Meta beats estimates but AI spending rattles

Meta’s quarter was strong. The market still flinched. That tells you what investors care about now — not whether Meta can make money today, but how expensive its AI push is getting and how clearly Mark Zuckerberg can explain the payoff. (investor.atmeta.com) ### What happened? Meta reported first-quarter 2026 revenue of $56.31 billion, up 33% from a year earlier, with diluted EPS of $10.44. Strip out a big tax benefit, and adjusted EPS was $7.31 — still ahead of expectations. The company also guided second-quarter revenue to $58 billion-$61 billion. On the surface, that is a very good quarter. (prnewswire.com) ### So why did the stock drop? Because Meta also raised its full-year capital expenditure forecast to $125 billion-$145 billion, up from $115 billion-$135 billion. That is the number that changed the mood. Investors can live with huge AI spending when the story feels crisp. They get jumpier when (prnewswire.com)(investor.atmeta.com) ### What is Meta spending all that money on? Mostly the plumbing for AI — chips, servers, networking gear, and data centers. Meta said the higher capex outlook reflects pricier components and extra data-center costs to support future capacity. Basically, the company is buying the compute it thinks it will need before the revenue from new AI products fully shows up. (investor.atmeta.com) ### Isn’t AI already helping Meta? Yes — but mostly through the old business. Meta’s ad machine keeps getting better as AI improves targeting, recommendations, and engagement. Ad impressions rose 19% year over year, and average price per ad rose 12%. That helps explain the 33% revenue growth. The catch is that this is still mostly “AI makes ads better,” not “AI is already a giant standalone business.” (prnewswire.com) ### Where does Muse Spark fit in? Muse Spark is the new piece investors are trying to price. Meta introduced it in early April as the first model from Meta Superintelligence Labs, and it looks like a strategic break from the old Llama playbook. Llama was released freely to the open-source communi(prnewswire.com)d Google are doing. (cnbc.com) ### Why does that strategy shift matter? Because open models can win developer goodwill, but they do not automatically produce clean revenue. A more closed, paid-access model could. That is why Wall Street wants Zuckerberg to explain not just model quality, but product strategy — who pays, for what, and how soon. The market is basically asking Meta to prove this is a business plan, not just an arms race. (cnbc.com) ### Was there anything else messy in the quarter? Yes — user growth. Family daily active people came in at 3.56 billion, below Wall Street expectations, and down slightly from the prior quarter. Meta blamed internet disruptions in Iran and restricted WhatsApp access in Russia. That did not dr(cnbc.com)at. (investor.atmeta.com) ### What does this mean beyond Meta? Big Tech has moved into a new phase of the AI cycle. Last year, investors rewarded ambition. Now they want conversion — spending turned into products, products turned into usage, usage turned into revenue. Meta’s quarter shows the bar has changed. Beating estimates is nice. Explaining the return on $145 billion of infrastructure is the real job now. (cnbc.com)