Meta Q3 2026 Earnings Results: Revenue Beats, Profit Hit
Meta's Q3 2026 earnings results beat revenue expectations at $64.6B (+26%), but a $9.89B legal charge cut EPS to $3.83. What the numbers really mean.

Key Takeaways
- Revenue hit $64.6B, up 26% year over year, beating the roughly $63.3B consensus — the ad business is still growing fast.
- A one-time $9.89B legal charge over the August 2026 settlement with 52 state attorneys general cut GAAP EPS to $3.83; excluding it, EPS would have been $7.40.
- Meta narrowed 2026 capex guidance upward to $135–145B and spent $34.9B in Q3 alone, pushing free cash flow to -$2.1B.
- Zuckerberg called Muse and Muse Wonder 'our first frontier model'; more than 10M people in the US and Canada set up Muse agents in about six weeks.
- Q4 revenue guidance of $73–76B came in above consensus, and the stock rose about 2–4% in after-hours trading.
Meta Q3 2026 Earnings Results: Revenue Beats, Profit Hit
This is the follow-up to yesterday's earnings preview — the actual numbers are out, and here is what they changed.
Meta Platforms reported its third-quarter 2026 results after the US market close on Wednesday, October 21, and the quarter can be summed up in two numbers: $64.6 billion in revenue — a 26% jump and a clear beat — and $9.9 billion in one-time legal costs that gutted the headline profit figure. The real story, though, is the number management keeps raising: Meta now expects to spend $135–145 billion on AI infrastructure this year, and it is reshaping its share structure to keep the money flowing.
The headline numbers
| Metric | Q3 2026 | Year over year | |---|---|---| | Revenue | $64.6B | +26% | | Advertising revenue | $63.1B | +26% | | Family of Apps revenue | $64.2B | +26% | | Total expenses | $51.6B | +68% | | GAAP operating income | $13.0B (20% margin) | -37% | | Legal settlement charge | $9.89B | one-time | | GAAP diluted EPS | $3.83 | distorted by tax-base effect | | Adjusted EPS (ex-legal charge) | $7.40 | vs $7.25 a year ago | | Capex (incl. finance leases) | $34.9B | +80% | | Free cash flow | -$2.1B | first negative quarter since IPO | | Daily active people | 3.64B | +3% | | Reality Labs revenue | $443M | -6% | | Reality Labs operating loss | $4.1B | narrower than expected |
Revenue beat the analyst consensus of roughly $63.3 billion by close to 2%. On paper the EPS headline looks ugly, but it needs context: a year ago Meta paid an effective tax rate of 87%, so the reported EPS "surge" this quarter is mostly a tax quirk, not a profit boom. The honest comparison is adjusted EPS of $7.40 versus $7.25 in the same quarter last year — modest growth, but growth.
The $9.89 billion legal bill, explained simply
Every reader's first question: how does a company post record revenue and a halved operating margin at the same time?
In August 2026, Meta reached a settlement with 52 US state attorneys general over lawsuits concerning teen mental health and social media. Meta booked $9.89 billion of that settlement as an expense in Q3. The effect, in plain terms:
- Operating income dropped from what would have been $22.88 billion to $12.99 billion.
- The operating margin fell from 40% a year ago to 20%.
- Diluted EPS fell from an adjusted $7.40 to a reported $3.83 — a roughly $3.57-per-share hit.
As CFO Susan Li said on the call, excluding the charge, operating income would have risen about 11% year over year. That is the number the business was actually running at; the charge is a one-time legal event, not a deterioration of the ad machine. But it is real cash leaving the company, which is why investors read it carefully.
The $135 billion AI bet keeps getting bigger
Meta spent $34.9 billion on capital expenditure in Q3 alone — up 80% from a year ago — on servers, data centers, and network infrastructure. That pushed free cash flow to negative $2.1 billion, the first negative quarter since Meta went public. It was still much better than analysts' feared negative $9.4 billion, and operating cash flow remained a healthy $32.8 billion.
More importantly, Meta raised the floor of its full-year 2026 capex guidance from $130 billion to $135 billion (keeping the $145 billion top), a clear signal that the AI buildout is not slowing down. The company also raised full-year expense guidance to $175–179 billion to absorb the litigation costs. Meta ended the quarter with $82.4 billion in cash against $83.7 billion in debt — enough to keep spending, for now.
Muse becomes the earnings narrative
The most-watched part of the call was AI. Mark Zuckerberg opened with a line that will define this quarter:
"We hit major milestones this quarter toward delivering personal superintelligence for everyone — including launching Muse and Muse Wonder, our first frontier model. Our investments are paying off, the core business is strong, and we're seeing velocity across the board."
This ties directly to Muse, Meta's personal AI assistant, which launched in early September. On the call, Zuckerberg gave the first official usage numbers: more than 10 million people in the US and Canada set up Muse personal agents in about a month and a half, with Meta "number one in the App Store almost every day." Sensor Tower data reported this week puts it at 10.2 million US downloads in 40 days — a pace only 11 other apps have ever matched, most of them mobile games — with notably better early retention than ChatGPT, Claude, or Grok.
The commerce angle stood out: shopping is Muse's most common use case, over a quarter of users have made purchases, and the average purchase exceeded $200. Management's monetization roadmap is a free tier funded by business commerce fees, paid subscriptions for advanced models, and enterprise offerings. Zuckerberg acknowledged Muse is "very compute-intensive to run" today but expects cost efficiency to improve more than 10x over the next year. That efficiency claim is the linchpin of the whole AI-spending thesis — watch whether it materializes.
The rest of the business: still strong, still changing
The core ad engine remains the story investors can trust. Ad impressions grew 10% while the average ad price rose 14%, a combination that means both more people seeing ads and advertisers paying more per view. WhatsApp paid messaging and subscriptions drove "other revenue" up 59% to $1.1 billion — a small but fast-growing line.
On engagement, Zuckerberg said Facebook and Instagram hit all-time highs, Threads has now surpassed X in monthly active users, and WhatsApp reached all-time highs in messaging activity. The 3.64 billion daily users across the family of apps is Meta's moat, and it is still widening.
Reality Labs posted $443 million in revenue, down 6% — lower Quest headset sales were partly offset by strong growth in AI glasses. Its $4.1 billion operating loss came in narrower than the roughly $5.1 billion analysts expected, and the company expects the segment's full-year loss to narrow to $18–19 billion, with further improvement toward roughly $15 billion in 2027.
New Class C shares: funding the future without losing control
In a major governance move, Meta's board approved a one-time dividend of one non-voting Class C share for every Class A or B share held, taking effect in January 2027. It functions like a 2-for-1 stock split, but voting control stays exactly where it is. The purpose is practical: a new currency for raising capital, compensating employees, and funding acquisitions — all without diluting control — while Meta navigates the AI race's enormous capital demands.
Guidance and market reaction
For Q4, Meta guided revenue of $73–76 billion, with the midpoint above consensus — and the company expects a roughly 1% foreign-exchange headwind. The stock closed the regular session up $3.27 at $697.52 and rose roughly 2–4% in after-hours trading, suggesting investors bought the quarter despite the profit hit.
The verdict from Wall Street was mixed but tilted positive: Pivotal Research raised its price target from $930 to $960, citing stronger revenue and capex execution plus Muse's usage data, while BMO kept a Hold rating on concerns about the visibility of returns on 2026–2027 capital spending.
What to watch next
- 2027 capex guidance — the raised $135–145B floor for 2026 already implies the answer, but confirmation moves the stock.
- Muse monetization — the first revenue or subscriber figures, whenever Meta discloses them.
- The cost-efficiency claim — if Muse doesn't get 10x cheaper to run, the AI math gets harder.
- Q4 delivery — $73–76B guidance must land, with the teen-litigation overhang now mostly in the rearview mirror.
Yesterday we asked whether Meta's AI spending was paying for itself. Tonight's answer: the ad business is still funding the future, the legal bill was a detour, and Muse — Meta's "first frontier model" — is now officially the company's growth story. The spending, though, is only going one direction.
Sources: Meta Q3 2026 coverage verified against the earnings call transcript (Motley Fool), and reporting from Shacknews, The Markets Daily, Zacks, AlphaSense, Particle, XTB, ChainCatcher, Finnhub market data, KuCoin News, Biggo Finance, and TechCrunch/Sensor Tower. Links: earnings call transcript · Shacknews report · Markets Daily call highlights · Zacks key metrics · Particle overview · XTB analysis · Biggo Muse coverage · TechCrunch on Muse downloads · KuCoin flash report
FAQs
Why did Meta's profit fall if revenue beat expectations?
A one-time $9.89 billion legal charge from the August 2026 settlement with 52 state attorneys general over teen mental-health litigation was booked in Q3. It dragged GAAP operating income down 37% to $13 billion. Excluding that charge, operating income would have grown about 11% year over year, according to CFO Susan Li.
What is the difference between Meta's GAAP EPS of $3.83 and adjusted EPS of $7.40?
GAAP EPS ($3.83) includes the $9.89 billion legal charge, which knocked roughly $3.57 off per-share earnings. Adjusted EPS ($7.40) removes that one-time item, giving a cleaner read on the underlying ad business. Analysts generally compare the adjusted number — $7.40 versus $7.25 a year ago — when judging the quarter.
What are the new non-voting Meta Class C shares?
Meta's board approved a one-time stock dividend: one non-voting Class C share for each existing Class A or B share held, taking effect in January 2027. It works like a 2-for-1 split but changes no voting control. Meta says the new share class gives it flexibility to raise capital, pay employees, and fund acquisitions while it pours money into AI infrastructure.
Is Meta's Muse AI making money yet?
Not that Meta has disclosed — no revenue figures were given on the earnings call. Management outlined a monetization roadmap instead: a free consumer tier funded by commerce fees paid by businesses, paid subscriptions for advanced models and extra tokens, and enterprise offerings including an API, Muse Code, and the Meta Enterprise platform. Early usage signals are strong: over a quarter of Muse users have made purchases, averaging over $200 per purchase, according to Pivotal Research.
Why did Meta's free cash flow turn negative?
Capital expenditure hit $34.9 billion in Q3 alone (up 80% year over year), outrunning the $32.8 billion in operating cash flow. The result was negative $2.1 billion in free cash flow — the first negative quarter since Meta went public, though still far better than the roughly negative $9.4 billion analysts had feared.
Written by
AI Trends Daily Team
The AI Trends Daily team researches AI and technology daily, then explains it in plain English — no jargon, no hype, just what matters.
Related Articles
TrendingMeta Q3 2026 Earnings Preview: The $42B AI Test
Meta reports Q3 2026 earnings tonight after the close. Three numbers — revenue, capex, and a Muse monetization signal — will show if its AI bet is paying off.
October 21, 2026
TrendingAI Agents Hacked 100+ Websites: Microsoft Fights Back (2026)
One hacker used AI agents to breach 100+ company websites for $5,000. Microsoft now treats agents as insider threats. What happened, what it means for you.
October 20, 2026
TrendingAolani's 500MW Malaysia AI Data Centre Deal Explained
Aolani signed a 500MW AI data centre deal in Malaysia — among the region's biggest AI infrastructure bets. Why it matters, and what it means for AI costs.
October 19, 2026