Meta Q3 2026 Earnings Preview: The $42B AI Test

Meta Platforms reports its Q3 2026 results tonight after the US market closes, and this quarter is about more than the usual ad-revenue beat or miss. It is the first earnings report since Meta launched Muse, its autonomous AI assistant, and Wall Street wants an answer to one uncomfortable question: is the company's record AI spending actually paying for itself?

A quick honesty note up front: the results are not out yet. They land after the market close tonight — roughly 1:00am PKT on Thursday, October 22 — and the earnings call follows at 4:30pm ET (about 1:30am PKT). So this is a preview: what to watch, why it matters, and which numbers will actually move the stock. We'll update this post with the actuals once they land.

When the numbers come out — and one timing detail worth noting

The schedule is settled: results after the close on Wednesday, October 21, with the conference call at 4:30pm ET, streamed on Meta's investor relations site. The first full trading session to price the numbers opens Thursday morning.

One thing earnings watchers noticed: Meta moved this report up by a full week — the earliest third-quarter report in a decade, according to one earnings tracker. Some analysts read that as a potential confidence signal, a break from the usual "Mega-Cap Tech Week" clustering with peers like Amazon and Alphabet. That is a guess, not evidence — but it is the kind of small tell investors are parsing because the stakes are high.

Number one: revenue against a $63.4 billion consensus

The number everyone will quote first: analysts expect roughly $63.3–63.4 billion in revenue — FactSet data points to about $63.34 billion, up roughly 23.6% from $51.24 billion a year earlier, per TradingKey's preview. That sits near the top of Meta's own $61–64 billion guidance.

Some analysts are more bullish than the crowd. BofA's Justin Post reiterated a Buy rating with an $810 price target and forecasts $63.8 billion in revenue, arguing AI-driven ad growth should help Meta beat the consensus. Bernstein's Mark Shmulik is at $64.06 billion — about 1% above consensus — with an $800 target, also Buy-rated.

Why revenue matters more than usual this quarter: advertising — Facebook, Instagram, WhatsApp — is still the engine that funds everything else. A Bank of America survey found ad demand held solid in Q3, and in Q2 ad impressions grew 14% while the average price per ad rose 12%, which Meta attributed to its AI-powered ad systems. The catch: growth is decelerating. Revenue grew 33% in Q1, 28% in Q2, and consensus implies roughly 24% in Q3. Landing at the top of guidance would validate the AI-advertising story; landing at the bottom would revive the "AI spending is masking a slowdown" worry.

Number two: the roughly $42 billion capex bill

This is the number the market will fight over. Analysts tracked by Visible Alpha estimate Meta's Q3 capital expenditure at about $41.7 billion — a huge step up from the $31.1 billion spent in Q2. Full-year 2026 capex guidance is $130–145 billion (including finance lease payments), raised at the lower end from $125 billion back in July. For perspective, that range more than doubles Meta's capital-spending pace of 2024.

The cash math is what worries people. In Q2, $31.1 billion in capex left free cash flow at just $784 million — versus $8.55 billion a year earlier. Estimates cited by trendintech suggest Q3 free cash flow could turn outright negative, around an $8.5 billion burn. That estimate may be off, but the direction is real: the AI buildout is now eating directly into the cash Meta generates.

As OANDA's preview puts it, the market is no longer asking whether demand for AI infrastructure is strong — it wants evidence the returns justify the financing and operating costs. S&P Global has flagged concerns about hyperscaler AI financing commitments broadly. Expect the single most important line on tonight's call to be any signal on 2027 capex guidance — whether the spending keeps climbing or plateaus.

Meta is clearly trying to build confidence ahead of that conversation. Days before the report, the company published a technical deep dive on Prometheus, its gigawatt-scale AI supercluster in New Albany, Ohio, built with Nvidia — sharing real reliability data, including a reported 99% reduction in daily interruptions per 8,000 GPUs since mid-2025, per Benzinga's coverage. Tech analyst Patrick Moorhead called it "co-engineering to the max." It is marketing, yes — but marketing with actual engineering numbers, which is notable heading into earnings.

Number three: any Muse monetization signal

Muse, Meta's personal AI agent, launched September 8, and its debut numbers were striking: over 5 million downloads in 22 days, per Sensor Tower data cited by CNBC — faster than ChatGPT's 56 days and Grok's 103 days to the same mark, according to TalkMarkets. It hit #1 on the App Store on September 18 and Google Play the next day, and Meta's stock jumped 27% in September, its best month since 2022.

But downloads are not dollars, and that is what tonight's call must address. The known pricing structure: a free tier, Power at $20/month, and Maximum at $100/month — the paid tiers buy more weekly usage, not extra features, according to Meta's help-page details as reported by the press. The tiers have not been officially confirmed in Reuters' earnings preview, so treat them as reported, not official.

The honest counterpoint, from the Motley Fool: subscriptions "won't come anywhere close" to offsetting up to $145 billion a year in AI data-center spending — especially since Muse is a consumer product, while rivals offset infrastructure costs with enterprise cloud deals. Bernstein's Shmulik sees Muse as the core of Meta's AI strategy but warns against assuming early interest converts quickly into a large business, pointing to past over-hyped bets. As Wedbush's Dan Ives framed it in Benzinga's coverage, the focus is "whether Meta can begin to monetize Muse on top of the ad engine."

The $20/month figure is also becoming the industry's default price for AI subscriptions — for comparison, our breakdown of ElevenLabs' new voice-agent plans launched this same week shows the whole industry converging on that tier.

One more line item: Reality Labs is expected to post an operating loss of around $5.2 billion for the quarter, per TradingKey. The drag continues, but it is the known drag — capex and Muse are the new questions.

Why the stock feels fragile going in

Context for tonight's reaction: Meta's stock is down over 5% in the past five trading days, per TipRanks, and it fell about 9% the day after the Q2 miss. Options markets are pricing roughly a ±7.1% earnings move, below the stock's 10-year average post-earnings move of ±10.4%, according to OANDA.

And a warning about the headline EPS number: it will look strange. A roughly $10 billion one-time legal settlement with state attorneys general, not in the July guidance, will weigh on GAAP EPS — BofA's Post forecasts reported EPS of just $3.79 because of it. Adjusted consensus sits around $6.50–$6.77. Do not read the headline EPS as a verdict on the business; revenue and cash flow are the cleaner reads.

The checklist for tonight

If you are watching the call or the headlines, here is what actually matters:

1. Revenue at or above ~$63.4B — near the top of guidance validates the AI-ads engine.
2. Q3 capex vs the ~$41.7B estimate — and whether full-year $130–145B guidance narrows or widens.
3. Free cash flow — positive or negative, and how management frames the 2027 trajectory.
4. Any Muse revenue or subscriber signal — the first hard monetization data point.
5. 2027 capex guidance — the number the stock will trade on tomorrow.

The results land around 1:00am PKT tonight; Tesla reports the same night. Check back here — we'll update this post with the actual numbers and what changed in the story.

---

Sources: Meta Q3 earnings previews and timing details via OANDA, TradingKey, TipRanks, TradingView/Benzinga, mondeumcapital, Tickeron, primexbt, trendintech, shattered.io, WSNext, IG, Seeking Alpha, Motley Fool, TalkMarkets, UseCarly, and goSearch.ai. All figures are analyst estimates or company guidance as of October 21, 2026, unless stated — Meta's actual results were not yet published at the time of writing.