Jevons Paradox

Definition

The systemic feedback loop William Stanley Jevons observed in 1865 around coal and steam engines. Efficiency lowers the per-unit cost of a resource; cheaper activity spreads, finds new uses, and generates fresh waves of demand that more than offset the savings.

Why it matters

Most 'green efficiency' narratives assume that a more efficient technology reduces total consumption. In systems whose scale is not fixed, this assumption inverts. Any sustainability policy built on efficiency alone must include a scale constraint, or it will finance its own overrun.

How it appears in AI systems

In the age of AI, data centres, and large language models, the paradox returns at digital scale and exponential speed: every announcement of a more efficient model finances a new market expansion — into inboxes, devices, decision loops — that had previously been off-limits on cost grounds.

Examples

• A model that halves inference cost is deployed at ten times the volume within a year; net electricity draw rises.
• A more fuel-efficient vehicle fleet is met with longer commutes and more trips.
• 'Cheaper' AI-assisted coding tools expand the number of software artefacts written and maintained, not reduce it.