Analysis
Meta Platforms shares were down again in Thursday premarket trading, extending a slide that has now become the longest daily losing streak in the company's history as a publicly traded stock. The stock fell for nine consecutive sessions heading into Wednesday's earnings report, sliding from a July intraday high of $681.90 to $593.41 and wiping out roughly $223 billion in market capitalization before a single second-quarter number had been reported. Rather than snapping the streak, Wednesday's after-hours print and Thursday's follow-through pushed the losses further, with shares falling as much as 9.64% after hours to $529.15 and then sliding another 7-9% in premarket trading Thursday morning.
The irony is that the underlying numbers weren't uniformly bad. Second-quarter revenue came in at $60.80 billion, up 28% year-over-year and ahead of the roughly $59.5-60.2 billion analysts had modeled, powered by a 14% jump in ad impressions and a 12% rise in average price per ad. The core advertising engine is still compounding four years into the AI infrastructure buildout -- that part of the story hasn't changed.
What spooked the market was everything beneath the top line. Adjusted earnings per share of $6.18 missed the roughly $7.13-7.22 consensus by a wide margin, and net income fell 14% year-over-year to $15.8 billion. Free cash flow -- the number growth investors watch most closely when a company is in heavy capex mode -- collapsed 91% to just $784 million, down from $8.55 billion a year earlier. Meta also raised the low end of its full-year capex range to $130-145 billion, from $125 billion previously, and guided full-year expenses to $165-169 billion. Third-quarter revenue guidance of $61-64 billion landed with a midpoint some desks read as soft relative to where consensus had drifted.
“Adjusted earnings per share of $6.18 missed the roughly $7.13-7.22 consensus by a wide margin, and net income fell 14% year-over-year to $15.8 billion.”
The timing made the divergence impossible to ignore. Hours after Meta's print, Microsoft reported that Azure had crossed $100 billion in annual revenue for the first time, grew 43% in the quarter, and guided FY2027 capex to $255-260 billion -- and its stock jumped roughly 7-9% on the news. Two hyperscalers, both raising AI spending guidance in the same 24 hours, produced opposite stock reactions: investors rewarded Microsoft's growth-and-monetization story and punished Meta's margin compression and cash-flow collapse. That's a market drawing a sharper line than "AI capex is bad" -- it's pricing execution and payback, not spending levels alone.
The investor unease predates this week's print. Meta's nine-day slide has been driven largely by capital-allocation concerns: a sprawling set of bets across Llama, Reality Labs, Quest, Orion smart glasses, and now cloud infrastructure hints from Zuckerberg, without a clean line from any one of them to near-term profit. Mark Zuckerberg's own stake lost roughly $15 billion in value as the stock slid, a visible marker of how directly the founder is now underwriting the AI infrastructure bet with his own balance sheet.
For founders and operators watching the hyperscaler earnings as a leading indicator, the read-through matters beyond Meta's cap table. If public markets are now willing to punish a 28%-growth quarter over free-cash-flow compression, the bar for AI infrastructure capex is rising even as the dollar amounts keep climbing -- a signal worth carrying into any GPU-heavy fundraising conversation this fall. The bear case is straightforward: if Meta's Q3 print in October doesn't show free cash flow stabilizing, the losing streak narrative hardens from "rough week" into "structural re-rating." The next data point lands tonight, when Amazon reports after the close -- the last of the "Big Four" this season, and the one that will determine whether Wall Street treats capex anxiety as a Meta problem or an industry one.
