Meta’s advertising growth beats expectations as new AI businesses lift valuation and add cash pressure

What was strong about the third quarter, and why was that not reflected in profits?

Meta’s third-quarter revenue was 64.6 billion dollars, up 26% year over year, above the market forecast of 63.3 billion dollars and the upper end of company guidance of 64 billion dollars. Excluding currency effects, advertising revenue grew 26% year over year, matching the prior quarter’s pace despite a 3 percentage-point tougher year-over-year comparison.
GAAP earnings per share were 3.83 dollars, below the market forecast of 5.52 dollars, primarily due to a one-time legal settlement charge of 9.9 billion dollars. Bank of America said operating profit and earnings per share exceeded expectations excluding that charge, helped by research and development expenses of 13.5 billion dollars, below the market forecast of 16.5 billion dollars.
Fourth-quarter revenue guidance was 73 billion–76 billion dollars, with the midpoint above the market forecast of 73.6 billion dollars. The upper end implies constant-currency revenue growth of 28%, 2 percentage points faster than in the third quarter, although the year-over-year comparison also became 2 percentage points easier.

The link is recommendation and conversion efficiency: more advanced models improve content recommendations and ad matching, raising user engagement and advertisers’ returns, thereby attracting more budget. Disclosed results include a nearly 2% conversion improvement in some Facebook Feed scenarios from generative recommendation models, and increases of 1%–1.7% in offsite conversions and 9.6% in app-ad conversions from GEM.
Bank of America believes improvements will continue as large language models become more deeply integrated into advertising systems and additional computing capacity supports broader deployment. Its higher revenue forecast therefore remains based mainly on core advertising growth, rather than factoring in the full potential of new AI businesses in advance.

Muse has been available in the US and Canada for about six weeks, with more than 10 million setups; the company’s long-term target is over 1 billion users. International expansion, stronger models and Meta’s existing ecosystem of 3.6 billion users are the main paths to wider adoption.
Shopping is the first commercial link: more than a quarter of users use the feature, with an average purchase above 200 dollars. Management has identified transaction fees, consumer and enterprise subscriptions, and advertising personalization as monetization paths; Bank of America expects subscriptions and e-commerce commissions could begin contributing to revenue in the second half of 2027.
On the enterprise side, the focus is on model API licensing and services such as Muse Code. Whether profitability improves with scale also depends on costs: the company expects Muse computing cost efficiency to improve more than 10-fold over the next year, through utilization, model efficiency and inference optimization.

Bank of America raised its 2027 revenue forecast by 1.3% to about 319 billion dollars, but cut its earnings-per-share estimate from 34.24 dollars to 34 dollars. Muse expansion, enterprise AI development, and a higher tax rate and share count offset part of the revenue increase.
Cash spending is heavier: Bank of America raised its 2027 capital expenditure forecast by about 5 billion dollars to 210 billion dollars and expects negative free cash flow of 33 billion dollars. Actual third-quarter free cash flow was already negative 5.3 billion dollars, showing expansion pressure in cash flow.
Meta announced new non-voting Class C shares, but disclosed no specific financing plan on the call. Bank of America identifies potential dilution from equity financing as a risk; meanwhile, about 98% of revenue still comes from advertising, and higher fixed costs would amplify pressure on earnings if advertising demand slows.

Bank of America maintained its buy rating and raised its price target from 810 dollars to 845 dollars, while raising its 2027 forward price-to-earnings multiple from 24 times to 25 times. Its reasoning is that stronger advertising growth and a clearer AI commercialization path merit a higher valuation; the price-target increase does not come from an improved earnings forecast.
The report used an after-hours share price of about 713 dollars, equivalent to a 2027 forward GAAP price-to-earnings multiple of about 21 times, in line with the nine-year historical average. Bank of America expects monetization of new AI services to lift the valuation, but accelerating investment, financing and AI platforms competing for advertising budgets will all affect that path.