Two companies, one thesis about the future, completely opposite market reaction. Microsoft added to its capitalization in a single day a sum that exceeds the value of most companies in the S&P 500 entirely. Meta lost approximately the same percentage. The reason is not general pessimism about AI — it's that investors finally started asking a specific question: who pays for this infrastructure, besides the company itself.
Microsoft has an answer, and it's simple. Azure grew 43% year-over-year, and Copilot paid users reached 30 million — 10 million more than four months ago. This is not an abstract bet on the future, but a subscription that someone pays for every month right now. The company is converting AI spending into revenue that can be counted in a financial report, not in a presentation for investors.
Meta has no such line. The company's free cash flow fell 91% — to $784 million, and this is not due to a one-time event but due to continuous increases in capital expenditures for data centers. Revenue guidance for the current quarter also came in below analyst expectations. Zuckerberg acknowledged in his comments a strange situation: companies offer to rent Meta's computing power at a price that significantly exceeds the cost of creating it, but Meta refuses because it wants to use these resources for its own products.
This is where a practical nuance lies that is easy to miss. Microsoft sells access to its infrastructure directly — through Azure and Copilot, and sees money immediately. Meta, meanwhile, sits on an asset that the market is ready to rent at a premium price right now, but is holding onto it for the sake of a product whose results are unknown. This is the difference between monetization that can be measured this quarter and monetization promised a year or two from now.
For the market, this became a moment of truth regarding the entire wave of capital expenditures on AI in large tech companies. If Meta does not show in the coming quarters its own analog of "30 million subscribers" — a concrete metric that turns billions of investments into revenue — investor pressure will demand either accelerating the monetization of AI-based products, or starting to lease part of the infrastructure to those who are already ready to pay for it today.
