Meta Platforms closed Friday at $709.33, down from its September 24 peak of $779.82, a pullback that has left the stock sitting roughly 9% below its 52-week high heading into Wednesday’s after-close report. The earnings figure the market sees will look alarming. The figure worth focusing on is entirely different.

Wall Street expects Q3 EPS of approximately $6.76, compared to $1.05 a year ago. That year-over-year jump is almost entirely a function of distorted comparables: Q3 2025 included a $15.93 billion one-time, non-cash tax charge. Strip that away and the comparison is far less dramatic. Bank of America, which reiterated a Buy with an $810 target ahead of the report, estimates underlying EPS will actually exceed consensus once a roughly $10 billion legal accrual tied to a settlement with state attorneys general is excluded.

The Number That Actually Matters

Advertising revenue. Meta’s Q2 2026 ad revenue grew 27% to $59.363 billion. Bank of America projects Q3 ad revenue of $62.2 billion, representing 24% year-over-year growth, and total revenue of $63.8 billion, above the $63.3 billion street consensus. The company’s own guidance called for Q3 revenue of $61 to $64 billion.

The ad business is growing because AI is making it work better. Meta told investors on its Q2 2026 earnings call that AI-driven upgrades to its ad-ranking system produced an 8.3% increase in ad clicks and a 15.7% uplift in conversions on Facebook. Advertisers are paying more per ad because the return justifies it. That is the structural argument for the stock.

The Cost Side Is the Complication

Meta expects 2026 capital expenditures of $130 to $145 billion, a range narrowed upward from $125 to $145 billion. Bank of America has already raised its 2027 capex forecast to $205 billion, above Wall Street’s $195 billion estimate, citing infrastructure requirements for the Muse AI products. Free cash flow in Q2 fell to $784 million from $8.55 billion a year earlier, a 91% collapse that reflects how much cash the AI buildout is consuming right now.

Muse, the AI agent launched on September 8, drove a roughly 27% September stock rally. Citigroup has estimated that Muse could generate more than $27 billion in annual revenue by 2030. But monetization at scale is unproven, and the market has already paid for the promise once. Wells Fargo raised its target to $1,000 on October 6. That is optimism, not evidence.

The tension between an AI product capturing investor imagination and an ad engine under pressure is not unique to Meta. Baidu’s experience shows how quickly markets discount an ad business when AI revenue timelines remain unproven.

What Could Disappoint

A legal charge near or above $10 billion will compress reported EPS sharply. If management widens 2027 capex guidance meaningfully without a corresponding revenue acceleration signal, the stock could revisit its early-October lows near $718. And the Q3 2025 comparison for operating margins is genuinely difficult: operating margin ran at 40% a year ago and fell to 31% in Q2 2026 as costs surged 55%.

Analyst consensus remains Strong Buy based on 45 ratings, with an average price target near $815. At $709.33, that implies roughly 15% upside before the report even lands.

The Bottom Line

Meta enters Wednesday with a business growing advertising revenue at roughly 24%, an AI product that captured investor imagination in September, and a legal charge that will make the headline EPS nearly unreadable. The stock is about 9% off its recent high. The question for investors is whether the ad growth rate, and what management says about 2027 spending plans, justifies paying roughly 27x trailing earnings for a company burning cash to build infrastructure it may not fully monetize for years.