Oct 7, 2026 · 9 min
Data centers test the promise of innovation against local power costs
One economist's answer to data center anxiety
AI infrastructure could make electricity more expensive in the near term, even if higher prices eventually encourage efficiency and new energy supplies.
- 1The failed 1980 Global 2000 forecast shows how badly energy predictions can miss innovation’s response to scarcity.
- 2Data-center companies are pursuing new power sources while improving AI efficiency and lowering the cost of queries.
- 3Communities may bear short-term costs, making local decisions about subsidies, regulation, and grid upgrades consequential.
Don't miss
Julia Cartwright explains why the 1980 Global 2000 energy forecast failed, arguing that high prices changed technology and supply in ways it overlooked.
The brief
AI data centers are fueling fears of strained grids and higher household electricity bills, reviving an old question: can energy supply keep pace with demand?
Economist Julia Cartwright revisits the 1980 Global 2000 Report, whose prediction of severe scarcity and a 150% rise in real energy prices failed in magnitude and direction.
Cartwright argues that high prices can trigger fuel-efficient cars, new drilling methods, kerosene, fiber optics, and other innovations that loosen resource constraints.
Data-center companies are investing in power plants, including nuclear facilities, while rapid improvements in AI efficiency reduce the cost of running queries.
The long-run case for innovation does not erase near-term pain: communities may face higher costs and must weigh subsidies, tax breaks, grid upgrades, and regulation.