Copper Supply and Demand: Research Notes

Core thesis

• Before 2025, refined and recycled copper supply and demand were roughly 27–28 million tonnes: about 23.5 Mt of mined copper and 4 Mt of recycled copper.
• By 2030, both supply and demand are projected near 33 Mt. Mined supply is expected to peak around then, creating deficits afterward.
• S&P’s current forecasts still assume linear growth in copper demand from grids and AI data centers; actual growth could materially exceed them.

Primary source for the charts and forecasts: S&P Global, Copper in the Age of AI: Challenges of Electrification, January 2026.

S&P copper supply and demand

Sources of incremental demand

The fastest-growing uses are electric vehicles, AI data centers, defense stockpiles, and grid renewal in Europe and the United States. Estimated annual demand in 2026–2028:

• EVs: 3.2–4.1 Mt
• AI data centers: 1.2–1.6 Mt
• Defense stockpiles: 0.4 Mt
• Grids: 0.75–0.80 Mt in Europe and 0.45–0.55 Mt in the US, or about 1.3 Mt combined. This is not reflected in the T&D line in the table above.

Copper learning notes

These notes emphasize facts and conclusions to reduce cognitive load.

1. Copper conducts electricity and heat well and is corrosion-resistant and ductile. Silver conducts better but costs more than 200 times as much.
2. Aluminum conducts less well but can replace copper in some uses. Strong substitution incentives historically appeared above a copper/aluminum price ratio of 3; recently the threshold has moved toward 3.5 as urbanization, renewables, and data centers change requirements.
3. More than 60% of copper exists as wire, showing the importance of electricity transmission.
4. Economic development and electrification are closely related. Per-capita electricity use in Nigeria may be below 2% of that in the US. About 730 million people lacked electricity in 2024.
5. Per-capita electricity use in developed countries such as the US and Japan peaked around 2007 and has been flat to slightly lower, possibly because industry moved to developing countries.
6. China has been the largest copper consumer for decades. Mining is distributed across South America, Africa, and elsewhere, while mid- and downstream processing is concentrated in China. Stagnant Chinese property and infrastructure would weigh on demand.
7. Copper demand grew about 2% annually over the past 30 years and about 4% over 75 years. Recessions and monetary tightening have historically reduced demand and price. EVs, grids, and data centers are the main incremental sources today.
8. In 2025, global refined copper demand was 27–28 Mt, supplied by more than 23 Mt of mined copper and over 4 Mt of recycled copper.
9. Copper can be recycled. Higher prices necessarily encourage scrap recovery and restrain price.
10. Copper exists in common rock, but a tonne of local rock may yield less than one kilogram—economically useless.
11. Thirty years ago, a 1–2% grade was considered good; today 0.5% can qualify.
12. Humans have used copper for millennia and explored much of the earth. Multiple new high-grade discoveries are unlikely.
13. Only 14 deposits totaling about 46 Mt were discovered in 2014–2023, versus about 225 deposits and 1.27 billion tonnes in 1990–2013. Expanding supply is difficult.
14. Falling grades mean moving more rock, digging deeper, and operating in remote areas—all of which require capital and raise costs.
15. A small mine can require at least US$1 billion; a large one often needs US$5–10 billion. Capital is locked up for years amid political and operating risks.
16. Scale, technology, and valuable by-products have kept C1 costs from rising in line with falling grades, partly hiding the higher capital barrier.
17. Technology could make lower-grade ore economic or expand recycling, raising supply.
18. The copper price needed to trigger new projects depends on government take and interest rates. Projects generally require an IRR above 15%, including funding costs during lengthy approval and construction periods.
19. The top 12 miners produce half of mined copper. Some modern mining companies have survived for centuries; imprudent operators have already disappeared.
20. Competition among miners is intense but less visible, occurring through M&A, resource access, and political relationships.
21. Miners currently prefer acquisitions and brownfield development to greenfield exploration. Capital seeks certainty, while mine development contains many uncertainties—government often being the largest.
22. Brownfields take roughly 3–7 years to generate cash; greenfields often take 10–20. Arizona’s Resolution deposit grades 1.5%, was found in 1995, taken over in 2004, and entered permitting in 2013; litigation continues and mining has not begun.
23. Chinese miners generally accept more risk than Western peers and invest more readily in politically unstable regions such as parts of Africa.
24. Governments can appropriate mining economics because subsoil resources belong to the state. Many miners now face effective tax burdens around 30–50%, with the burden rising.
25. Supply surprises are mostly negative—accidents and geopolitical disruptions. One of the few positive surprises is finding more economic ore around existing mines.
26. Western capital markets favored asset-light internet and technology companies for two decades. Since the AI boom, those companies have redirected large cash flows into hard assets.
27. Defense uses relatively little copper against the enormous base, but copper is indispensable. It is therefore exposed to both defining variables of the era: AI and geopolitics.