META · Q3 2025 Earnings

Cautious

Headline numbers

EPS

Q3 FY2025

$1.05

Revenue

Q3 FY2025

$51.24B

+26.0% YoY

Gross margin

Q3 FY2025

82.0%

Free cash flow

Q3 FY2025

$10.63B

Operating margin

Q3 FY2025

40.0%

Key financials

Q3 FY2025

Metric	Q3 FY2025	YoY	Q2 FY2025	QoQ
Revenue	$51.24B	+26.0%	$47.52B	+7.8%
EPS	$1.05	—	$7.14	-85.3%
Gross margin	82.0%	—	82.1%	-10bps
Operating margin	40.0%	—	43.0%	-300bps
Free cash flow	$10.63B	—	$8.55B	+24.3%

Guidance

Q3 revenue beat guide by ~1.5%; FY2025 capex and expense guidance raised at low end, signaling accelerated infrastructure investment.

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q3 FY2025	$47.5-50.5 billion	$51.242 billion	+$0.74 billion above guide high end	Beat

New guidance

Metric	Period	Guide	YoY
Operating Margin	Q3 FY2025	40.0%	—
Revenue	Q4 FY2025	$56-59 billion	—
Tax rate	Q4 FY2025	12-15%	—

Changes to prior guidance

Metric	Period	Prior guide	New guide	Δ	Result
Total expenses	FY2025	$114-118 billion	$116-118 billion	Low end raised from $114B to $116B (+$2B); high end reaffirmed at $118B	Raised
Capital expenditures	FY2025	$66-72 billion	$70-72 billion	Low end raised from $66B to $70B (+$4B); high end reaffirmed at $72B	Raised

Segment performance

Q3 FY2025

Segment	Q3 FY2025	YoY
Family of Apps	$50.772B	+25.8%
Reality Labs	$0.47B	+74.1%
Advertising Revenue	$50.08 billion	—
Advertising Revenue YoY Growth	+26% (25% ex-FX)	—

Platform metrics

Q3 FY2025

Segment	Q3 FY2025
Family Daily Active People (DAP)	3.54 billion (Sept 2025)
Ad Impressions YoY Growth	+14%
Average Price Per Ad YoY Growth	+10%
Headcount	78,450 (+8% YoY)

Profitability

Q3 FY2025

Segment	Q3 FY2025
Operating Margin	40.0%
Free Cash Flow	$10.63 billion

Management tone

Narrative arc: Q2 "2026 will be similarly significant" → Q3 "2026 will be significantly faster." The capex commitment has hardened from a directional preview into a concrete pre-commitment, while the language around 2026 expenses has escalated meaningfully in just one quarter.

In Q2, Li's framing of 2026 capex was "similarly significant dollar growth" — a qualitative anchor. In Q3, the language shifted to "further upward pressure on capital expenditures and expense plans next year" with expenses growing at a "significantly faster percentage rate." From Susan Li in the press release: "We anticipate total expenses will grow at a significantly faster percentage rate in 2026 than 2025, with growth driven primarily by infrastructure costs." The shift signals that internal capacity planning has moved beyond the ranges contemplated three months ago — and management chose to telegraph this before guidance is formally set, presumably to absorb the reaction over multiple quarters rather than at the FY2026 guide drop.

The framing of ROI evidence has also shifted from defensive to specific. In Q2, Li was asked about Gen AI revenue and said it wouldn't be "meaningful this year or next" — a hedge. In Q3, responding to Brian Nowak's CapEx ROI question, Li offered concrete operational evidence: "value-weighted conversion rates are showing very strong year-over-year growth, with weighted conversions growing faster than impressions." This is a more confident posture — pointing to measurable lift in the ad system attributable to AI investment, rather than asking for patience.

A new and important admission emerged in Doug Anmuth's exchange: infrastructure currently underinvests both the core business and MSL — management explicitly said both are compute-constrained with ROI-positive opportunities going unfunded. This reframes the capex super-cycle from "discretionary frontier bet" to "catching up to demand." It's a more defensible framing for the cost trajectory, but it also implies the spending floor is higher than previously communicated.

On Reality Labs, the Q4 RL revenue headwind disclosure (lapping Quest 3S) was acknowledged unprompted in the press release — a small but notable shift toward proactively flagging composition of growth rather than letting analysts triangulate it. Ray-Ban Meta and Oakley Meta were repeatedly cited as selling well, but Zuckerberg's "hundreds of millions to billions of users for sustained profitability" framing for wearables makes clear this is a 5+ year horizon, not a 2026 story.

Q&A highlights

Brian Nowak · Morgan Stanley

What early quantifiable signals from A/B tests on core AI improvements in 2026 demonstrate ROIC confidence on infrastructure CapEx? How large is the Reality Labs revenue headwind in Q4 guidance?

Susan detailed that value-weighted conversion rates are showing very strong year-over-year growth, with weighted conversions growing faster than impressions, demonstrating ad performance improvements. New model architectures enable leverage of more data and compute. Reality Labs revenue headwind in Q4 not specifically quantified but attributed to lapping Quest 3S introduction and retail Q3 procurements; expects significant AI glasses revenue growth will not fully offset Quest headsets declines.

Value-weighted conversion rates showing strong YoY growthWeighted conversions growing faster than impressionsQ4 Reality Labs revenue expected lower YoY due to Quest 3S lap and retail timing shiftAI glasses revenue expected to grow significantly in Q4

Doug Anmuth · JP Morgan

How do you triangulate CapEx dollar growth and significantly faster expense growth in 2026 with core business growth impact on earnings and free cash flow? Do you have targets for cash on hand or net cash?

Susan and Mark explained they are still in early budget planning for 2026 with dynamic capacity planning. Core business and MSL both have ROI-positive use cases for additional compute. Infrastructure underinvestment currently constrains both MSL and core business opportunities. No specific targets for cash/net cash disclosed; strategic priority is ensuring sufficient compute capacity for AI success. Worst case involves pre-building for couple years with depreciation absorbed as growth occurs.

Still in early 2026 budget planning processBoth MSL and core business compute-constrained with ROI-positive opportunitiesCapEx dollar growth expected notably larger in 2026 vs 2025Total expenses expected to grow at significantly faster rate in 2026 vs 2025

Eric Sheridan · Goldman Sachs

What consumer signals are you seeing on Meta AI adoption today, and how will scaling frontier models from MSL efforts change utility and behavior around Meta AI in years ahead?

Mark noted Meta AI has over 1 billion monthly users and sees clear correlation: as models improve, usage increases. Frontier models from MSL will unlock new products and capabilities beyond assistant functionality. Meta's track record scaling new products to billions of users positions it well. Novel capabilities across reasoning, video generation, and content creation will apply to assistants, business AI, advertising tools, and feed ranking/recommendations.

Over 1 billion monthly Meta AI usersClear correlation between model quality improvements and usage growthFrontier MSL models expected to drive new products and capabilitiesNovel capabilities will apply across multiple product surfaces and use cases

Mark Shmulek · Bernstein

How does the scale of ad performance and engagement improvements expected in 2026 compare to progress over last two years? Should we expect frontier model launches from MSL in 2026, or focus on products like Vibes?

Susan explained Meta doesn't use large foundation models like GEM directly for inference due to cost; instead transfers knowledge to lightweight runtime models. Even single-digit percentage improvements in ads performance and basis-point improvements in conversions drive meaningful absolute revenue growth on large advertiser base. Mark declined to provide specific timing on frontier models or products but expects both new models and novel products will be released; emphasized sharing more detail when ready.

Large foundation models transfer knowledge to lightweight inference models for cost efficiencySingle-digit percentage improvements in ads conversion drive material revenue growthBasis-point improvements in performance metrics yield meaningful absolute dollarsBoth new frontier models and novel products expected but no specific timing provided

Yusuf Squali · Truist Securities

Will Ray-Ban Meta/Oakley Meta hardware sales recoup investments or is success dependent on new revenue avenues like advertising and commerce? What about on-balance vs. off-balance sheet financing of AI initiatives, including Blue Owl Louisiana deal impact on CapEx guidance?

Mark stated wearables investment will eventually be very profitable through device sales plus services and AI revenue; AI will become primary value driver as products mature. Long-term target is hundreds of millions to billions of users for full profitability. Susan explained Blue Owl JV provides external capital partnership for data center co-development, giving optionality. Prior CapEx included portion of data center costs; going forward, Meta contributes 20% of remaining construction costs (matching ownership stake) recorded as investing cash flows, not CapEx.

Ray-Ban Meta and Oakley Meta products selling well post-launchWearables expected eventually very profitable from devices plus services and AIAI expected to become primary value driver for wearables over timeTarget scale: hundreds of millions to billions of users for sustained profitability

Answers to last quarter's watch list

Q3 revenue landing within or above the $47.5–50.5B guide range — $51.24B (+26% YoY), $740M above the high end. The ad business not only landed above guide but accelerated YoY growth from +22% to +26%. Status: Resolved positively

Ad pricing growth sustaining above high-single digits — Pricing +10% YoY (up from +9% in Q2), with impressions also accelerating to +14% (from +11%). Both legs of the engine improved simultaneously, the healthiest possible answer. Status: Resolved positively

FY2026 capex range disclosure — No specific dollar range provided. Management instead telegraphed "further upward pressure" on capex and "significantly faster" expense growth than 2025's 22–24% — directionally bracketing 2026 above current consensus but without committing to a number. The market-moving disclosure remains pending. Status: Continue monitoring

Reality Labs operating loss trajectory — RL revenue $470M (+74% YoY) but the operating loss line wasn't called out in the extraction, and Q4 RL revenue was flagged as down YoY (lapping Quest 3S). The Blue Owl JV treatment moves some infrastructure spend off the CapEx line into investing cash flows, complicating the loss-line read going forward. Status: Continue monitoring

Concrete announcement on data-center co-development partnerships — The Blue Owl JV for the Louisiana data center is the first named structure. Meta contributes 20% of remaining construction costs as investing cash flows, not CapEx. Material capital-structure signal: off-balance-sheet financing is now operational, not just being explored. Status: Resolved positively

Q4 revenue deceleration magnitude — Q4 guide $56–59B (midpoint $57.5B). Against the just-reported $51.24B that's +12% QoQ at midpoint; against Q4 2024's $48.4B that's roughly +19% YoY at midpoint — a step-down from Q3's +26% but still strong. The press release attributed the deceleration to lapping tougher comps and lower YoY RL revenue, not core ad softness. Status: Resolved positively

What to watch into next quarter

FY2026 expense and capex ranges when formally guided on the Q4 call — Q3 telegraphed "significantly faster" than 22–24% expense growth and "further upward pressure" on capex. Watch whether the formal 2026 capex range starts with an 8 or 9 (vs. 2025's $70–72B) — the single most market-moving disclosure of next quarter.

Q4 revenue landing within $56–59B with ad pricing sustaining at or above +10% YoY — anything below midpoint with the explicit RL headwind already telegraphed would suggest core ad demand softening. Pricing dropping back to mid-single digits would re-open the volume-mix question.

Operating margin trajectory — 40% in Q3 vs. 43% in Q2 is the first visible margin compression from the capex ramp. Whether Q4 margin stays at or above 40% (despite the heavier expense run-rate) determines if the 2026 setup is "managed compression" or "accelerating compression."

Blue Owl JV scale and subsequent off-balance-sheet structures — first quarter with the JV mechanic operational; watch the disclosed split between CapEx and investing cash flows in the 10-Q, and any announcements of additional partner structures that would systematically understate reported CapEx growth.

Reality Labs Q4 revenue decline magnitude and operating loss — Q4 RL revenue is guided down YoY; the loss line will determine whether AI glasses momentum is materially offsetting Quest declines or just optically improving the mix.

Any quantification of Meta AI commercialization — Zuckerberg flagged 1B+ monthly users but provided no revenue contribution. First disclosure of monetization (ads in AI, business AI revenue) would be a meaningful new line of sight on Gen AI ROI.

Sources

Meta Q3 2025 Press Release (Form 8-K Exhibit 99.1), filed October 29, 2025 — https://www.sec.gov/Archives/edgar/data/1326801/000162828025047114/meta-09302025xexhibit991.htm
Meta Q3 2025 Earnings Call Q&A (transcript excerpts)
Meta Q2 2025 Tapebrief (prior quarter watch list and guidance baselines)

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.