Understanding rebound effects in digital sustainability
2026-05-17
Cutting a datacenter's energy consumption by 30% sounds like a clear win. But what happens when that efficiency makes hosting so cheap that demand takes off? That is precisely what rebound effects describe.
Rebound effects are often overlooked when studying the environmental impact of digital technologies. They describe situations where efficiency gains lead to increased usage, partially or entirely offsetting the expected environmental benefits.
The Jevons paradox
The concept has historical roots: in 1865, economist William Stanley Jevons observed that improvements to steam engine efficiency led to more coal consumption, not less. Cheaper operation made coal more accessible, which stimulated broader use. The same logic applies to digital infrastructure today.
William Stanley Jevons (1835–1882)
Four types of rebound in the digital sector
Direct rebound occurs when a more efficient service becomes cheaper to run or use, leading to higher consumption. Streaming video, for instance, uses less energy per hour than it did a decade ago, but total viewing hours have grown dramatically.
Indirect rebound happens when savings from efficiency are redirected toward other energy-intensive activities. The money saved doesn't disappear; it gets spent elsewhere.
Structural rebound refers to how digital efficiency reshapes entire industries. AI-optimised logistics enables on-demand delivery models that didn't previously exist, generating a new category of environmental cost.
Economy-wide rebound operates at the macro level: digital productivity gains stimulate economic growth, which drives higher overall energy and material consumption.
AI as a new catalyst
The rise of generative AI illustrates this dynamic clearly. As inference becomes cheaper, access to models widens and the number of requests scales accordingly. When the marginal cost of generating text or images approaches zero, usage expands into contexts where it was previously unthinkable. Efficiency enables scale, and scale consumes resources.
Why this matters for measurement
Most environmental assessments of digital systems rarely account for rebound effects, because they occur outside the directly measured perimeter. An efficiency gain looks like a win on paper while driving increased consumption in practice.
Some emerging responses include decoupling energy pricing from efficiency gains, incorporating real-world usage metrics into digital environmental labels, and setting absolute consumption caps rather than relative intensity targets.
The key takeaway
Technological efficiency is necessary but not sufficient. Without mechanisms to constrain demand growth, environmental gains risk being absorbed by the expanded usage they make possible. Understanding rebound effects is essential for anyone trying to reduce the digital sector's environmental footprint in absolute terms, not just relative ones.