Three month copper on the London Metal Exchange reached $14,728 a tonne on September 8, an all time high, with the American futures contract setting its own record near $6.74 a pound. The metal is up close to 18% this year.

The demand story is the one everybody tells, and it is real. The harder half is on the other side of the ledger: global mine production actually fell 1.1% in the first half of this year, and output of concentrate fell 2.6%.

A price signal that cannot summon supply

In most commodity markets a record price is self correcting. High prices bring on new production, the surplus arrives, and the price falls. Copper is the clearest case of a market where that mechanism barely works.

The reasons are physical and they compound. Average ore grades have fallen about 40% since 1991, so the same tonnage of rock yields considerably less metal and requires more energy to process. The capital cost of expanding a project has risen about 65% since 2020. And the average time from discovering a deposit to producing from it is around seventeen years.

Seventeen years means the mines that would respond to today's price were decided on in the late 2000s. Nothing a buyer does in 2026 changes the supply available before the 2040s, which is why analysts are forecasting deficits rather than surpluses: the International Energy Agency projects a 25% shortfall against demand by 2035 on currently announced projects, and S&P Global has put a potential gap near 10 million tonnes a year by 2040.

Individual producers illustrate the point. Codelco, the Chilean state company and the largest producer in the world, mined 564,000 tonnes in the first half, down 11% on a year earlier after operational restrictions at one of its principal mines.

What is pulling on the other end

Whitmore's Watchlist:
> FCX (Freeport-McMoRan Inc.): The largest listed pure play on the copper price.
> SCCO (Southern Copper Corporation): Low cost Latin American production with long reserve life.
> COPX (Global X Copper Miners ETF): The basket, for exposure without single mine risk.

Electrification is the structural demand, and artificial intelligence has recently become the most visible part of it. The International Energy Agency expects data centre electricity use to reach roughly 945 terawatt hours by 2030, about double the 2024 level, and every one of those watts arrives through copper.

BHP has put a figure on it that makes the scale legible: each additional $200 billion a year of data centre investment requires roughly one new mine producing 150,000 tonnes annually. There are not many such mines, and as above, there is no way to build one quickly.

Clean energy demand on its own is projected to rise from 7.7 million tonnes in 2024 to 10.9 million by 2030 under the agency's central scenario, with total demand potentially reaching 34.1 million tonnes by 2040.

The policy layer that is distorting the price right now

A 2025 presidential proclamation directed an assessment of whether to impose a 15% tariff on refined copper imports from January 2027, rising to 30% in 2028. Traders have responded by moving metal into the United States ahead of any decision.

That has two effects worth separating. American inventories have risen, which looks like comfort. Everywhere else inventories have thinned, which is the opposite. A tariff has not yet been imposed and the price is already being shaped by the possibility of one, which is a reminder that in a tight physical market the anticipation of a policy moves metal before the policy exists.

For an investor the practical distinction is between owning the metal and owning the miners. The metal expresses the deficit directly. A producer adds jurisdiction risk, operating risk and cost inflation, and this year has shown that even the largest can miss its own production target by eleven percent. In a market where supply cannot respond, the companies that already own producing assets have an unusual kind of scarcity value, and that is the part the price has been recognising.

Whitmore's Take: Copper is the rare commodity where a record price buys you almost no new supply for a decade. Worth deciding whether you want the metal or the miners, because only one of them can disappoint you operationally.

Written by Daniel Whitmore
Millionaire Insiders