Jevons Paradox vs Deflationary Technology

Cause and Effect: The Deflationary-Jevons Flywheel

The most important insight about these two concepts is that they are not alternatives — they are sequential stages of the same economic cycle. Deflationary technology creates the conditions for Jevons' Paradox to operate. When inference costs drop 92%, that is a deflationary technology event. When that drop triggers a surge in total compute spending that exceeds what was saved, that is Jevons' Paradox in action. Together with Wright's Law, they form a flywheel: increased production drives costs down, lower costs expand demand, and expanded demand funds even more production.

This flywheel has operated across every major technology wave. In semiconductors, Moore's Law drove deflationary cost reductions for decades. Jevons' Paradox ensured that cheaper transistors didn't shrink the chip market — they created entirely new categories (PCs, mobile phones, IoT, AI accelerators) that consumed orders of magnitude more silicon than the mainframe era. The same pattern is now unfolding in AI, but at a pace that compresses decades of disruption into years.

Demand Elasticity: Why AI Is Different

Not every efficiency improvement triggers a full Jevons effect. The paradox only holds when demand is sufficiently price-elastic — meaning that a cost reduction generates a proportionally larger increase in consumption. What makes AI unusual is that demand for compute appears to be super-elastic. Research from Zhang and Zhang (January 2026) formalizes this as a Structural J-curve in their paper The Economics of Digital Intelligence Capital, demonstrating that the addressable market for machine intelligence expands faster than costs fall.

This super-elasticity stems from the nature of intelligence as a resource. Unlike coal or steel, intelligence can be applied to an essentially unbounded set of problems. Every cost threshold that falls opens entirely new application categories: agentic commerce, AI code generation, personalized education, real-time translation, and autonomous research. The demand curve for intelligence doesn't flatten — it steepens, because each new application reveals further applications that were previously inconceivable at higher price points.

The Incumbent Trap: Deflation Destroys, Jevons Rebuilds

Deflationary technology is both creator and destroyer. It follows a predictable pattern where incumbents resist cost decline to protect margins while insurgents build natively on the cheaper substrate. The SaaSpocalypse is a current example: SaaS companies that charged premium subscriptions for features that AI can now commoditize face the same structural pressure they once imposed on on-premise software vendors. But Jevons' Paradox tells us that the total market doesn't shrink — it reallocates and grows.

The 2025 capital expenditure arms race illustrates this perfectly. Despite — and because of — collapsing inference costs, Meta, Alphabet, Microsoft, and Amazon collectively committed over $300 billion in AI infrastructure spending. The deflationary force made each dollar of compute cheaper; the Jevons effect ensured those savings were reinvested many times over. Incumbents who understand both dynamics can ride the flywheel rather than be crushed by it.

The Energy and Infrastructure Paradox

One of the most consequential applications of the Jevons-deflation interaction is in AI energy consumption and data center infrastructure. Algorithmic efficiency improvements have reduced the energy cost per inference by roughly 40% annually at the hardware level. But total energy consumption by AI workloads continues to climb dramatically, because the volume of inference calls is growing far faster than per-call efficiency improves. A January 2025 paper in arXiv examined this directly, warning that efficiency gains in AI create rebound effects that undermine sustainability goals.

This has direct implications for infrastructure planning and climate policy. Organizations cannot assume that more efficient models will reduce their total compute footprint. The historical pattern — from coal to cars to cloud computing — consistently shows that efficiency gains lower the barrier to broader adoption, expanding total resource use. Strategies for managing AI's environmental impact must account for this dynamic rather than relying on efficiency alone.

The Creator Economy and Democratization Effect

Where deflationary technology and Jevons' Paradox intersect most visibly is in the expansion of who can participate in the economy. When the cost of writing code approaches zero via AI code generation, the bottleneck shifts from implementation to product vision. When the cost of running AI on every customer interaction approaches zero, agentic commerce replaces traditional interfaces. The Creator Economy expands to include millions of new participants because deflationary technology removes cost barriers while Jevons' Paradox ensures total creative output explodes rather than consolidates.

This democratization is the most optimistic implication of the flywheel. Each deflationary wave has historically expanded participation: cloud computing enabled the SaaS revolution, mobile enabled the app economy, and collapsing AI costs are enabling a generation of solo founders and micro-teams to build products that previously required large engineering organizations. The Jevons effect guarantees that this expansion creates net new demand rather than simply redistributing a fixed pie.