Meta's $60.8B Revenue Fails to Halt Stock Drop as AI Capex Crushes Free Cash Flow

Woofun AI reports that Meta’s Q2 financial results exposed a critical paradox: record-breaking revenue failed to sustain stock price stability, signaling market rejection of the company’s aggressive AI ledger strategy despite hitting historical highs in earnings.

The financial snapshot presented a stark divergence between top-line performance and market reaction. Meta generated $60.8 billion in revenue, marking a 28% year-on-year growth that surpassed Wall Street consensus. Advertising revenue alone reached $59.4 billion, driven by a 27% growth rate, with ad impressions rising 14% and average prices increasing 12%. Despite these metrics showing no negative news on the surface, the stock fell 7% in after-hours trading. In contrast, Microsoft, which reported earnings on the same day, saw its shares rise over 8% after its cloud business exceeded expectations.

The core of the market’s anxiety lies in the evaporating free cash flow. Operating cash flow for the quarter stood at $31.86 billion, but after deducting $31.08 billion in capital expenditures, free cash flow was reduced to a mere $784 million. This represents a 91% evaporation compared to the $8.55 billion recorded in the same period last year. If this burn rate persists, free cash flow is projected to turn negative in the next quarter. Google has already entered negative cash flow territory in its recent earnings, marking the first time in its history, highlighting a sector-wide pressure point.

Expense structure analysis reveals that one-time items do not account for the entire cost surge. CFO Susan Li explained that total expenses for Q2 were $42 billion, a 55% increase year-on-year. This included $2.4 billion in legal provisions and $1.18 billion in severance pay. Excluding these specific items, operating profit would have increased by approximately 9% year-on-year.

However, the deeper driver is the R&D expenditure, which surged to $21.66 billion, a 67% increase. The proportion of R&D to revenue jumped from 27% in the prior year to 36%, indicating that this spending is structural, tied to monthly AI payments rather than transient legal or staffing costs.

Capital expenditures have locked in a massive annual investment baseline. Capex soared from $17.01 billion in the same period last year to $31.08 billion, an 83% increase. The company adjusted its full-year capex guidance from a range of $125 to $145 billion to $130 to $145 billion. While the upper limit remained unchanged, the lower limit was raised from $125 to $130 billion. This adjustment confirms that at least $130 billion in annual AI investment has been secured as the bottom line, removing flexibility to scale back spending in the near term.

Woofun AI data shows that AI implementation is delivering tangible engagement metrics, yet monetization lags. Zuckerberg highlighted the Advantage+ AI advertising tool, which has an annualized revenue run rate exceeding $75 billion. Nine million advertisers now use at least one Meta AI tool, and LLMs are integrated into recommendation systems. Instagram users are experiencing double-digit growth in time spent, validating the utility of these tools.

However, the growth rate of the advertising business is not keeping pace with the growth rate of capital expenditures, creating a structural imbalance in the cost-revenue equation.

The monetization gap is widening as revenue growth slows while costs accelerate. Revenue growth decelerated from 33% in Q1 to 28% in Q2. The Q3 guidance median is $62.5 billion, corresponding to about 22% year-on-year growth, confirming a clear deceleration trend.

Meanwhile, capex growth stands at 83%, a three-to-one gap against revenue growth. Operating profit margin dropped from 43% in the same period last year to 31%. Unlike Microsoft Azure and Google Cloud, which monetize computational redundancy directly, Meta lacks a mature cloud business to offset inference costs, forcing it to bear the full burden of infrastructure spending.

Strategic divergence between Microsoft and Meta highlights different risk profiles. Microsoft Azure benefits from a short, clear commercial cycle where corporate demand for AI computing power translates directly into cloud revenue. Meta operates a longer, closed loop: AI improves ad recommendations, advertisers see better ROI, and ad revenue grows. While every link exists, the cycle from investment to return is longer and more uncertain. Microsoft is effectively selling shovels, collecting cash quickly. Meta is using shovels to dig its own mine. Ad prices increased 12% and impressions rose 14%, with Instagram surpassing 2 billion daily active users, but the cost of mining is rising at three times the rate of price increases.

User scale serves as both a moat and an inference burden. Meta’s family of apps has 3.6 billion daily active users, a 3% year-on-year increase. Threads has over 500 million monthly active users, and WhatsApp peaked at 30 million messages per second during the World Cup final. This base provides the largest AI distribution and testing scenario globally, offering massive training data and feedback loops for recommendation algorithms and content generation.

However, the demand for AI inference from 3.6 billion users burns computing power at an equivalent scale. Every personalized recommendation and real-time bidding optimization increases inference costs, which are rising faster than advertising efficiency gains.

The market’s 7% sell-off reflects a shake-up of the core assumption that AI investments would be covered by revenue growth within 12 to 18 months. The Q2 data shows the opposite: revenue growth slowing to 22% in Q3 projections, capex accelerating at 83%, and profit margins compressing to 31%. Susan Li indicated that 2026-2027 will be the "full-scale expansion" phase, with only room for adjustment after 2028. This implies at least another 18 months of high-intensity investment. Q3 is a critical window; with free cash flow at $784 million and capex rising, turning negative is imminent. Zuckerberg’s pivot to providing APIs and intelligent agents offers a potential escape route, but without a timeline for cost convergence, Meta remains a heavy-asset AI infrastructure operator facing a valuation premium fade as patience runs out.