Jevons Paradox refers to the phenomenon where improvements in energy efficiency—intended or assumed to reduce overall consumption—can actually lead to more total resource use, not less.

      **This paradox was first articulated by economist William Stanley Jevons in 1865, who observed that as coal-burning steam engines became more efficient in England, coal consumption actually increased. The reason: efficiency made coal cheaper to use per unit of output, which in turn made coal-powered applications more economically attractive, expanding their use.**

       **This concept applies to modern energy systems and global economic growth. When we make energy use more efficient—whether in lighting, transportation, manufacturing, or computing—it often lowers costs and stimulates demand elsewhere in the system. For example, fuel-efficient cars might reduce fuel use per mile, but if people drive more or buy more cars as a result, total energy use and emissions can rise.**

        **Thus, Jevons Paradox undermines the common assumption that technological efficiency alone will solve ecological overshoot or climate change. Instead, it highlights the need to focus on absolute reductions in energy and material throughput, not just relative efficiency gains.**

       **A specific modern case of Jevons Paradox in renewable energy is the example of data centers powered by cheap solar and wind energy, especially in regions like the southwestern U.S. and parts of Northern Europe.**

       **As solar and wind became cheaper—often subsidized or backed by government incentives—large tech companies like Google, Microsoft, and Amazon built massive data centers designed to run on “100% renewable energy.” On the surface, this looks like an environmental win.**

       **But here’s the paradox: the low cost and positive optics of “green power” enabled the rapid expansion of data center infrastructure. More processing capacity invited more cloud services, more AI model training, more streaming, and more consumer usage. Some of these services, such as generative AI, are extremely energy intensive—training a large language model can consume as much electricity as hundreds of U.S. households use in a year.**

       **So rather than reducing global energy demand, renewable-powered data centers have contributed to a surge in digital energy consumption, much of which is still backed by natural gas or coal at night or during demand spikes. This is especially true in grids that lack sufficient storage or long-distance transmission.**

Global Energy, Our World in Data, Website

**https://ourworldindata.org/grapher/primary-sub-energy-source**

Jevons' Paradox, Great Introduction. Nate Hagens, you tube, 2 minutes

**https://www.youtube.com/shorts/FHzIZcaaYXw**

Jevons' Paradox and the clean energy transition, Wisconsin Energy Institute, Youtube 60 minutes

**https://www.youtube.com/watch?v=6_LMkeclN7E**

Charts / Drawings

Global Energy Use 1800-now.pdf

There is no energy transition.pdf