Copper is holding near recent highs because the mine problem is no longer theoretical. Chile is still the center of the story, and its biggest producers are showing you exactly why supply is struggling to keep up with data centers, grids and electric vehicles.
Copper is not rising on an abstract energy-transition story. It is sitting in Chile, where mines that supply roughly a quarter of the world's mined copper are producing less metal just as buyers need more of it. Comex copper moved back above $6.30 a pound in July, near the elevated levels traders have been watching since the spring. The fresh production numbers gave the market a real reason to stay nervous.
Antofagasta said on July 15 that first-half copper production fell 9% from a year earlier to 285,000 tonnes, with lower output at Los Pelambres and Centinela. The company kept its 2026 guidance at 650,000 to 700,000 tonnes, but that still asks a lot from the second half. You don't need to overcomplicate this. When a major Chilean miner says output is down and the rebound depends on better grades and higher processing later, the market hears risk.
The mine problem is in the grades
BHP made the same point from a different mine. Reuters reported that the company expects fiscal 2027 copper production of 1.65 million to 1.8 million tonnes, down from 1.9528 million tonnes in fiscal 2026, mainly because of lower forecast grades at Escondida in Chile. Fastmarkets noted that Escondida's concentrator feed grade is expected to fall to around 0.70% in fiscal 2027 from 0.90% in fiscal 2026 and 1.02% in fiscal 2025.
That is the hard part. You can spend more, move more rock and improve recoveries, but thinner ore still makes every tonne harder to win. BHP said fiscal 2026 production at Escondida was helped by record material mined and record concentrator throughput, with better recoveries on top. Even with that, the grade decline is now big enough to pull next year's group guidance lower.
Chile's central bank has been blunt about the same pressure. In its June monetary policy report, the Banco Central de Chile said weaker first-quarter activity was largely tied to natural-resource sectors, with copper mining hit by lower ore grades and downtime at some sites. It also raised its copper price assumptions for 2026 through 2028, citing higher trading prices and strong global demand tied to defense spending and the energy transition, plus investment in new technologies.
UBS is taking the squeeze seriously. According to Investing.com, the bank expects a global copper market deficit of 520,000 metric tons in 2026, with uncertainty around US copper tariffs pulling material into the Comex market and affecting inventories outside the United States. That is not a neat analyst theme. It is a supply chain with fewer easy tonnes available.
AI still needs metal
The AI infrastructure story usually gets told through chips, power contracts and cooling systems. Copper sits inside all of it. Data centers need wiring and grid connections. Transformers too. EVs and renewable projects pull from the same market. If you are building anything that moves or uses electricity at scale, copper is not optional in the way a software subscription is optional.
Here is the thing: high prices are already forcing buyers to adapt. Reuters reported on June 30 that Ferrari and BMW are rolling out models with aluminum wiring, following earlier moves by Tesla and Chinese EV makers. JPMorgan expects substitution to affect about 2% of global copper demand this year, and the same Reuters report said further switching could come if copper prices stay structurally high.
Aluminum has a place. It is cheaper and lighter, which matters in cars and some air-conditioning systems. But it conducts less efficiently than copper, so engineers often need more material to do the same job, and some uses are less forgiving. Fine wiring inside dense electronics and server equipment is not the same as a heavier cable run where there is room to redesign.
That is why substitution buys time rather than solving the Chile problem. Antofagasta can approve water projects, BHP can upgrade processing and miners can squeeze more from existing pits, but the market is still dealing with assets that take years to expand and decades to replace. A new data center campus can move from board approval to construction far faster than a new copper mine can move through permitting, financing and buildout.
For now, the arithmetic is plain. Antofagasta has lower first-half output. BHP is guiding lower for fiscal 2027. UBS sees a 2026 deficit. Buyers are testing aluminum where they can. If copper prices stay high, it will not be because traders suddenly discovered electrification. It will be because the mines are telling them, in tonnes and grades, that the easy supply is not there.
Also read: Alibaba Rushes Out Qwen3.8 Two Days After Kimi K3 Rattled Chip Stocks • Alibaba's Qwen3.8 Arrives With 2.4 Trillion Parameters After Kimi K3 Selloff • Iran Kills Two US Troops in Jordan and Oil Markets Are Bracing for Worse
