment margins compressed because R&D — overwhelmingly allocated to FoA-supporting AI infrastructure depreciation, MTIA chips, and Meta Superintelligence Labs headcount — grew 46% YoY. The cash-flow lever in our thesis is FoA margin, and FoA margin is going the wrong way. The mitigating factor: 48% on $55.9B is still extraordinary in absolute dollars and the segment is producing $26.9B of OI per quarter ($107.6B annualized) — easily enough to fund the entire AI build and still produce growing absolute OI. But "compressing margin on accelerating revenue" is not the textbook bull setup we had through Q2 2025.
Other Revenue (WhatsApp Business / Threads): +74%
FoA Other revenue of $885M (+74% YoY) is the small line item that backs the long-term monetization-diversification thesis. WhatsApp Business paid messaging crossed a $2B run rate by Q4 2025 per the backfill summary; +74% growth here implies the run rate has stepped up materially again in Q1, putting FoA Other on track for ~$4B+ in 2026 from $1.4B in 2024. Threads monetization is mostly inside the ad-impressions line at this point but the surface area is real and growing. This is the "Meta AI / WhatsApp Business become real 2026+ monetization vectors" pillar of the bull case — confirmed and accelerating.
DAP: Geopolitical Soft Spot, Engagement Holds
DAP of 3.56B (+4% YoY) is below the ~3.62B Street estimate and declined Q/Q vs. Q4 2025's 3.43B+. The press release attributes the Q/Q softness specifically to internet disruptions in Iran and a restriction on access to WhatsApp in Russia. Both are plausible, exogenous, and likely transitory. The more important read is that ad impressions grew +19% on a DAP base of +4% — implying impressions-per-DAP up ~14% — which suggests engagement deepened materially even as headline user count was geopolitically pressured. The Street will probably look past the DAP miss given the impressions trajectory; we agree.
Reality Labs — Loss Narrows for the First Time in Years
| RL Metric | Q1 2026 | Q1 2025 | YoY |
|---|---|---|---|
| Revenue | $402M | $412M | -2% |
| Operating Loss | $(4.03)B | $(4.21)B | +$182M improvement |
| Implied OpEx (incl. COGS) | ~$4.43B | ~$4.62B | -4% |
Reality Labs is the segment that produced the best directional surprise in this print, even though the headline numbers look unremarkable. Operating loss narrowed by $182M YoY — the first YoY loss reduction in Reality Labs in many quarters and the first quarter testing Zuckerberg's Q4 2025 framing that 2025 marked "peak" RL operating losses. Q1 2026 is inside that envelope. At a $4.0B quarterly loss run rate, RL is on track for ~$16B of full-year operating loss, which would be marginally lower than 2025's run rate — exactly the trajectory the Q4 thesis update assumed.
RL revenue of $402M (-2% YoY) is the air pocket between Quest 3 saturation and the next-generation AR Glasses ramp (Ray-Ban Meta is presumably the dominant contributor here at this point but is too small to break out). The product cycle here is well-understood — RL hardware revenue is lumpy, gated on launch timing, and the loss trajectory is the more important driver of segment value than the revenue line.
Why it matters: The RL pillar of the bear case from Q1 initiation through Q3 2025 was "open-ended losses with no visible inflection." Q4 2025 introduced the language of a peak; Q1 2026 is the first piece of evidence that the peak language was real. If subsequent quarters confirm the loss trajectory bends down, this removes one of the three structural bear pillars from the thesis. We don't update the rating on one quarter, but the directional read is positive.
AI Capex — Raised Again, Now $125–145B for FY26
| Capex Framing | Q4 2025 Guide | Q1 2026 Guide | Change |
|---|---|---|---|
| FY26 Capex Range | $115–135B | $125–145B | +$10B at midpoint |
| FY26 Capex Midpoint | $125B | $135B | +8% |
| FY26 Opex Range | $162–169B | $162–169B | Unchanged |
| Q1 Capex Actual | n/a (forecast) | $19.84B | — |
| Implied 9M Remaining Capex (midpoint) | n/a | ~$115B | ~$38B/quarter run rate |
This is the headline negative of the print and the data point that defines the market reaction. Three months after Q4 2025 quantified FY26 capex at $115–135B (already above our Q3 model range and the data point that confirmed the bear-case capex math), management raised the range $10B at the midpoint to $125–145B. The stated drivers are (1) higher component pricing — broadly believable given the broader AI-infrastructure cost-curve commentary across hyperscalers this cycle, and (2) "additional data center costs to support future year capacity," which is the same forward-build language management has used for three consecutive quarters and which has consistently translated into raises rather than holds.
What concerns us: The capex raise is decoupled from any incremental forward-revenue commitment. Q2 revenue guidance is in line with Street rather than raised; FY26 opex is unchanged. The $10B incremental capex flows through D&A starting roughly 2027, which is the year where the OI-growth pledge becomes mechanically harder to honor. At a notional 5–7 year useful life on AI infrastructure, an incremental $10B of capex generates ~$1.4–2.0B of incremental annual D&A starting in 2027 — which is non-trivial against an OI base that needs to grow off 2025's $83.3B and 2026's implied $90B+.
What does not concern us (yet): Q1 capex of $19.84B is on a ~$80B annualized pace, well below the $125–145B FY26 envelope. This is consistent with management's standard pattern of back-half-loaded capex deployment as data-center capacity comes online. It also leaves room for the FY26 actual to land closer to the midpoint than the high end if component pricing normalizes — but that is a contingent improvement, not a base-case improvement.
Capital Returns — Buybacks Zeroed for Second Quarter
| Capital Return | Q1 2026 | Q1 2025 | YoY Change |
|---|---|---|---|
| Share Repurchases | $0 | $12.75B | -100% |
| Dividends + Equivalents | $1.35B | $1.33B | +1% |
| Total Return to Shareholders | $1.35B | $14.08B | -90% |
| Diluted Share Count (mm) | 2,564 | 2,590 | -1% |
Q1 2026 is the second consecutive quarter (after Q4 2025) with zero share repurchases. Q1 2025 had $12.75B of buybacks; the year-on-year delta is a $13B swing in capital return. The dividend was held flat. Cash-and-marketable-securities of $81.18B is roughly stable Q/Q because operating cash flow funded capex and tax/SBC withholding rather than buybacks.
This is the single cleanest signal in the print that capital allocation has fundamentally shifted. For most of the post-Q4 2022 era, Meta was a buyback-led capital-return story — the share count had been on a steady downward trajectory for eight quarters straight before Q4 2025. The pause is now a two-quarter pattern, which means it is policy, not noise. Diluted share count actually ticked down YoY (2,564M from 2,590M) reflecting net SBC vs. tax-withholding share retirements — but this is roughly de-minimis vs. the 2023–early 2025 pace of share-count reduction.
Implication: Multiple compression risk. A buyback-supported tape is a different valuation construct from a capex-funded tape. Removing the bid is not a fundamental issue — fundamentals are fine — but it removes a marginal-buyer floor that the stock has had through every prior earnings drawdown of the AI investment cycle.
Threads / Llama AI / Generative AI Products
The press release is light on specific Threads/Llama/Meta AI metrics — a notable omission given the Q4 2025 emphasis on the Meta Superintelligence Labs (MSL) launch. The CEO quote highlights "the release of our first model from Meta Superintelligence Labs" but provides no specifics: model name, parameter scale, benchmark results, or deployment plan. This is a question for the call.
What the release does tell us about AI commercialization is indirect but consistent with the Q4 thesis: ad impressions +19% (highest of cycle) and price/ad +12% are both materially driven by AI-recommendation gains (GEM/Andromeda extending into Threads, Reels, and creator content); FoA Other revenue +74% is materially driven by WhatsApp Business AI agent monetization. These are the ROI data points that justify the capex even if they are presented as headline ad metrics rather than as "AI product revenue" line items.
The absence of any quantification of (1) Threads MAU/DAU, (2) Meta AI assistant MAU, or (3) Llama download/deployment metrics in the press release is conspicuous. Each of these was disclosed at various points in 2025 prints. Their absence here may simply reflect Q1 conventions or it may suggest the metrics are not where the company wants them. Worth pressing on the call.
Notable Items in the Release
$8.03B Tax Benefit Distorts GAAP — Treasury Notice 2026-7
The press release foregrounds (in the same numbered footnote across multiple tables) that Q1 results include an $8.03B income tax benefit recognized as a result of U.S. Treasury Notice 2026-7, which addressed CAMT treatment of previously capitalized U.S. R&D costs. This benefit partially reverses the $15.93B non-cash CAMT charge taken in Q3 2025 upon enactment of the One Big Beautiful Bill Act. Excluding this benefit, the effective tax rate would have been ~37 percentage points higher (i.e., approximately +14% rather than the reported -23%) and diluted EPS would have been $3.13 lower at ~$7.31.
Assessment: This is bookkeeping unwind of a prior bookkeeping charge. The combined Q3 2025 + Q1 2026 net effect is a ~$7.9B residual CAMT charge versus the original $15.93B headline — meaningfully smaller than the original framing implied. PMs should treat both the Q3 2025 charge and the Q1 2026 benefit as non-operating noise. Operating EPS power for Q1 is ~$7.31, not $10.44.
Q2 Tax Rate Guide: 13–16% — Below 2026 YTD But Forward-Looking
Management guides "tax rate for the remaining quarters of 2026 to be between 13–16%" — meaningfully below the Q1 2025 reported rate of 9% but consistent with the post-OBBBA structural rate. This implies that for Q2–Q4, modeled EPS should reflect ~14–15% tax against ~9% in 2025, which is a ~5–6 pp tax-rate headwind to YoY EPS comparisons even on growing operating income.
Assessment: Modeling implication. Q2 2026 EPS comp will be tougher than the operating story suggests because of the tax-rate normalization. Underwrite OI growth, not GAAP EPS growth.
Material-Loss Language on Youth Trials Repeats
The CFO outlook block explicitly states: "we continue to see scrutiny on youth-related issues and have additional trials scheduled for this year in the U.S., which may ultimately result in a material loss." This is the same "material loss" language Susan Li introduced in Q3 2025 — which is when we downgraded — and which produced the Q4 2025 G&A spike of +386% YoY on legal accruals. Q1 2026 G&A was $2.61B (+15% YoY), considerably more measured than Q4's spike, which suggests the legal accrual flow is normalizing rather than accelerating in Q1.
Assessment: The legal/regulatory tail risk that emerged in Q3 remains live. Q1 G&A is not the spike Q4 was, but the language is identical and the trial calendar is now described as "this year." Watch the call for trial-date specificity.
Headcount: +1% YoY to 77,986
Headcount of 77,986 (+1% YoY) is essentially flat. Given R&D opex grew 46% YoY, the cost growth is overwhelmingly compensation per head (Meta Superintelligence Labs reportedly hired at premium packages) plus infrastructure depreciation flowing through R&D rather than headcount expansion.
Assessment: Headcount discipline holds; compensation inflation is the real R&D story. The "AI talent war is bidding up package economics" thesis is empirically true here.
Long-Term Debt Holds at $58.7B; Balance Sheet Capacity Still Wide
Long-term debt of $58.75B is essentially unchanged from the $58.74B at year-end 2025 — no incremental issuance in Q1 even though the Q4 2025 framing teed up the possibility of debt-funded capex. Cash-and-securities of $81.18B against $58.75B of debt leaves net cash of ~$22B. The balance sheet has more than enough capacity to support the raised capex envelope without needing to lean further on debt issuance in 2026.
Assessment: Balance sheet is fine. The buyback pause is a deliberate prioritization, not a liquidity-driven necessity.
Guidance & Outlook
| Metric | Prior Guide (Q4 2025) | New Guide (Q1 2026) | Change |
|---|---|---|---|
| Q2 2026 Revenue | n/a (one-quarter forward) | $58–61B (mid $59.5B) | Street ~$59.5B — in line |
| FY26 Total Expenses | $162–169B | $162–169B | Unchanged |
| FY26 Capex | $115–135B | $125–145B | +$10B at midpoint |
| Q2–Q4 Tax Rate | n/a | 13–16% | — |
| FY26 Operating Income | "Above 2025 OI" ($83.3B) | "Above 2025 OI" — reaffirmed | Unchanged pledge |
Implied Ramps
The Q2 guide of $58–61B (mid $59.5B) represents +6% Q/Q growth from Q1's $56.31B and ~+22% YoY ex-FX. This is a deceleration from Q1's +29% ex-FX, consistent with the law-of-large-numbers framing the bears have been pushing for the last three quarters. To hit FY26 revenue of ~$245–250B (the implied range required for the OI-growth pledge to work given the opex envelope), H2 needs to roughly average $65B/quarter — meaningful sequential acceleration from the Q2 guide. That is achievable with continued FX tailwind plus the typical Q4 seasonality, but it is not effortless.
The capex raise without an offsetting opex cut means the operating-income pledge is now slightly tighter math than it was at Q4. FY26 capex range $125–145B + opex range $162–169B = total spend $287–314B, against a revenue base that needs to be roughly $245–250B to produce OI > $83.3B with the indicated cost structure. The pledge survives, but the margin of safety has narrowed.
Our Interpretation
The guide is best characterized as "steady on the revenue line, harder on the spending line" — exactly the Q4 2025 setup repeating with a higher capex number. The Street had been positioned for a possible capex hold (the bull case) or a $5B raise (the consensus case); $10B at the midpoint is at the high end of plausible outcomes. Combined with the in-line Q2 revenue guide rather than a beat-and-raise, the print provides no offset to the spending escalation.
Questions for the Call
- Capex: what changed in 90 days? The $115–135B Q4 framing was given on January 28, 2026; the raise to $125–145B is given on April 29, 2026. Three months. Bullish: management identifies a specific incremental hyperscale-revenue contract or AI product launch that justifies the additional $10B as a forward-revenue commitment, not just a forward-capacity option. Bearish: the explanation is exclusively component pricing and "future year capacity," in which case the spending magnitude continues to compound without an external demand anchor — the same bear pillar from our Q3 downgrade.
- Reality Labs loss trajectory: is the peak confirmed? Q4 2025 framing called 2025 "likely the peak" for RL operating losses. Q1 2026 narrowed YoY by $182M. Bullish: Susan Li commits to FY26 RL operating loss being lower than FY25's. Bearish: management hedges with "lumpy quarter-to-quarter" or pre-launch-investment language that allows full-year losses to be flat or higher than 2025 — which would functionally retract the Q4 peak language.
- FoA segment-margin compression — structural or transitory? FoA op margin compressed ~400 bps YoY despite +33% revenue. Bullish: management frames the compression as front-loaded R&D against future ad-monetization gains (i.e., ROAS benefits show up in 2027). Bearish: management acknowledges this is the new normal — segment margin in the mid-to-high 40s rather than the low 50s — which would require resetting our long-term FoA OI model down ~$5–10B annualized.
- Buyback strategy: when does the bid return? Two consecutive quarters of zero buybacks. Bullish: management commits to a buyback resumption framework — e.g., once capex peaks or once a specific FCF threshold is cleared. Bearish: management defers indefinitely, framing capital allocation as "AI build first" without a clear waypoint for shareholder return resumption.
- Meta Superintelligence Labs first-model release — what is it? The CEO quote calls out "the release of our first model from Meta Superintelligence Labs" with no specifics in the release. Bullish: a clear positioning statement — open-weights or proprietary, parameter scale, benchmark performance, deployment surface (consumer Meta AI, ads platform, third-party API). Bearish: vague capability framing without competitive benchmarking, which would suggest the model is not yet competitive with frontier-tier peers.
- DAP weakness: how much of the Q/Q decline is Iran/Russia recoverable? The release attributes the Q/Q DAP softness to specific geopolitical events. Bullish: management quantifies the impact and frames it as transitory with normalization expected in Q2. Bearish: the impact is larger than implied or there is non-geopolitical engagement softness in mature markets bleeding through.
- Threads / Meta AI / Llama metrics — why no disclosure? Conspicuous absence of any of the AI/social product metrics that were disclosed in 2025 prints. Bullish: management provides updated MAU/DAU/inference-volume figures on the call. Bearish: management defers, which would be a tonal shift from the Q4 disclosure pattern.
Market Reaction
After-hours tape at time of writing is -7% (stock $625.44 vs. ~$669 close, per StockTitan; ~$118B of market cap removed). The reaction is consistent with our read: top line beat, FoA strong, but capex raise + Q2 in-line + DAP miss + buyback pause produce a "fine quarter, worse forward" composite. We expect the AH move to hold or extend modestly into the call; if Susan Li delivers a credible reframing of the capex raise (specific revenue commitment, hyperscaler/sovereign co-investment language), the move could narrow to -3% to -5% by tomorrow's close. If the call confirms that capex-without-offset is the new normal, the move could deepen to -8% to -10%.
Model Implications
| Item | Prior View (Q4 2025) | Post-Q1 2026 | Reason |
|---|---|---|---|
| FY26 Revenue | ~$245–253B | ~$248–255B | Q1 beat plus Q2 in-line guide; FX continues as tailwind through H1. |
| FY26 Operating Income | ~$84–88B (above 2025's $83.3B) | ~$85–90B | Q1 OI of $22.87B annualizes to ~$91B before seasonality; pledge intact with modest cushion. |
| FY26 Capex | ~$125B (midpoint of prior $115–135B) | ~$135B (midpoint of new $125–145B) | Direct pass-through of guide raise. |
| FY26 Free Cash Flow | ~$70B | ~$60–65B | $10B incremental capex flows directly to FCF reduction; partially offset by OI strength. |
| Reality Labs FY26 Loss | ~$17B | ~$15–16B | Q1 loss inside Q4's "peak" framing; first datapoint supporting the bend. |
| FY26 EPS Power (ex one-time tax items) | ~$26–28 | ~$28–30 | OI strength + share count flat; ex-tax-benefit Q1 EPS of $7.31 annualizes ~$29 against historical seasonality. |
| FY27 Capex Trajectory | "Higher than 2026" | "Higher than 2026" (no change) | Q1 raise pulls forward the spend question; FY27 framing not yet touched. |
Valuation: At ~$625 AH, META trades at ~21–22x our $29 mid FY26 EPS. The multiple has compressed materially from the ~28x level it held through most of 2025 — appropriate given the capex/buyback shift, but not punitive. Fair value range is roughly $600–700 depending on whether the FY27 capex trajectory bends or extends; the wide range reflects exactly the visibility gap that produced our Q3 downgrade. We see neither a compelling buy nor a compelling sell at $625.
Thesis Scorecard
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: AI investments translate into ad pricing power | Confirmed | Price/ad +12% on impressions +19% — the cleanest volume-and-price quarter of the cycle. Five consecutive quarters of confirmation. |
| Bull #2: Operating leverage holds through the capex cycle | Challenged | FoA op margin compressed ~400 bps YoY despite +33% revenue. Total op margin held at 41% only because RL loss narrowed; FoA-specific leverage is still moving the wrong way. |
| Bull #3: Meta AI / WhatsApp Business / Threads become real 2026+ monetization vectors | Confirmed | FoA Other +74% YoY to $885M. Six consecutive quarters of acceleration. |
| Bull #4: Absolute-dollar OI growth pledge for FY26 is credible | Confirmed (with margin) | Q1 OI of $22.87B (+30% YoY) annualizes to ~$91B before seasonality vs. 2025's $83.3B. Pledge has cushion at this run rate. |
| Bear #1: Capex compounds without external demand anchor | Confirmed | FY26 capex raised $10B at midpoint with no incremental forward-revenue commitment. Five quarters of escalation. |
| Bear #2: Buyback pause signals a fundamental capital-allocation shift | Confirmed | Second consecutive quarter of zero buybacks. The pause is now policy, not noise. |
| Bear #3: Reality Labs is open-ended-loss with no inflection | Rejected (partial) | Q1 loss narrowed YoY by $182M — first directional confirmation of the Q4 "peak" language. Need 2–3 more quarters to mark this fully rejected. |
| Bear #4: Youth-trial litigation produces material P&L impact | Neutral | "Material loss" language repeats; Q1 G&A is +15% YoY rather than the +386% Q4 spike. Reserve build is normalized but live. |
Overall: Thesis essentially balanced — one bear pillar (RL losses) softened, one bull pillar (operating leverage) further challenged, and the capex/buyback bear pillars confirmed at higher magnitude. The Q3 2025 downgrade thesis is intact. The Q4 framing of FY26 as the year that quantifies the spend bet is unchanged — Q1 just adds $10B to the bet without adding offsetting upside.
Preliminary Action: Hold (Maintained). Do not chase the AH selloff but do not sell into it either. The ~21–22x forward multiple is a reasonable entry for incremental positioning if the call reframes the capex raise credibly; it is a reasonable exit if the call confirms capex-without-offset as the new policy. The call is the swing factor — pending review.