Sometimes progress works in surprising ways. We often assume that making technology more efficient will reduce resource use but history repeatedly shows the opposite.
Jevons Paradox describes this counterintuitive effect: when something becomes more efficient and cheaper to use, total consumption can actually increase. Efficiency lowers costs, unlocks new uses, drives up demand, and ultimately expands overall resource use.
The idea dates back to 1865, when economist William Stanley Jevons observed an unexpected trend in Britain’s coal industry. James Watt’s improved steam engine consumed far less coal than the older Newcomen design. Logic suggested that coal demand would fall. Instead, it soared. The cheaper, more efficient engine made steam power accessible, fuelling rapid industrial expansion. Factories, railways, shipping and mining all adopted the new technology, and national coal consumption grew dramatically.
Jevons realised that efficiency can act like an accelerator, making resources more attractive and spreading their use further.
The Modern Twist: AI and the New Efficiency Trap
Over time, we have seen this play out number of times:
- As it became cheaper to make clothes, brands make more of them and as more of them are made, demand is created to purchase them.
- As food became cheaper to produce and more people didn't have to go hungry anymore, the abundance has driven unhealthy eating habits and production patterns.
- In manufacturing, efficiency and new materials inventions made everyday objects like furniture and electronics cheaper to produce and they quickly became seen as disposable.
Now, the Jevons Paradox is resurfacing in the world of artificial intelligence.
AI chips and models become more efficient, they become cheaper to train and deploy. This accelerates adoption - more models, more data centres, more embedded AI services across every sector. Even if each model requires less energy per operation, the total energy demand of the expanding ecosystem keeps rising. In Ireland, for example, data centres now consume around 21% of the country’s electricity, up from just 5% in 2015. *
This is efficiency-driven growth, not efficiency-driven reduction.
How Falling Energy Costs Can Undermine Climate Gains
The environmental impact of rising digital demand has been partially cushioned by growth in renewable energy. Wind and solar, with very low marginal costs, often replace more expensive fossil generation and lower wholesale electricity prices.
But cheaper electricity has an unintended consequence: it encourages more consumption - more cooling, more manufacturing, more digital services, more applications that would have been uneconomical under higher prices.
This is the rebound effect. In some cases, rebound can offset much of the benefit of efficiency or clean energy. In extreme cases, it can even reverse it; a situation known as “backfire.”
For example, according to Ember, since 2015, solar generation has increased eightfold, while wind generation more than tripled **. At the same time, fossil-fuel electricity generation also continues to rise in absolute terms: Ember reports a +0.8% increase (~135 TWh) in fossil generation in 2023 compared to 2015‑2022 trends***.
In other words, most new renewable energy was additional; meeting new demand, rather than replacing fossil fuel.
More efficiency + more renewables ≠ lower total energy use, unless demand is addressed.
When Efficient Products Increase Total Emissions
When products become more resource-efficient or lower-carbon, companies often sell more of them. The per-unit footprint drops, but total units sold rise, pushing absolute emissions up.
Large furniture and homeware retailers (including companies such as IKEA) have openly acknowledged this challenge; Products are becoming more energy-efficient, lower-carbon, and more circular, but selling more units can still drive up absolute emissions, supply-chain impacts, and material use. In other words, efficiency gains don’t always reduce total impact, they can sometimes fuel further growth.
To solve this, companies are exploring how to decouple financial performance from environmental impact, for example by:
- shifting revenue toward services (repair, rental, reuse, refurbishment)
- designing long-lasting products that reduce turnover
- increasing revenue per unit of material used
- tracking emissions intensity per unit of revenue or margin, not only per product
In a way - using resources more efficiently so that the value generated grows, but not the use of resources themselves. These metrics help companies understand whether they’re truly reducing their negative impact, or simply scaling efficient products to a level that cancels out the gains.
Why This Matters: Absolute vs Relative Metrics
Lowering the carbon footprint per product is not enough if the total number of products sold keeps rising.
Companies therefore measure impact in two ways:
1. Absolute emissions
- Total tonnes of CO₂e emitted.
- This shows whether the company is actually reducing environmental impact.
2. Intensity metrics
- Emissions per revenue
- Emissions per gross margin
- Emissions per unit sold
- Emissions per square metre of product
Tracking both provides a realistic picture of progress, and prevents efficiency from masking growing impacts.
If efficiency reliably leads to greater demand (as history shows), then companies need business models that enable growth without scaling environmental harm.
This is the essence of absolute decoupling, increasing revenue while reducing total planetary impact.
Conclusion: Efficiency alone isn’t enough.
To avoid rebound effects and hidden growth in resource use, companies need to track and report both relative and absolute emissions; from production to use to end of life.
The Jevons Paradox reminds us that technological efficiency, while essential, is not a guaranteed path to sustainability. Real progress requires pairing efficiency gains with smarter design, responsible scaling, demand-side measures, and policies that keep total resource use within planetary limits.
Only by tackling both efficiency and total consumption can innovation genuinely lighten our impact - rather than quietly increasing it.
References:
To demonstrate that this phenomenon is not specific to humans and is part of nature:
🐜 Ant Colonies & Foraging Efficiency
- As an ant colony grows, it evolves or develops more efficient foraging trails (shorter, more optimized pheromone paths).
- The increased efficiency allows the colony to grow larger.
- A larger colony needs more food overall, so it forages more, spreading further and consuming more resources.
🐺 Predator Hunting Efficiency → Prey Overconsumption
- A predator that evolves a more efficient hunting strategy (cooperative hunting, better sensory abilities) can capture prey much more cheaply (in energy terms).
- This efficiency allows the predator population to grow.
- Larger predator populations consume more prey overall, potentially stressing the ecosystem.
🔍 Why this is like Jevons:
Efficiency → population expansion → total consumption increases.
Wolves & Cooperative Hunting
Study on wolves hunting bison
- Research has shown that wolf pack size strongly influences their success rate when hunting very large and dangerous prey like bison. PMC+1
- Specifically: as the number of wolves in a pack increases, their success at killing bison increases beyond the point where small packs plateau. PMC
- This suggests cooperative hunting adds a kind of “efficiency boost” for very challenging prey. Large groups can tackle prey that would be hard or very costly for smaller groups to subdue. PLOS
- Because they can take down more formidable prey effectively, this may support (or at least permit) larger wolf populations, since high-payoff prey is more reliably available.
You have come to the place for thought leadership.
