Amazon and Google Lead the $400B AI Capex Arms Race — But Where’s the ROI?

The AI industry is in the midst of a data center arms race. Giants believe that controlling the most computing power will determine the winner in future AI products. Amazon is leading the charge, projecting $200 billion in capital expenditures for 2026; Google follows closely ($175-185 billion); Meta, Microsoft, and others are also making massive investments.

(Google CEO)

The underlying logic is that high-end computing will become a scarce future resource, and only those who build their own supply chains will survive. However, the market has reacted strongly—every company announcing huge spending has seen its stock price drop immediately, with higher investments correlating to steeper declines.

This is not just a problem for companies without a clear AI strategy (like Meta). Even firms with mature cloud businesses and clear monetization paths, such as Microsoft and Amazon, are facing pressure. Expenditures reaching hundreds of billions of dollars are testing investor patience.

While Wall Street’s nervousness may not alter the tech giants’ strategic direction, they will increasingly need to downplay the true cost of their AI ambitions. Behind this computing power contest lies the ultimate between technological innovation and capital’s patience.

Roger Luo said:The current AI computing power race has transcended mere technology, evolving into a capital-intensive strategic game. While giants are betting that computing power equals dominance, they must guard against the potential pitfalls of heavy-asset models—capital efficiency traps and innovation stagnation.

All articles and pictures are from the Internet. If there are any copyright issues, please contact us in time to delete.

Inquiry us