The Copper Constraint: Price Formation Under Truncated Supply Elasticity in the AI Infrastructure Buildout

Abstract

Copper prices rose 56 percent above their 2024 average within two years while the refined market recorded a surplus in 2025 and never departed meaningfully from balance. This paper reconciles that pattern with a market-clearing model in which the binding constraint is not the current balance but the supply elasticity available inside a finite window. Motivated by discovery-to-production lead times that now exceed the forecast horizon, the central restriction truncates the twenty-year supply elasticity of 0.77 to a prior centred near 0.42 for 2024-2035. A 20,000-draw Monte Carlo over 29 literature-calibrated inputs yields median 2035 prices of 6.75 USD per pound on a 2024 anchor and 8.42 USD per pound on a 2026 anchor. A Shapley decomposition with an embedded null control finds supply-side parameters dominant and, on the ex-ante anchor, every AI-specific input indistinguishable from noise: an AI capex bust moves 2035 by 0.47 USD per pound, permitting reform by 2.25 USD per pound. AI is the marginal buyer of a market that cannot expand quickly; the constraint is a property of supply, not of AI demand.