Argument
The Jevons Paradox — first observed in 1865 when William Stanley Jevons showed that more efficient steam engines consumed more coal, not less — explains why Google’s memory breakthrough (LPDDR5X flash, April 2026) crashed chip stocks, then made them rally a week later. Cheaper, more efficient AI doesn’t reduce demand for compute, energy, chips, or labor inputs; it scales appetite for the underlying resources.
The piece applies the paradox in three registers: hardware (more efficient memory → more AI deployed → more memory demanded), energy (Jevons at grid scale), and labor (the jobs that survive AI are the jobs AI makes more in demand, not the ones it replaces).
Structure
Opening hook: Every efficiency gain in your adult life — faster internet, bigger drives, more powerful chips — produced more consumption, not less. Efficiency scales appetite, not capacity.
The Jevons core: Introduces the 1865 coal/steam engine paradox. “It is wholly a confusion of ideas to suppose that the economical use of fuel is equivalent to a diminished consumption.” The coal is now compute.
Google memory breakthrough application: The April 2026 chip stock crash → rally sequence is the contemporary test case. Cheaper memory made AI more deployable at scale, which drove demand for memory up, which reversed the crash.
Energy implications: The same dynamic at the grid level — cheaper AI inference drives deployment scale, which drives energy consumption up, not down.
Labor implications: The piece’s most ambitious register — the jobs that AI makes cheaper to do get done more, not less. The displacement fear is correct directionally but inverted in magnitude: the at-risk jobs are the ones AI makes abundant, and the scarce jobs are the ones AI creates demand for.
Key Claims
- Jevons Paradox is the correct framework for AI’s resource consumption trajectory — efficiency gains accelerate resource demand
- Google’s LPDDR5X memory breakthrough caused stocks to crash (cheaper memory = chip suppliers lose margin) then rally (cheaper memory = more AI deployment = more memory demanded overall)
- Cheaper AI inference won’t reduce energy consumption at the grid level
- Labor market implication: the jobs that survive are those that become more demanded as AI scales the tasks they support, not the jobs AI can’t do
Connected Research
- AI-Proof Majors Anxiety — AP — companion piece on Gen Z anxiety about AI job displacement (70% see AI as threat); the Jevons frame is the analytical response to the anxiety
- Without data centers, GDP growth was 0.1% in H1 2025 — Furman analysis that 92% of 2025 GDP growth came from AI/data centers; direct evidence of Jevons at GDP scale
- Cheaper AI Won’t Use Less of Anything — this article
Entities Referenced
- William Stanley Jevons (British economist, 1835-1882, discoverer of the paradox)
- Google (memory breakthrough catalyst)
Concepts to Link
- Jevons Paradox (no dedicated concept page yet — candidate for creation)
- AI Infrastructure (see Space-Based Computing, China Space Computing Constellation — SCIO)
Open Questions This Raises
- At what scale does the Jevons dynamic in AI hit a natural resource ceiling (energy, rare earth materials)?
- Does the paradox apply differently to white-collar knowledge work than to physical resource consumption?
- Is the chip stock crash → rally sequence repeatable as a predictive pattern for future AI efficiency breakthroughs?