Data centers and the Jevons Paradox

The Jevons Paradox dictates that as technology makes a process more efficient, demand for its inputs rises rather than falls. William Stanley Jevons first observed this when James Watt’s coal-efficient steam engine democratized steam power, causing coal demand to surge (Divya Venn from observing happiness). This dynamic now repeats in artificial intelligence, where the new Jev model performs classification tasks 200 times faster and 400 times cheaper than frontier LLMs, yet fuels a surge in demand for compute (Divya Venn from observing happiness).

Data centers face physical constraints as regional power grids struggle to supply sufficient electricity to meet this demand. Operators like Fjernvarme Fyn in Denmark demonstrate a physical response by supplying surplus heat to 12,000 homes, a move supported by the abolition of surplus heat taxes (Global Data Center Hub). While traditional control systems struggle with the rapid instabilities inherent in fusion reactors, startups like Radical AI and Citrine Informatics now use AI models to predict these disruptions milliseconds before they occur, demonstrating how technology can be redirected toward solving the energy supply chain bottlenecks it creates.

These physical interventions contrast with the symbolic abstraction of financial markets, where 90% of wealth owners prioritize AI-driven labor replacement over the physical reality of diesel-dependent logistics networks (Charles Hugh Smith's Substack). As the issue on Wednesday showed, corporations often deploy capital into infrastructure before confirming that financial returns justify the expenditure, ignoring that electrons in a data center cannot refill an aquifer or replace the physical maintenance required for rail and road networks.