Meta lost roughly $200 billion in market value in a single day over AI spending fears, then clawed most of the sentiment back four days later. Wall Street still hasn't decided what an $80 billion capex bet is actually worth.
Meta Platforms reported second-quarter earnings on July 29, 2026, and the market's first reaction was brutal. Shares fell as much as 10% the next trading day, July 30, wiping out tens of billions in value as investors zeroed in on one number: Meta raised the low end of its full-year capital expenditure guidance to a range of $130 billion to $145 billion. Free cash flow, meanwhile, collapsed. It dropped roughly 91% year over year to just $784 million, according to CNBC, as $31.1 billion in quarterly capex consumed nearly all of Meta's operating cash flow.
Revenue told a different story. It grew 28% year over year to $60.8 billion. But earnings per share came in at $6.18, about a dollar short of the roughly $7.17 Wall Street expected, and that miss, layered on top of the capex number, was enough to spook the room.
The same week, Microsoft posted its own AI-heavy earnings and jumped as much as 15%, according to CNBC. The split wasn't really about who spent more. It was about who could point to paying customers: Microsoft's Azure cloud business showed clear external demand for its AI infrastructure, backed by a growing order backlog. Forbes framed the divergence bluntly: Wall Street is rewarding proven AI demand and punishing AI spending that looks internal and speculative. Meta's AI investment, by contrast, is mostly funding its own ad-ranking systems and Llama models, not sold to outside customers the way Azure or AWS capacity is.
Meta's Reality Labs division didn't help the mood. It lost $4.62 billion in the quarter, pushing cumulative losses toward $88 billion since the unit's creation, and the loss actually widened from the $4.03 billion it posted in the first quarter. That's not a business turning a corner. It's a bet still deep underwater, eight years into Mark Zuckerberg's metaverse push.
Then, four days later, the reversal #
On August 3, the stock reversed hard. Shares jumped as much as 6.7%, closing near $594 and pushing Meta's market cap back above $1.5 trillion, according to TradingKey and The Motley Fool. The rally reflected a shift in how analysts were reading Meta's balance sheet: the core advertising business throws off enough cash to fund the AI buildout on its own, without the company needing to lean on debt or spook shareholders further. WhatsApp, still largely unmonetized for ads, stands out as one of the clearest untapped revenue levers inside that core business - giving investors a durable growth story that didn't exist in the panic of July 30.
Nothing about Meta's underlying business actually changed in those four days. The capex guidance didn't shrink. The free cash flow didn't recover. What changed was the story analysts told about where that money eventually leads. Ad impressions were up 14% and price per ad up 12% in the same quarter, proof the core business isn't stalling even while it funds an AI buildout that hasn't found outside customers yet.
What it means for anyone else pitching AI #
Here's the thing worth sitting with: this isn't really a Meta story. It's a preview of how every AI infrastructure bet gets priced from here. Microsoft sells AI compute to other companies, and got credit for it. Meta doesn't, not yet, and got punished. For now, its AI spend mostly makes Meta's own products better, nothing more. That distinction is the whole ballgame if you're a founder pitching AI infrastructure to investors. Spending on AI is no longer impressive by itself. Spending on AI that someone else is already paying to use - that's what moves a stock 15% in the right direction instead of 10% in the wrong one.
Meta's next earnings report, due in late October, will be the real test of whether Monday's rally was conviction or just relief.
Also read: INC Ransomware Exploits Maximum Severity SonicWall SMA1000 Flaws • JPMorgan and Accenture Back a New Group to Standardize AI Token Costs • Sonera Turns 20 Years of Compliance Data Into the Foundation for AI-Era Outbound Communications