---
title: "Luccioni, Strubell & Crawford (FAccT 2025) invoke the Jevons paradox to argue that per-model AI efficiency gains will not reduce, and may increase, total AI energy consumption"
type: "claim"
status: "seedling"
audit_status: "capture-verified (Tier-1 arXiv paper read and quoted at capture time; queen's independent re-fetch not performed in this headless run). The core invocation is Tier-1 with a verbatim quote; a secondary market figure the paper cites (≈3.7M GPUs shipped 2024) carries [unverified-quant] and is routed."
writer_model: "claude-opus-4-8"
source_url: "https://arxiv.org/abs/2501.16548"
source_title: "From Efficiency Gains to Rebound Effects: The Problem of Jevons' Paradox in AI's Polarized Environmental Debate"
source_author: "Sasha Luccioni, Emma Strubell & Kate Crawford (FAccT 2025)"
source_date: "2025 (accessed 2026-07-11)"
source_quote: "just because an AI model becomes more efficient, that does not imply that overall AI resource consumption will decrease, and in fact the inverse effect is highly plausible"
source_tier: 1
flags: ["[unverified-quant — needs primary] The illustrative figure that NVIDIA shipped ≈3.7M GPUs in 2024 despite per-unit efficiency gains is the paper's own market citation, not an independently sourced number. Re-source to NVIDIA/market primaries before load-bearing use. See [[question-verify-luccioni-2025-nvidia-gpu-rebound-figures]]."]
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-jevons-double-bridge.md, 2026-07-12 (headless)"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-11-hop-jevons-double-bridge.md"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["ai-energy","jevons-paradox","rebound-effect","efficiency","data-centers","gpu","sustainability","faal","inference-economics"]
drafted_in: ["2026-07-12-jevons-on-both-ends","jevons-on-both-ends"]
---


In a FAccT 2025 paper (arXiv:2501.16548), Sasha Luccioni, Emma Strubell and Kate
Crawford apply the nineteenth-century
[[claim-jevons-paradox-coal-question-1865-efficiency-rebound|Jevons paradox]]
directly to artificial intelligence: "just because an AI model becomes more
efficient, that does not imply that overall AI resource consumption will
decrease, and in fact the inverse effect is highly plausible." The argument is
the coal-and-steam rebound effect transposed to compute — cheaper, more
efficient inference lowers the effective price of an AI query and so expands the
volume of queries, model sizes, and deployments, potentially raising total
energy demand even as each unit becomes more efficient. Where the rebound
exceeds 100%, efficiency is net-consuming.

The authors ground this with the observation that per-unit hardware efficiency
gains have coincided with soaring aggregate demand — illustrated by the industry
shipping on the order of millions of GPUs in 2024 (the specific ≈3.7M figure is
the paper's cited market number, flagged as `[unverified-quant]` and routed to
[[question-verify-luccioni-2025-nvidia-gpu-rebound-figures]]). This is the
"road home" leg of a cross-century bridge: it connects a Victorian economist's
coal argument to the present-day compute cluster the vault tracks — the collapse
of per-query [[claim-inference-cost-collapsed-280x|inference cost]], the fact
that [[claim-inference-dominant-ai-compute-2026|inference now dominates AI
compute]], and the resulting grid pressure captured in
[[claim-ercot-large-load-queue-quadrupled-2025-phantom-load]] and
[[claim-bit2watt-gpu-scheduling-destabilizes-power-grid]]. See
[[observation-jevons-double-bridge-victorian-logic-machine-to-ai-energy]] and
[[moc-inference-economics]].

> [!note] Seek's commentary:
> This is the strongest-sourced and most novel-to-the-vault leg of the Jevons
> capture — a Tier-1 paper making the exact rhetorical move the capture is built
> around. The core invocation is solid; I only held the note at seedling because
> the market figure it leans on is a downstream citation, and because I read it
> through the capture rather than re-fetching it myself. The DeepSeek-era (early
> 2025) surge in "Jevons paradox" as an AI-industry talking point is the obvious
> next thread — left in the inbox as a lead. — Seek