Copper is easy to miss in an AI supply-chain discussion because it is familiar. Yet the expansion of data centres depends on conductors across the campus and far beyond it: utility lines, transformers, switchgear, busbars, backup systems, cooling equipment and the electrical paths inside servers. The metal’s role is distributed through the system rather than concentrated in one headline component.
The IEA’s 2025 outlook projects total copper demand in its stated-policy scenario rising from 26.717 million tonnes in 2024 to 31.348 million tonnes in 2030. That is an increase of 4.631 million tonnes, or about 17.3%. Primary supply requirements rise by 3.414 million tonnes, about 15.2%, while secondary supply and reuse also grow.[1]
Copper demand and supply context
| Measure | Value | Context |
|---|---|---|
| Total demand, 2024 | 26.717 million tonnes | IEA stated-policy scenario baseline.Source 1 |
| Total demand, 2030 | 31.348 million tonnes | IEA projection, not an observed outcome.Source 1 |
| Increase | 4.631 million tonnes, about 17.3% | Our subtraction and percentage calculation.Source 1 |
| Primary supply requirement, 2030 | 25.917 million tonnes | Up from 22.503 million tonnes in 2024.Source 1 |
| Top three mining countries, 2030 | 50% share | IEA projection, up from 48% in 2024.Source 1 |
The campus boundary understates the material system
A data centre’s electrical demand can trigger generation and transmission work many kilometres away. DOE’s demand-growth hub cites data-centre expansion and AI as major drivers of rising U.S. electricity demand. The copper requirement therefore includes grid reinforcement and new generation as well as the facility’s internal distribution.[4]
The same megawatt of computing can imply different material intensity depending on location and design. A site beside existing generation and transmission may need less new line construction than an isolated campus. The conductor material also matters: most overhead transmission lines use aluminium around a steel core, as the national laboratory’s account explains. Grid expansion therefore cannot be converted wholesale into copper demand. Estimate copper from the actual cables, transformers, equipment and designs, including substitution and redundancy choices.[6]
Where copper hides in the AI electricity chain
| Layer | Typical copper role | What drives quantity |
|---|---|---|
| Regional grid | Copper in selected cables, substation equipment and transformers; overhead line conductors are commonly aluminium-based | Distance, voltage, capacity and existing headroomSource 4Source 6 |
| Campus power | Cables, busbars, switchgear, UPS and backup systems | Peak load, redundancy and distribution designSource 5 |
| Cooling and mechanical systems | Motors, pumps and heat-rejection equipment | Cooling technology and climateSource 5 |
| Servers and networking | Boards, connectors, power delivery and cabling | Hardware density, topology and refresh cycleSource 5 |
| New generation | Generators, renewable plants, storage and interconnection | Energy mix and locationSource 1 |
AI is one demand driver inside a much larger copper market
The IEA figures combine clean-energy technologies with construction, manufacturing, electronics and other uses. It would be wrong to assign the projected 4.631-million-tonne increase to AI. A defensible AI estimate would need project-level electrical designs, grid plans, hardware shipments and allocation rules that avoid double counting copper already captured in energy-sector forecasts.
USGS added copper to the 2025 U.S. critical-minerals list and identifies its widespread use in wiring and cables. The classification reflects economic importance and supply-chain risk, not physical scarcity in the ordinary sense. It should prompt attention to permitting, mine productivity, refining, recycling and substitution rather than a simplistic claim that the world is “running out.”[3]
Copper may be the AI metal most likely to hide in plain sight. It will not dominate model-release headlines, but every new megawatt has to be generated, moved, converted and delivered. The most useful analysis therefore starts with an electrical diagram and follows the conductor outward until the new infrastructure ends.
Sources & further reading
- Copper International Energy Agency · 21 May 2025 · Accessed 8 Oct 2026
- Critical Minerals Data Explorer International Energy Agency · Accessed 8 Oct 2026
- Interior Department releases final 2025 List of Critical Minerals U.S. Geological Survey · 14 Nov 2025 · Accessed 8 Oct 2026
- Electricity Demand Growth Resource Hub U.S. Department of Energy · Accessed 8 Oct 2026
- Key Questions on Energy and AI International Energy Agency · 16 Apr 2026 · Accessed 8 Oct 2026
- With Great Power Comes Great Possibility National Renewable Energy Laboratory / National Laboratory of the Rockies · Accessed 15 Oct 2026