Copper's First Supply Decline Since 2017: Real, but Narrower Than the Price Says

Copper keeps making history, and this time the reason is on the supply side. Three-month copper on the London Metal Exchange touched $14,779 a ton this month, New York futures crossed $3.74 a pound, and the metal is up roughly 51% over the past year — about four times the gain of the Magnificent Seven stocks over the same stretch. The headline everyone reaches for is scarcity: global mine output fell 1.1% in the first half of 2026, a drop that raises the prospect of the first annual decline since 2017, according to the International Copper Study Group.

That much is true, and it matters. But "supply is shrinking" is where the useful part of the story begins, not where it ends. The answer to whether this rally is built on a real, durable chokepoint or on a recoverable stumble decides what the trade is worth — and the evidence points to a shortage that is real but narrower than the price implies.

The fall was real, but look at what actually moved

The surprising detail is that the capacity to dig copper was there all along. Mine capacity grew about 3.8% in the first half, yet utilization fell from 81.0% to 77.1%. Mines produced less not because the shovels were missing but because they ran at lower rates — a loss of roughly 570,000 tonnes relative to the prior year's operating efficiency.

A large share of that is event-driven and, in principle, recoverable. Chile, the source of just under a quarter of world copper, posted its weakest second quarter in at least 19 years and cut its full-year forecast for a second straight quarter, now guiding to a 2.6% annual decline. The Democratic Republic of Congo's concentrate output fell 34% after a 2025 seismic event at Kamoa. Indonesia's Grasberg, hit by a mud rush, saw concentrate output drop 32%. Concentrate production fell 2.6% overall, offset only partly by 4.3% growth in solvent extraction-electrowinning output.

If these were the whole story, you'd call this a bottleneck — temporary capacity trouble that snaps back as operations recover. The reason not to dismiss it is the part that will not snap back quickly.

That part lives in the geology and the calendar. Ore grades keep slipping, and the mines running them are old and tired, as BlackRock's Evy Hambro put it. The discovery pipeline is thinner than the industry's own history: S&P Global found that discoveries since 2000 (687 million tonnes) have not matched the roughly 715 million tonnes found in the 1990s alone, that the cost to find a tonne of copper in the ground has risen roughly 225-fold since then, and that the average trip from discovery to production now runs about 17.5 years. A deposit found today produces in the 2040s. Permitting queues are long enough that meaningful new mine output is unlikely before 2030. Scarcity of this kind does not respond to a price spike.

Where the shortage actually bites today

Here is where the dependency picture sharpens. The clearest signal that mine output — not processing capacity — is the binding node shows up one layer down the chain, in the smelting market.

Smelters convert concentrate into refined copper, and their economics are measured by the treatment charge they earn for doing so. That benchmark settled at zero dollars per tonne for 2026, down from $21.25 in 2025. Read that plainly: Chinese smelters, which hold most of the world's newly built smelting capacity, have bid the price of concentrate up until they process it for nothing. Smelters are the replaceable, squeezed layer with idle capacity and no feed. Concentrate is the scarce, indispensable node their entire capacity depends on. When a whole processing industry works for free because the upstream material is short, that is the signature of a chokepoint, not just a tight quarter.

The consumer hasn't felt it yet

The discipline test is to notice what has not happened. None of this has reached the person buying copper wire, cable, or a transformer yet. Refined production is still growing — by roughly 0.9% this year per Morgan Stanley — because processors that can't get enough concentrate are feeding more scrap into the system, and scrap-based output rose 4.3%. Refined copper is not in deficit; Bloomberg Intelligence shows surpluses in 2025 and 2026, with the balance only turning short later in the decade.

This is the constraint-migration point. Right now the squeeze is stacked at the top of the chain, buffered by scrap and by the very smelting expansion that is being paid to work for free. It becomes a shortage that actually touches the economy only if demand keeps rising — electrification, grid buildout, and data centers, where each incremental megawatt embeds 60 to 75 tonnes of copper — while mine output stays flat into 2030.

The price may have already absorbed the structure

That leaves the question of whether the market is paying for what it is getting. The clean exposure to this chain is the mining company that sells concentrate into a scarce market, and the stocks have moved accordingly: Freeport-McMoRan is up roughly 41% this year, Southern Copper and Teck Resources are up around 45%. The leverage is real — Freeport's annual EBITDA moves by roughly $390 million for every $0.10 move in the copper price, and it now carries a market cap above $100 billion with an EV/EBITDA near 12.

The structure is real. That is the part this article is willing to bank. But structure and price are two different statements, and they have been drifting apart: copper at a record, miners up 40%-plus this year, and a refined market that is still in surplus on scrap. The discovery is no longer secret, and the valuation has started to assume the shortage is both permanent and everywhere. The asymmetry that the prospect of a first supply decline since 2017 once offered is thinner now — the honest reading is a real, concentrated squeeze at the mine layer that the market has largely decided to believe in, with the risk that the recoverable half of the decline returns before 2030 brings new tonnes.