Analysis
Meta's second-quarter results landed as a genuine split decision. Revenue came in at $60.80 billion, up 28% year-over-year and ahead of the $59.50 billion Wall Street expected, powered by a 14% jump in ad impressions and a 12% increase in average price per ad -- proof the core advertising engine is still compounding even four years into the AI infrastructure buildout. But adjusted earnings per share of $6.18 missed the $7.13 consensus by a wide margin, and net income fell 14% year-over-year to $15.8 billion, as elevated AI infrastructure and operating costs ate directly into the bottom line.
The capex signal is the one investors are parsing most closely. Meta raised the low end of its full-year capital expenditure range to $130-145 billion, from a prior $125-145 billion, and guided full-year expenses to $165-169 billion. Operating cash flow of $31.86 billion for the quarter looked healthy, but free cash flow of just $784 million shows how thin the margin has gotten once AI infrastructure spend is netted out -- a dramatic compression from the cash-generative Meta of two years ago.
“Meta raised the low end of its full-year capital expenditure range to $130-145 billion, from a prior $125-145 billion, and guided full-year expenses to $165-169 billion.”
This result lands in the middle of an earnings week that's becoming a referendum on how each hyperscaler is managing the same tradeoff. Alphabet's July 22 print saw its stock drop 7% after a capex raise came with negative free cash flow for the first time since its 2004 IPO. Meta's mixed beat-miss-raise combination fell somewhere in between -- shares moved lower on the print, per Benzinga, reflecting the EPS miss more than the revenue beat. Hours later, Microsoft's own capex raise to $255-260 billion for FY2027 sent MSFT up 7% after hours, a sharper positive reaction built on an Azure business growing 43% and crossing $100 billion in annual revenue.
For founders and operators, the read is that ad-driven revenue is holding up even as AI infrastructure investment intensifies -- Meta's core business is not showing demand weakness, it's showing cost pressure. That's a different risk profile than the chip-sector selloff unfolding the same week, which is about demand uncertainty rather than margin compression on proven demand. For LPs modeling exposure to consumer-facing AI plays, Meta's Q3 guide of $61-64 billion revenue suggests growth continues; the open question is whether margin recovers once the current infrastructure buildout phase matures or whether this is the new steady state.
What to watch: whether Meta's Q3 print in October shows margin stabilizing as Llama and Meta AI monetization ramps, and whether the market continues rewarding hyperscalers more for cloud/infrastructure revenue growth (Microsoft's model) than for maintaining a consumer ad business while absorbing AI costs internally (Meta's model).
