Deep Report · Copper

The Copper Constraint Behind the Data-Center Boom: Why Recycling and New Mines Both Matter

Data centers add a new copper demand vector to an already electrifying world. Mine supply adjusts slowly and recycling is becoming essential — but near-term refined surpluses argue against chasing the trade at any price.

The data-center buildout is usually discussed as a race for chips, power plants and grid connections. Copper sits underneath all three.

Servers and cooling systems need wiring. Substations, transformers, switchgear and busbars need conductive metal. New generation has to connect to the grid, and distribution networks have to move more electricity to sites whose loads can resemble those of small cities. The result is a second-order consequence of the data-center boom: even if computing demand remains the headline, the physical bottleneck can migrate into the materials that make electrification possible.

That does not mean every mile of new transmission is copper. Aluminium is widely used in overhead high-voltage lines because it is lighter and cheaper. But copper remains heavily used in underground and distribution cable, transformers, switchgear, substations and the equipment inside data centers. The investment chain therefore runs well beyond chips: data centers → electricity → generation and grids → copper.

Data centers are becoming a real copper demand category

S&P Global estimates global copper demand at about 28 million metric tons in 2025 and projects roughly 42 million tons by 2040. AI and data centers are not the largest part of that increase — grids, electric vehicles and broader electrification matter more — but they are a new source of incremental demand arriving at the same time.

Its study estimates direct copper demand from data centers rising from about 1.1 million metric tons in 2025 to 2.5 million tons by 2040. The associated power infrastructure needed for data centers could require roughly another 1 million tons of copper a year by 2040. Typical non-crypto data centers can use roughly 30 to 40 metric tons of copper per megawatt of capacity, with some AI-training facilities more copper-intensive.

The harder side of the equation is supply

Copper demand can move faster than mine supply. S&P Global estimates that new copper mines take about 17 years on average to move from discovery to production. The International Energy Agency reaches a similar structural conclusion: even after incorporating announced projects, its 2026 outlook projects copper supply in 2035 about 25% below primary supply requirements in its base case.

Long lead times are only part of the constraint. Ore grades have been declining, new projects require large amounts of capital, and major deposits face permitting, water, infrastructure and political risks. Existing mines can be expanded, but brownfield additions alone may not be enough if electrification demand keeps compounding.

S&P's baseline illustrates the scale of the problem. It sees today's roughly 23 million metric tons of primary mined supply slipping toward about 22 million tons by 2040 as existing assets age. Even with end-of-life recycled supply rising from roughly 4 million tons to 10 million tons, its scenario still leaves an unfilled gap of around 10 million tons unless more mine projects and expansions become viable.

Recycling is becoming part of the supply base, not an afterthought

Copper's recyclability is one reason the market can adjust. Scrap can return to the system much faster than a greenfield mine can be permitted and built. The IEA says stronger recycling is one of the critical responses to future copper bottlenecks, while also noting that the availability of end-of-life scrap itself limits how quickly secondary supply can scale.

That makes recycling economically important, not merely environmentally desirable. Higher copper prices improve incentives to collect and process scrap; better sorting and recovery can lift yields; and more material becomes available as old buildings, vehicles, electrical equipment and grids reach the end of their lives. But recycling cannot conjure metal that is still locked inside equipment that remains in service. For the next decade, the market still needs both secondary copper and new primary production.

Why this is not an automatic “buy copper” signal

The structural story is bullish. The near-term balance is more complicated.

The International Copper Study Group's October market forecast still expects refined copper supply to exceed refined use in both 2026 and 2027. That can coexist with a long-run mine constraint: smelter capacity, scrap flows, inventories and the timing of individual projects can leave refined metal relatively comfortable even while the upstream concentrate market is tight and the longer-term project pipeline looks inadequate.

Price matters too. Copper is already trading around record territory. CME's Oct. 21 technical review describes an ascending triangle in COMEX copper with repeated selling near $6.80 per pound and progressively higher pullback lows. In other words, the market is already aware that something has changed.

What would make the thesis stronger — or break it

The competing case is straightforward: investors may be capitalizing a 2030s scarcity story too early. If industrial demand softens, recycling expands rapidly and currently announced mine projects arrive on schedule, copper can correct substantially even while the long-term electrification thesis remains intact.

Market takeaway

Copper is one of the clearest examples of why second-order analysis matters. The first-order data-center story is about compute. The second is about power. The third is about the physical materials required to deliver that power.

That makes copper a credible beneficiary of the data-center capital cycle, but the better trade may depend on patience. The long-run supply problem is increasingly visible; the short-run refined market is not yet telling the same story with equal force.

Sources

• S&P Global — Copper in the Age of AI
• International Energy Agency — Global Critical Minerals Outlook 2026, executive summary
• International Energy Agency — Global Critical Minerals Outlook 2026, outlook
• International Copper Study Group — Oct. 14, 2026 copper market forecast
• CME Group — Oct. 21, 2026 Copper Futures technical review