Home / Wiki / A / AI Infrastructure Cycle

AI Infrastructure Cycle

Investment and Capital MarketsUpdated 16 Nov 2026

The AI infrastructure cycle is the multi-year wave of investment in the data centers, chips, power supply and networks needed to train and run artificial intelligence models. Analysts use the phrase to treat spending on AI infrastructure as a capital expenditure cycle with phases: an initial build-out led by a few large technology companies, a broadening to suppliers and financiers, and at some point a slowdown as capacity catches up with demand.

The framing borrows from earlier investment booms, such as railways and the telecom networks of the late 1990s, in which heavy spending was followed by excess capacity and falling prices. Investors watch company spending plans, chip orders, electricity contracts and measures such as data center systems spending for signs of where the cycle stands. Opinions differ on its length and on whether it will end gradually or as a bubble. Forecasts of future annual spending vary widely and should be read as estimates.

Last updated 16 Nov 2026