How rock that is more than 99% waste becomes the metal that carries electricity through almost everything.
You are almost certainly within a few metres of some right now.
One ore, two very different roads. What the copper is bonded to — sulphur or oxygen — decides whether it meets a furnace or a vat of acid. Both end at the same pure metal plate. Tap any stage to look closer.
↓ tap any stage to explore ↓
Copper ore is about half a percent copper. Almost everything that follows exists to solve that one problem.
A typical open-pit copper mine works ore grading around 0.5% copper, and the world average grade has fallen roughly 40% since 1991. To make one tonne of copper you must crush and grind something in the order of 200 tonnes of rock. The furnace is not the hard part — the grinding is.
About 80% of the world's copper starts here — mostly chalcopyrite (CuFeS2), where copper is bonded to sulphur and iron. Found in huge, low-grade deposits in Chile and Peru. It must be concentrated before it can be smelted.
Weathered ore nearer the surface — malachite, azurite, chrysocolla — where copper is bonded to oxygen instead. It is usually processed by acid leaching rather than flotation — a completely different route. Roughly a fifth of mined copper.
Flotation needs chemicals that make copper minerals cling to bubbles. Smelting needs silica flux to carry the iron off as slag, and oxygen to burn the sulphur. Leaching needs sulphuric acid — much of it made from the smelters' own waste gas.
Which route a mine uses isn't a choice about technology. It's decided by what the copper is bonded to — sulphur, or oxygen.
Both roads end in the same product: a sheet of 99.99% pure copper cathode, about a metre square. The smelting route concentrates sulphide ore and burns the sulphur out at around 1,250°C, and handles roughly 80% of mine output. The leaching route never lights a furnace: it trickles dilute acid through heaped ore and pulls the copper out with electricity, accounting for about 20%.
Leaching is not a lesser version of smelting — it is what you do when smelting cannot pay. It suits oxide ore, which flotation handles poorly, and very low-grade rock that would never justify a concentrator. It needs little capital, no furnace, and far less water than milling.
Geography follows the geology. The heartland is Chile and the south-western United States, where arid conditions produced thick oxidised caps over sulphide deposits, and where sulphuric acid is cheap because smelters nearby make it from their own waste gas. The scale of the tilt shows in the American plant list: the United States runs 14 electrowon refineries but only two primary electrolytic refineries.
Leaching supplies about a fifth of the world's mined copper — and it is the route that ore too poor for a smelter depends on.
Then they converge. A cathode is not the finished article — it is the form copper is traded in. Both routes hand over to the same last step: melt the cathode again and cast it continuously into rod, billet or cake, which are then drawn, extruded or rolled into what manufacturers actually buy.
A single cathode can become kilometres of conductor. Rod is drawn down through successive dies into wire; billet is extruded into tube; cake is rolled into sheet and strip. Alloying happens at this stage too — add zinc for brass, tin for bronze.
Only silver conducts electricity better, and silver is far too expensive to bury in a wall.
Copper combines outstanding electrical and thermal conductivity with enough ductility to be drawn into hair-thin wire, good corrosion resistance, and a price the world can afford at 28 million tonnes a year. In the United States, copper products break down roughly as follows.
Wiring, plumbing tube, roofing and cladding. Copper demand tracks construction cycles more closely than clean-energy headlines suggest.
Grid cables, transformers, motors and circuitry. Electricity networks are one of the two largest sources of global demand, and grid investment has driven most recent growth.
Wiring looms, radiators and motors. An electric vehicle uses several times the copper of a petrol car. EVs are the fastest-growing use of all, rising from 2% of global demand in 2024 to about 10% by 2050 on IEA projections.
Industrial equipment (7%) and consumer and general products (10%): heat exchangers, valves, coinage, cookware, brass fittings.
The newest line on the demand chart: data centres. Copper carries their power and cooling, and AI server racks need more of it than conventional ones. The IEA estimates data centres could use 250,000–550,000 tonnes of copper in 2030 — about 1–2% of global demand. Estimates of AI data-centre copper intensity vary by as much as tenfold, so this is among the least settled numbers in the market.
The countries that dig copper and the country that refines it are almost entirely different places.
Copper is a highly concentrated business. Chile alone mines nearly a quarter of the world's supply; China refines close to half of it.
What the world's copper mines produced, in money. 23.2 Mt at 2025's average price of $9,945 a tonne. The United States accounted for about $11 billion of it from 1.0 Mt of mine output — 10% more value than in 2024 from 5% less metal, which is the price story of the last two years in one line.
Copper's supply chain splits geographically, and the reason is the concentrator. Because nobody ships rock that is 99% waste, the mill that upgrades ore to concentrate sits at the minesite — so the map of miners and the map of concentrators are effectively the same map. The smelter is somewhere else entirely, and concentrate is what travels between them. The Andes and the African Copperbelt do the digging; China alone refines about 48% of the world's copper, from ore it mostly does not own.
Thousand tonnes of copper content, 2025 estimates; bars are scaled within each column. The gap between the two columns is what moves by ship — and it is why control of smelting, rather than control of mines, has become the strategic question in copper. Chile mines more than three times what it refines; China refines nearly eight times what it mines. The DR Congo is the exception that proves the rule, having become the world's second-largest refiner and overtaking Chile in 2024. Japan and South Korea together refine over two million tonnes with essentially no domestic ore at all.
Reserves, million tonnes of contained copper, 2025 estimates. World reserves total about 980 Mt — roughly 42 years of production at 2025's rate. That figure is easy to misread. Reserves are an economic category, not a countdown: they mean copper that is known about and profitable to extract with today's technology at today's price. They have been rising for decades as prices rose and methods improved, and identified resources not currently economic already run to about 1.5 billion tonnes. Chile leads on both counts, mining the most and holding the most. Australia is the mirror image, holding roughly a tenth of world reserves while mining only 730,000 tonnes a year — though on the stricter JORC standard its reserves are 27 Mt rather than 100 Mt, a reminder that these figures are not measured identically everywhere.
Water. Copper is concentrated in some of the driest inhabited places on earth. The IEA found that 7% of global copper supply was at risk of disruption from floods or droughts in 2024, and expects that share to rise. Chilean operations have responded by piping desalinated seawater hundreds of kilometres inland and uphill.
Time. A greenfield project typically takes 17 years from discovery to production. Approval, financing, construction and ramp-up each add years, and none of them compress easily.
Discovery. Of 239 copper deposits found between 1990 and 2023, only 14 came in the last decade. On the IEA's base case, mined supply peaks in the late 2020s at a little over 24 Mt, then falls below 19 Mt by 2035 as reserves deplete and assets retire.
Copper's carbon problem is not in its furnace. It's in the diesel and electricity spent moving and grinding rock — which means clean power can fix most of it.
The smelter's reducing agent is the ore's own sulphur, and the waste gas is SO2, captured and sold as sulphuric acid. Roughly 80% of copper's carbon footprint sits upstream of the smelter, in mining and concentrating; only about 20% comes from turning concentrate into metal. Copper's hardest problem is not carbon at all — it is grade.
Feed dried sulphide concentrate into a flash furnace with oxygen and the reaction gives off enough heat to sustain itself with little or no added fuel.
Estimates of copper's overall footprint cluster around 3.4–4.5 tonnes of CO2e per tonne of metal, with a huge spread between mines. The best-performing tenth emit under 0.8 t; a low-grade mine on coal-fired power can emit 3–5 t.
The average grade of copper mines has fallen about 40% since 1991. Poorer ore means more rock moved and more energy spent per tonne of metal, and the effect compounds: in Latin America, the cost of expanding an existing mine has risen 65% since 2020, approaching the cost of building a new one.
There is a chemical wall behind the economic one. Conventional heap leaching works on oxide ore, but it barely touches primary sulphides like chalcopyrite — the mineral that holds most of the copper still in the ground. Those ores have to be concentrated and smelted, however poor they get.
Break that wall and low-grade sulphide ore becomes leachable — which is exactly what the new technologies are built to do.
Bacteria and catalytic chemistry that open up sulphide ore without a furnace.
Clean power and battery haulage attack the 80% of the footprint that sits upstream.
Copper loses nothing in recycling — the limit is collection, not metallurgy.
Where the recycled metal ends up is not always wire. Cathode-grade purity is unforgiving — the traces electrorefining strips out are exactly the ones that spoil a conductor — so the cleanest scrap returns to the rod mills, and the rest goes to the brass mills, where the copper is being alloyed with anywhere from 5% to 45% zinc anyway and a little residual tin or lead is tolerated, or even specified. In the United States, brass and wire-rod mills together account for about 80% of all copper recovered from scrap (USGS, 2026): brass is what makes the messier half of the scrap stream worth collecting.
The one to watch: bioleaching at industrial scale. Microbes and engineered leach chemistry can attack the sulphide ores that ordinary heap leaching cannot touch. Rio Tinto's Nuton process reached its first industrial-scale deployment at the Johnson Camp mine in Arizona; Jetti Resources and the Chilean firm Ceibo are pursuing the same goal. The significance is structural rather than incremental: it makes 99.99% cathode at the mine gate, removing the concentrator, the smelter and the refinery from the chain altogether. If it scales, it cuts emissions and turns ore currently written off as too poor into ore worth mining.
Copper's chemistry is comparatively benign, and clean electricity can take most of its emissions away. What the industry is running short of is good rock. Which is why the technologies that matter most are not those that make copper cleaner from the ore we already mine, but those that make ore we currently ignore worth mining at all.