Big Four's AI Capital Expenditure May Exceed 700 Billion USD This Year

Alphabet, Meta, Microsoft, and Amazon successively announced their Q1 reports this week. The combined capital expenditure plans of these four companies for this year have risen to $725 billion, with a primary focus on AI data center equipment.

Specifically, Alphabet and Meta both raised their full-year capital expenditure guidance, while Microsoft provided its first full-year forecast, which is on par with Alphabet at $190 billion. Amazon is the only one among the four that maintained the original plan, with a full-year capital expenditure of $20 billion. Due to the significant increase in actual expenditures in the first quarter, free cash flow has accordingly narrowed significantly.

Meta CEO Mark Zuckerberg stated in an analyst call that the main reason for increasing the expenditure plan is the rise in hardware costs, especially the price of memory. He also emphasized that multiple internal and external indicators at the company validate the reasonableness of this investment. Meta increased the upper limit of its full-year capital expenditure range to $145 billion.

The Q1 financial performance exceeded market expectations overall, but market reactions were not consistent.

The performance of Amazon and Alphabet was more striking, while Meta's expenditure structure received more scrutiny. Unlike Microsoft, Amazon, and Google, Meta does not provide cloud computing services to external customers and cannot offset the risks of excessive investment by renting out idle computing power, making its large-scale capital expenditures appear riskier in the eyes of the market.

For investors, the core issue to balance at present is: before the commercialization benefits of AI are fully realized, how long can this round of expenditure competition last, and whether each company's monetization path is clear enough.

Cloud vendors at least have the "build more to rent more" buffer mechanism, while players who purely bet on self-use AI infrastructure have stricter requirements for the return schedule.

The subsequent points of attention include: whether the price trend of memory and other core hardware will further push up expenditures in the second half of the year, the actual realization speed of AI-related revenue of various companies, and where the boundary of the capital market's patience for this expenditure cycle lies.