---
title: "Wright's law: unit cost falls a constant fraction with every doubling of cumulative production"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://www.newthingsunderthesun.com/pub/6b1c38y9/release/5"
source_author: "New Things Under the Sun (Matt Clancy); cross-referenced against Our World in Data's learning-curve explainer (ourworldindata.org/learning-curve)"
source_date: "2026-07-11T00:00:00.000Z"
source_quote: "if the progress ratio is 0.8 ... the unit cost is 80% of what it was before"
source_tier: 3
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-real-disruptions-fall-monotonically.md, 2026-07-12"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-11-hop-real-disruptions-fall-monotonically.md"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["economics","wrights-law","learning-curve","experience-curve","price-dynamics","resource-economics"]
drafted_in: ["the-line-no-one-walks"]
---


Wright's law — also called the experience curve or learning curve — states that as cumulative production of a good doubles, its unit cost falls by a constant percentage, the "progress ratio": "if the progress ratio is 0.8 ... the unit cost is 80% of what it was before" each time cumulative output doubles. Under competition, price tracks cost down continuously across the whole scaling trajectory, rather than holding at an old level until a late collapse.

This is the proposed general mechanism behind two documented monotonic-decline cases: [[claim-aluminium-price-fell-monotonically-after-hall-heroult]] (Hall-Héroult aluminium, 1880s–90s) and [[claim-inference-cost-collapsed-280x]] (AI inference, 2022–2024). It is the structural opposite of the hold-then-collapse shape predicted for asteroid-mined platinum-group metals in [[claim-asteroid-pgm-price-holds-then-collapses]] — see [[claim-cheaper-extraction-disruptions-fall-monotonically-not-hold-then-collapse]] for the synthesis across all three.

**`[unverified-mechanism — needs primary]`.** Both cited restatements (New Things Under the Sun, Our World in Data) are secondary explainers, not the originating economic literature — T.P. Wright's 1936 aircraft-manufacturing cost paper, or the Boston Consulting Group's original 1960s–70s experience-curve studies. Per the vault's sourcing floor, a specific technical-mechanism claim needs a Tier 1–2 primary; a named-analyst blog post and a data-journalism explainer clear the floor for "what is this concept" but not yet for "this is exactly how it works and how well it's supported empirically." Verification routed to [[question-verify-wrights-law-primary-source]] — partially answered by [[claim-sahals-identity-equates-wrights-law-and-moores-law]], a Tier 1 primary (Lafond et al. 2017) that empirically tests this mechanism, though it is not the 1936 Wright original the question asks for.

> [!note] Seek's commentary:
> Wright's law is well-established enough in the economics and operations-research literature that I'd be surprised if the primary check turns up trouble — but "well-established" is exactly the kind of belief that should still cash out to a primary source before this goes evergreen. Flagged as a candidate seed by the capture itself (novelty 0.674); no atomic note existed for it before this one. — Seek