The Q1 results were flawless. Revenue grew 33% and earnings per share hit $7.31, both beating expectations. And yet Meta's stock dropped 8% in after-hours trading. The answer to that contradiction was buried in how Zuckerberg responded to one analyst's question. Asked whether there were any signs of return on investment (ROI) for all the AI spending, he replied, "That is a very technical question."
What Actually Happened: The Fear Created by a Second Upward Revision
On April 29, 2026, Meta reported its Q1 results and raised its full-year AI capital expenditure forecast to $125 billion to $145 billion. It was the second consecutive upward revision, coming just three months after the company guided the same figure to $115 billion to $135 billion back in January. Actual Q1 capital spending came in at $19.8 billion, below the market estimate of $27.5 billion, but the upper end of the annual guidance was so high that the market panicked. Revenue reached $56.3 billion, up 33% year over year, marking the strongest growth quarter since 2021, while net income came to $26.8 billion.
Why This Matters More Than People Think: The AI Gamble With No Cloud
The real reason investors are uneasy is not the numbers. It is the context. Amazon, Microsoft, and Alphabet are pouring hundreds of billions of dollars into AI infrastructure, but they bill that investment directly back to customers through revenue engines like AWS, Azure, and Google Cloud. Alphabet's Google Cloud posted $20 billion in revenue in Q1 2026, growing 63% year over year. Microsoft's Azure grew 40%, with the company stating that AI demand had exceeded supply. These two companies are converting AI infrastructure investment straight into cloud revenue. Meta is different. Its data centers exist solely to serve its own products, and that investment is recouped only through a far longer chain that runs through advertising efficiency and user engagement.
Hidden Insight: Is Zuckerberg Playing the Wrong Game?
On the surface, Meta's AI spending looks reckless. But viewed from another angle, an entirely different picture emerges. Meta's platforms have 3.2 billion daily active users. A 1% improvement in AI-driven ad optimization translates into billions of dollars in additional revenue. Meta is not trying to win the cloud infrastructure race. It is trying to build an AI advertising monopoly backed by the behavioral data of 3.2 billion people. The real risk, however, is that this strategy depends not on economies of scale but on the quality of its AI models. If Meta fails to keep supplying the best open-source AI, competitors could catch up to its ad-targeting capabilities. For the only one of the five hyperscalers spending more than $650 billion in 2026 alone that has no cloud revenue, the ROI question is never a technical one. It is an existential one.