Carlota Perez published Technological Revolutions and Financial Capital in 2002, right after the dot-com crash, which made her timing look uncanny. She describes five great surges of development since the 1770s, from the first mechanized cotton mills to the age of information technology.
Each surge follows a pattern. In the installation period, financial capital floods into the new technology and a frenzy builds until the bubble bursts. A turning point follows, usually with new institutions and new rules. In the deployment period the technology spreads through the whole economy and production capital takes the lead.
Where AI sits on her curve
Perez has argued for years that the information age is stuck at its turning point. Read that way, the AI investment boom is installation money arriving inside a revolution that started in 1971.
The most useful part of her book for anyone planning AI budgets is the role of the bubble. Speculative money builds infrastructure sober money would never fund, railways in the 1840s and fibre in the 1990s. The investors often lose their money, and later companies get the railways and cables cheaply. Many of the data centres built in this boom will outlast the companies that paid for them.