Spending Bombshell Lands, Stock Rallies Anyway — Maintaining Hold as the OI-Growth Pledge Anchors a Larger Bet

Results vs. Consensus

Quality of the Beat

• Revenue: Genuinely strong — +24% reported, +23% constant currency (FX a 1pt tailwind in Q4). Family of Apps revenue $58.94B (+25%); advertising revenue $58.14B (+24%). Notably, the third consecutive quarter of high-20s growth followed by a fourth at +24% indicates the AI-engine growth wave is sustained, not decelerating. Full-year 2025 revenue $200.97B (+22%) — Meta crossed $200B for the first time.
• Margins: 41.3% operating margin in Q4 vs. 48% in Q4 2024 — a 700bps compression that dwarfs Q3's 290bps. The driver: total costs and expenses +40% YoY ($35.15B vs. $25.02B) on a 24% revenue line. R&D +41% ($17.1B vs. $12.2B), G&A +386% ($3.70B vs. $0.76B) on legal accruals. This is the operating-leverage pillar of the bull case continuing to break. Full-year 2025 op margin 41% vs. 42% in 2024 — only a 100bps compression, but the Q4 exit rate is materially worse than the FY average suggests.
• EPS: $8.88 diluted GAAP, +11% YoY. Effective tax rate 10% in Q4 (a true Q4 number, with 2026 guide of 13-16% — so Q4 was below run rate). Adjusted for normalized tax, EPS would have been closer to $8.10 — still a beat but less dramatic than the headline.
• Capex: $22.14B in Q4, taking 2025 total to $72.22B — at the upper end of the $70-72B guide. The Q4 figure is the highest single-quarter capex in Meta's history.
• Cash flow: $14.08B Q4 FCF; $43.59B for full-year 2025 — down from $52.10B in 2024. FCF compression of $8.5B YoY is the direct mechanical consequence of the capex ramp; this trajectory continues sharply in 2026.

Segment Performance

Family of Apps

FoA delivered a clean Q4: ad impressions +18% YoY (the strongest quarterly impression line in over two years), price per ad +6%, FoA "Other" revenue $801M (+54%) marking a fifth consecutive quarter of acceleration on WhatsApp paid messaging — which Susan Li disclosed crossed a $2B annual run rate in Q4. US click-to-message ad revenue grew over 50% YoY. Threads ads are now expanding to all remaining countries (UK, EU, Brazil) starting this month.

"Our community across the family of apps continues to grow. We estimate more than 3.5 billion people used at least one of our family of apps on a daily basis in December." — Susan Li, CFO

The product engagement story strengthened: Q4 optimizations drove a 7% lift in Facebook organic feed/video views — "the largest quarterly revenue impact from Facebook product launches in the past two years" per Susan Li. Threads time spent +20% on recommendation improvements. Instagram original-content prevalence grew 10pp to 75% of US recommendations. Meta AI daily media-generation actives tripled YoY.

Assessment: The operational engine is firing at the strongest pace of the AI cycle. The Q1 2026 revenue guide ($53.5-56.5B, midpoint $55B implies +18% YoY at constant currency, with FX a 4pt tailwind taking the reported guide to ~+22%) is consistent with continued strength — the "strong start to 2026" framing is genuine, not promotional. This is the unambiguous bull pillar of the print.

Reality Labs

RL revenue $955M (-12% YoY), tracking the Q3 disclosure that Q4 would lap Quest 3S launch and the Q3 channel-fill pull-forward. Operating loss $6.02B in Q4 — wider than the ~$4.4B run rate of recent quarters and the largest single-quarter RL loss to date. Full-year 2025 RL operating loss of $19.19B vs. $17.73B in 2024. Zuckerberg disclosed that AI-glasses sales "more than tripled" in 2025, and committed to a strategic shift: "For Reality Labs, we are directing most of our investment towards glasses and wearables going forward, while focusing on making Horizon a massive success on mobile and making VR a profitable ecosystem over the coming years."

"I expect Reality Labs losses this year to be similar to last year, and this will likely be the peak as we start to gradually reduce our losses going forward while continuing to execute on our vision." — Mark Zuckerberg, CEO

Assessment: First time Zuckerberg has called a "peak" on RL losses. If credible, this is a meaningful long-term thesis-positive disclosure — the open-ended-losses bear point has finally been bracketed. The credibility test is whether 2026 RL operating loss actually lands at ~$19B rather than expanding again. The mobile-Horizon pivot is a notable strategic concession that VR-first never reached the user base it needed.

Key KPIs

Two notable shifts: (1) impression growth accelerated to +18% from +14% — the highest of the cycle, indicating the recommendation system improvements continue to compound; (2) price per ad decelerated to +6% from +10% — Susan Li attributed this to "improved ad performance and higher advertiser demand" but the deceleration alongside accelerating impressions implies inventory mix shift toward newer, lower-priced placements (Threads ads, WhatsApp, click-to-message). Net revenue effect remains strongly positive but the composition matters for forward modeling.

Key Topics & Management Commentary

Overall Management Tone: Markedly more confident than Q3. Zuckerberg led with "we are now seeing a major AI acceleration" and framed 2026 as the year products and revenue inflect together. Susan Li was less defensive than Q3 — her capex/opex disclosure was direct and accompanied by a hard OI-growth commitment, which inverts the Q3 dynamic where the spending escalation came without offsetting commitments. The +10.4% stock reaction reflects the market reading the OI-growth pledge as a credible counterweight to the spend.

2026 Capex and Opex: Quantified — and Even Larger Than Q3 Implied

This is the hinge of the print. Q3 framing: "notably larger than 2025" capex; "significantly faster percentage rate" opex. Q4 quantification:

• 2026 capex: $115-135B (midpoint $125B). 2025 actual: $72.22B. Dollar step-up: $43-63B (vs. $30B step-up from 2024 to 2025). Midpoint step-up of $53B is 77% larger than 2025's step-up — "notably larger" was, if anything, an understatement.
• 2026 opex: $162-169B (midpoint $165.5B). 2025 actual: $117.69B. Growth rate: 38-44% (vs. 24% in 2025). The "significantly faster percentage rate" framing translates to opex growth nearly double 2025's pace.

"We anticipate 2026 capital expenditures, including principal payments on finance leases, to be in the range of $115 to $135 billion, with year-over-year growth driven by increased investment to support our Meta Superintelligence Labs efforts and core business." — Susan Li, CFO

This was the moment the bear math from Q3 needed to be either contradicted or confirmed. It was confirmed at the high end of our Q3 model range ($110-130B) and arguably above it.

Assessment: If this disclosure had landed without an offsetting commitment, the rating action today would be a downgrade to Underperform. The reason it isn't is the OI-growth pledge below — and the stock's willingness to digest it.

The Operating Income Floor

Susan Li's most consequential single sentence on the call:

"Despite the meaningful step up in infrastructure investment, in 2026 we expect to deliver operating income that is above 2025 operating income." — Susan Li, CFO

She immediately clarified that this is an absolute-dollar commitment, not a growth-rate one — meaning 2026 OI > $83.28B. Working backward: with opex midpoint of $165.5B, 2026 revenue must clear ~$248.8B for OI to hit $83.3B at the implied gross-margin profile, and substantially more if depreciation (the largest expense growth driver) outpaces our model. Consensus 2026 revenue going into the print was ~$233-238B; the OI commitment effectively forces the buyside to mark up 2026 revenue assumptions or to assume management is highly confident in the cost-base mix.

Implicit in this commitment: management has visibility into 2026 revenue trajectory tight enough to make an absolute-dollar OI promise on a $165.5B cost base. That is meaningful — Susan Li does not make commitments she expects to retract. The Q1 2026 revenue guide of $53.5-56.5B (midpoint +14% YoY at reported, ~+10% constant currency) supports the math — annualized that's a ~$220B base before any acceleration.

Assessment: This is the soft floor that prevents a further downgrade. It transforms the 2026 capex/opex disclosure from a runaway-spending story into a "spending plus revenue acceleration" story — which is investable, just not at a premium multiple. We require Q1 2026 results to validate the pledge before considering an upgrade.

MetaCompute and Capital Structure Flexibility

Zuckerberg formally introduced "MetaCompute" — Meta's AI-infrastructure organization — and announced Dina Powell McCormick joined as president and vice chairman to lead partnerships with "governments, sovereigns, and strategic capital partners to expand our long-term capacity."

On the capital side, Susan Li confirmed Meta will continue to "look for opportunities to periodically supplement our strong operating cash flow with prudent amounts of cost-efficient external financing, which may lead us to eventually maintain a positive net debt balance." Q4 long-term debt rose to $58.74B from $28.83B at year-end 2024 — the company issued $29.9B of new long-term debt in Q4 alone, and buybacks went to zero in the quarter ($26.25B for full-year, down from $30.13B in 2024). The composition of capital allocation has clearly shifted from buybacks to debt-funded capex.

"I recently announced MetaCompute with the belief that being the most efficient at how we engineer, invest, and partner to build our infrastructure will become a strategic advantage. We will continue working with key partners while advancing our own silicon program. We're architecting our systems so that we can be flexible in the systems that we use, and we expect the cost per gigawatt to decrease significantly over time." — Mark Zuckerberg, CEO

Assessment: Two distinct moves here. First, Meta is openly preparing for sovereign and strategic-partner co-investment in data center capacity — a structural admission that even Meta's balance sheet cannot fund standalone the spend trajectory it is committing to. Second, the buyback step-down to zero in Q4 (from $13.4B in Q1, $9.76B in Q2, $3.16B in Q3) confirms the per-share-tailwind bear point from Q3: capital return is no longer a thesis pillar through this cycle. Debt-funded capex is now the model.

Ad-Side AI ROI: Continuing Compounding

The ROI evidence remained substantive in Q4:

• GEM extended to Facebook Reels; now covers all major Facebook and Instagram surfaces. Training cluster size doubled in Q4. 2026 plan: scale to "an even larger cluster" with new sequence learning architectures.
• New runtime ad-ranking model launched across Instagram feed, stories, and Reels in Q4 — drove a 3% conversion rate increase.
• Lattice consolidation brought Facebook stories and other surfaces into the unified Facebook model; Susan Li disclosed this drove a 12% increase in ad quality. 2026 plan: "consolidate more models than we had in the prior two years."
• WhatsApp paid messaging crossed $2B annual run rate; US click-to-message ad revenue +50% YoY.
• AI coding tools internal productivity: engineer output +30% since the start of 2025; power users +80% YoY. Susan Li expects this growth to "accelerate through the next half."

Assessment: The bull-case data points continue to mount, and the ad-side model architecture (GEM at the foundation, lighter-weight inference downstream, Lattice consolidating runtime models) now looks coherent rather than experimental. The challenge remains the same as Q3: the next $50B of incremental capex needs to produce proportional ad-side returns to justify the magnitude. The runtime-model 3% conversion lift and 12% ad quality lift are meaningful but small relative to the implied 2026 spending step.

Litigation: Material Loss Language Repeated, G&A Already Reflecting It

The youth-trial disclosure that was new in Q3 was repeated verbatim in the Q4 outlook:

"We continue to see scrutiny on youth-related issues and have a number of trials scheduled for this year in the U.S., which may ultimately result in a material loss." — Susan Li, CFO

The income statement is now reflecting the buildup. Q4 G&A was $3.70B vs. $0.76B in Q4 2024 — a 386% YoY increase. Susan Li attributed the legal expense growth specifically to "lapping legal accrual reversals in '24 and charges recorded in Q4 2025." Full-year 2025 G&A was $12.15B vs. $9.74B in 2024 — a $2.4B step-up, most of it concentrated in Q4. Without specifying the figures, this materially raises the implied legal-accrual reserve for 2026 trials.

EU LPA: management disclosed alignment with the European Commission on "further changes to our Less Personalized Ads offering, which we will begin rolling out this quarter" (Q1 2026). This converts the three-quarter "as early as next quarter" overhang into actual rollout — finally quantifiable in subsequent quarters. Susan Li flagged "potential headwinds from revised EU ad offerings as factors that may reduce growth rates later in the year."

Assessment: The G&A jump is the first concrete income-statement impact from the youth-litigation track. If 2026 G&A continues to absorb additional accruals at this pace, full-year 2026 G&A could approach $15-18B. The EU LPA finally moving from "negotiated" to "rolling out" removes one source of opacity but introduces a quantifiable revenue headwind starting in Q2 2026.

Glasses, Horizon, and the Hardware Pivot

Zuckerberg explicitly committed to a hardware pivot toward AI glasses:

"Sales of our glasses more than tripled last year, and we think that they're some of the fastest-growing consumer electronics in history. Billions of people wear glasses or contacts for vision correction, and I think that we're at a moment similar to when smartphones arrived. It was clearly only a matter of time until all those flip phones became smartphones. It's hard to imagine a world in several years where most glasses that people wear aren't AI glasses." — Mark Zuckerberg, CEO

The Horizon-on-mobile pivot is the more telling strategic concession: Meta is acknowledging that VR-first never reached the user base needed to justify the standalone investment, and is pivoting Horizon to where consumers actually are. Combined with the "peak losses" framing, this is the most coherent RL strategic posture Meta has articulated in the AI cycle.

Assessment: Net positive on RL. The peak-losses commitment plus glasses 3x in 2025 plus Horizon on mobile constitutes a real strategic narrative. We continue to value RL at zero in our model — but the path to it being net-additive over a 3-5 year horizon is more credible after this print than at any prior point.

Guidance & Outlook

The Q1 2026 revenue guide of $53.5-56.5B implies +10-16% reported YoY growth on Q1 2025's $42.31B (which would be the slowest growth quarter since Q4 2023, but Q1 has FX as a 4pt tailwind, so constant-currency growth is ~+6-12%). This is a deceleration from Q4's +24% — but management framed it as conservatism: "the range embeds an outlook for accelerated growth, and that's really underpinned by the strong demand that we saw through the end of Q4 and continuing into the start of 2026" (Susan Li).

Implied full-year 2026 revenue: If Q1 lands at midpoint ($55B) and the year compounds, FY26 revenue likely tracks $245-260B on the bull case — required to support the OI-growth commitment given the cost base. Consensus going in was ~$235B; this print reasonably moves consensus toward $245-250B.

Street at: Pre-print Q1 consensus was $53.5B; the guide midpoint is $55B — a 2.8% raise. The +10.4% stock reaction reflects this upward revision plus the OI-growth pledge plus the relief of finally quantifying the spend.

Guidance style: Aggressive on Q1 revenue (above consensus), aggressive on the OI commitment, aggressive on quantifying spend. This is a different posture than Q3 — management has clearly decided the way through the spending escalation is to commit publicly to absolute outcomes, not to leave the spending unmoored from any earnings anchor.

Analyst Q&A Highlights

2026 Operating Income Math

• Doug Anmuth (JPMorgan) asked whether Meta expects positive free cash flow in 2026 alongside the OI-growth commitment, and how to think about JV and partnership structures for data centers. Susan Li reaffirmed the OI commitment and indicated the business will "generate sufficient cash to fund our infrastructure investments in 2026" but explicitly opened the door to incremental external financing and "new ownership structures for some of our large data center sites."
  Notable: She did not commit to positive 2026 FCF. The implication is that 2026 FCF could go negative or near-zero if capex lands at the high end of $135B. This is a meaningful concession.
• Brian Nowak (Morgan Stanley) pressed on the drivers of accelerated 2026 growth. Susan Li pointed to FX tailwinds, ad-performance investment compounding, advertiser response to ranking improvements, and Advantage Plus adoption.
  Assessment: The drivers are credible but Q1's 4pt FX tailwind is a one-time effect; underlying constant-currency growth is the metric to track quarter-by-quarter through 2026.

Capacity Constraints and Compute

• Eric Sheridan (Goldman Sachs) asked whether Meta is still capacity-constrained internally. Susan Li confirmed: "We do continue to be capacity constrained." She framed the 2026 capex as still meeting demand, not creating slack.
  Notable: This is a key bull-case data point. If the company is still demand-constrained at $115-135B of 2026 capex, the marginal capex dollar is still earning a return. The bear case requires capacity to overshoot demand; that has not happened yet.
• Sheridan also asked whether the ads business has fully captured first-order returns from compute scaling. Susan Li: "We think that there is room for our larger models to benefit from having more compute" — i.e., the ad-side return curve has not flattened. This reinforces the marginal-ROI picture.

MSL Progress and Frontier Models

• Doug Anmuth and Mark Shmulik (Bernstein) asked for an MSL update and a path to a frontier model in 2026. Zuckerberg was explicit about offering "somewhat unfulfilling" answers, framing the period as one of rebuilding and noting initial models will roll out over the year. He emphasized the long-term nature of the work.
  Assessment: Six months into the MSL effort, no public-facing model output yet. The credibility test is mid-2026 model releases. Until then, the MSL investment is unmeasurable from outside.
• Ken Gawrelski (Wells Fargo) asked whether a leading general-purpose model is critical or whether use-case-specific capability suffices. Zuckerberg argued strongly for owning the underlying technology: "Frontier AI for many reasons, some competitive, some safety-oriented, are not going to always be available through an API to everyone."
  Assessment: Strategic logic is sound; financial logic depends on whether MSL produces a competitive frontier model on a useful timeline.

Capital Return

• Mark Mahaney (Evercore) noted Q4 had no buybacks and asked about capital return. Susan Li did not directly commit to resuming buybacks in 2026; the implication is that buybacks remain de-prioritized through the capex peak.
  Assessment: The per-share tailwind from buybacks is gone for the foreseeable future. Capital return through this cycle is the dividend ($5.32B in 2025) plus whatever remaining capacity FCF allows.

What They're NOT Saying

1. FY2026 revenue range: Susan Li gave Q1 explicitly but declined to provide a full-year revenue range despite the OI commitment effectively requiring one. The OI-growth pledge implicitly anchors FY26 revenue to a range, but management chose not to make that explicit.
2. 2026 operating margin: Despite committing to OI growth in absolute dollars, no operating margin disclosure. With opex midpoint $165.5B and OI midpoint > $85B, implied revenue is > $250B and operating margin lands ~34%, down from 41% in 2025. The 700bps compression that Q4 showed in isolation looks like it could become the FY26 reality.
3. Specific MSL deliverables: Zuckerberg explicitly declined to share specific milestones or quantitative deliverables — "more about showing the trajectory that we're on rather than being a single moment in time."
4. Youth-trial accrual size: The Q4 G&A jump signals significant accruals were taken, but no specific reserve amount disclosed. Estimating from the YoY G&A delta, accruals likely landed in the $1.5-2.5B range in Q4 alone.
5. EU LPA quantified impact: Now that the new offering is rolling out in Q1 2026, the impact will become observable — but management gave no impact estimate. Likely a 200-400bps drag on European ad revenue starting Q2 2026.
6. 2026 free cash flow: Susan Li conspicuously did not commit to positive FCF in 2026. With operating cash flow tracking ~$130B+ and capex midpoint $125B, FCF could be $0-15B — a fraction of 2025's $43.6B. This is a major change in the financial profile that was not flagged directly.
7. Specific MetaCompute partnerships: Powell McCormick joining and "sovereigns and strategic capital partners" are framed but no specific deals disclosed. The structures themselves are the new disclosure; the scale and counterparties remain to come.
8. RL operating loss path beyond 2026: Zuckerberg said losses will "gradually reduce" post-2026 but did not commit to a specific multi-year trajectory. The "peak" framing is the key new disclosure but its credibility is unverified.

Market Reaction

• Post-print move: +10.4% in the session following the print, adding ~$177B of market cap and taking Meta to ~$1.88T.
• Volume: ~1.5x average.
• Intraday peak: +6.9% before settling higher into the close.
• Sector context: Meta marginally up while peers GOOG (-0.19%), GOOGL (-0.24%), SNAP (-1.84%), SPOT (-0.15%), and BIDU (-1.09%) were all down — confirming the move was stock-specific to the print, not a sector tailwind.
• Analyst tone: Constructively repositioned — the OI-growth commitment broke the spending-overhang narrative for the bull camp. Most desks raised 2026 revenue estimates while flagging the sharp implied margin compression as the next risk to track.

This is a noteworthy reaction because the spending disclosure was, in absolute terms, worse than the bear case had positioned for ahead of the print (consensus 2026 capex was ~$95-105B; the actual was $115-135B). The +10.4% rally despite that is the market choosing to weight the OI-growth pledge and Q1 revenue guide above the spending magnitude. The implicit market view: if Meta delivers $85B+ of OI in 2026, the multiple compression that Q3's downgrade priced in is excessive. We are not yet willing to underwrite that bet without Q1 evidence — but the reaction does set a higher bar for further negative repositioning.

Street Perspective

Debate: Is the OI-Growth Commitment Credible?

Bull view: Susan Li doesn't make absolute-dollar commitments she expects to retract. The Q1 revenue guide above consensus, the still-capacity-constrained framing, and the 18% Q4 impression growth all suggest revenue acceleration is real. If revenue lands at $250B+ with disciplined cost composition, OI of $85-90B is achievable on top of the stepped-up cost base.

Bear view: The math is hostile. Opex of $165B + ~$45B of infrastructure depreciation (driven by the capex ramp) + the existing depreciation base produces a cost structure that requires revenue growth meaningfully above consensus. Any miss on the revenue side — from EU LPA, youth-trial settlement noise, or a normal 2026 macro wobble — sends OI below the 2025 base, breaking the commitment.

Our take: Coin flip leans bull-side. Meta has earned credibility on absolute-dollar guidance (the $116-118B 2025 opex guide landed at $117.7B). But the margin of safety is thin, and the sensitivity to revenue downside is the highest it has been in the AI cycle. We don't underwrite the commitment as a high-confidence bet; we treat it as a soft floor that prevents catastrophe rather than a guarantee of upside.

Debate: Is the AI Glasses Story Investable?

Bull view: Sales tripled in 2025. Zuckerberg's smartphone-arrival analogy, while marketing language, has internal logic — the form factor is genuinely new, and Meta is pacing well ahead of competitors. Peak RL losses in 2025 means the segment is no longer an unbounded liability.

Bear view: "Tripled" is on a small base. Ray-Ban Meta unit numbers still not disclosed. The smartphone analogy is being used by every AI hardware company. Until RL revenue clears $5B annual run rate (currently ~$2.2B FY25), this is a cost center.

Our take: Bull case is incrementally stronger after this print but still not investable as a primary thesis driver. The glass form factor is the most credible Meta hardware bet to date — but valuation does not require it to work for the stock to perform.

Debate: Litigation Reserve — Is Q4's G&A Jump the Bottom or the Start?

Bull view: Q4 G&A absorbed material accruals; if the youth-trial calendar resolves in 2026 with reserve adequacy, 2027 G&A normalizes and provides a margin tailwind. Tech-sector litigation tends to settle below initial reserve estimates.

Bear view: Q4's $3.70B G&A is likely the start, not the end. Multiple state AG actions plus federal scrutiny plus EU enforcement creates a multi-year G&A overhang. Tobacco-precedent thinking suggests reserves keep building.

Our take: Bears more likely correct in 2026. The Q4 jump is concentrated in a single quarter; if the litigation calendar is back-loaded, 2026 quarterly G&A is unlikely to step down materially.

Model Update Needed

Valuation impact: Net neutral to slightly positive vs. our Q3 model. The OI-growth commitment plus higher implied revenue offsets the larger capex/opex base; the tax-rate normalization (13-16% vs. 30% reported in 2025) drives EPS higher than the operating-margin-only view would suggest. The risk profile around the model has widened — more upside if revenue lands at $260B+, more downside if it stalls at $235B.

Thesis Scorecard Post-Earnings

Overall: Thesis is approximately balanced. Two new pillars added — Bull #5 (OI-growth commitment as soft floor) and Bear #6 (capital-return suspension confirmed). Bear #2 (RL) materially improved with the peak-losses framing. Bear #1 (capex magnitude) confirmed at the high end of our model; Bear #5 (youth trials) now showing in the income statement. Net: the bull case has gained a credible commitment but lost the operating-leverage pillar definitively; the bear case got the spending magnitude it modeled but lost the open-ended-RL-loss point.

Action: Maintain Hold. The case for an upgrade requires Q1 2026 to validate the OI-growth pledge and the ad-revenue acceleration trajectory; the case for a downgrade to Underperform requires a meaningful revenue miss on Q1 or a youth-trial settlement at the high end of the reserve range. We expect to be testing both directions through 1H 2026.