London copper ended last week near a record. On October 9 the cash price settled at $14,689 a tonne. The three-month price was $14,570. Buyers were paying $119 more for metal now than for metal in three months. That premium had lasted fifteen sessions. A sales note from UBS, published the next day by a market blog, called copper one of the highest-conviction themes in commodities. A chart beside the note showed a hole in mine supply that lasts into the early 2030s.
Here is the idea, and it is the only one. The conviction is an unbuilt tonne. Do not buy that tonne as if it were a mine. The copper already coming out of the ground is one thing. The copper sitting in a government warehouse is a second thing. The copper that no board has committed to build is a third. The note treats them as one story. An investor should not. This is not a recommendation to buy or sell any metal, miner, or share.
What the picture actually says
The chart is not a UBS drawing. The source line reads Wood Mackenzie and Morgan Stanley Research. It runs from 2025 to 2032. Demand for mined copper rises. The band called operating mines does not rise with it. It sags. A second band, committed projects, stacks on top of the mines that already run. Together they do not catch the demand line. The space above them is the gap.
Two dashed cases try to fill that space. One counts non-committed projects at 60 percent. The hole is still there in the early 2030s. The other counts every non-committed project, at 100 percent. Even then, demand and supply only meet near the end of the window. The shortage in the forecast is not a mystery of geology. It is a pile of projects that have not been approved, financed, and built. Call them committed and the picture changes. Leave them as hopes and the picture is the one the banks are selling.
That is the whole long-term claim. Supply growth lags demand for several years. Electrification and data centers want more copper than new mines deliver. Théa Ziegler, on the UBS equity advisory sales team, wrote that on Friday. She said the lag underpins high prices even with tariffs and trade policy unresolved. She called the theme scarcity. Scarcity, on this chart, is a polite word for projects that are still a decision.
The other shortage is a warehouse
Daniel Ghali, Deutsche Bank's head of metals research, is talking about a different pile. Last month he said inventories available to the rest of the world had fallen to unprecedented lows. The reason he gave was not a missing mine. It was stockpiling. China has been putting copper into a strategic reserve for decades. The threat of American import tariffs is pulling more metal into U.S. warehouses. He has estimated that the two countries could be sitting on about 71 percent of global inventories by year-end. On one telling, China holds on the order of two million tonnes in reserve, and U.S. warehouses are headed toward roughly 1.3 million tonnes.
His price is a date, and it has to stay his. He sees London copper at $10 a pound, which is $22,050 a tonne, in the second quarter of 2027. From prices near $14,600 when he spoke, that is about a 50 percent rise. He also said the extreme case, metal unavailable to other users by the end of 2028, probably does not happen. High prices would push some buyers into aluminum. Aluminum is a worse conductor and a cheaper one. He does not think today's price is high enough to force that switch. And if Washington never imposes the tariff, the metal may still stay in the United States. U.S. copper futures trade at a premium to London. The exchange has warehouses on American soil. Metal can move from one American shed to another and never reach a factory in Europe or Asia.
Hold the two shortages apart. The Wood Mackenzie picture is tonnes that mines have not promised. Ghali's picture is tonnes that mines already made, which two governments are keeping off the open market. One can be right and the other wrong. A warehouse can empty. A project that was never financed cannot pour. A target of $22,050 is a bet that the sheds stay full and the buyers do not switch. It is not a bet that a feasibility study gets built.
The strike is a third clock
Natalia Corfield, who runs Latin America corporate credit research at JPMorgan, pointed at Chile this week. A strike is on at Centinela, a large part of Antofagasta's copper. The two unions, with about 709 members, walked out after mediated talks failed. They are about 22 percent of the mine's direct workforce. They want the same benefits no matter which union a worker belongs to. The company had offered a 2 percent real raise and a bonus of about 22.5 million Chilean pesos, roughly $23,000. The unions rejected it. A strike vote passed with about 99 percent support.
The unions say the plants can take ore for a while. Mining, haulage, and development are already disrupted. They expect the cut to show after about two weeks, and they say November output could fall by as much as half. Bloomberg had carried a version of that warning. Antofagasta does not agree. It says it will keep talking, and that its production outlook is unchanged. Group guidance for 2026 is still 625,000 to 655,000 tonnes. Centinela made 240,400 tonnes in 2025 and 97,100 tonnes in the first half of 2026, about a third of the company.
A walkout at one Chilean pit can move a price that is already tight. It cannot fill, or create, the gap on a chart that runs to 2032. If the strike ends next week, the 2030s look the same. If it cuts November in half, the world is short one mine's tonnes for one month. That is a trading fact. It is not the structural fact the sales note is using. Investors who buy "scarcity" because a union posted a deadline have bought the wrong clock.
The four names are not the gap
Ziegler closed with a list. The team prefers copper-linked miners with real leverage to the theme. The names were Freeport-McMoRan, First Quantum, Hudbay, and Teck. She called them well placed for a multi-year upcycle driven by structural demand and a constrained supply response. Those are her words. They are not a research report you can audit from a blog, and they are not a reason to own the shares.
Read the list against the chart, not against the adjective. These are companies that already mine copper, or are spending money on copper they have already decided to build. They live in the bottom bands. Operating mines. Committed projects. If the price stays high because other projects never get built, a producer that is already pouring gets the margin. That is the only logic in the list that matches the picture. It is leverage to a price. It is not a claim that any of the four will close the gap by themselves.
The trap is the jump from that list to everything else with copper in the name. The white space on the chart is non-committed projects. A junior with a study, a permit fight, and a hope of a buyer is not "the same theme." It is the tonne the forecast is assuming might not arrive. Paying a scarcity multiple for the project that scarcity depends on failing is how a careful chart becomes a bad portfolio. UBS did not tell you to do that. The headline makes it easy.
What you can underwrite
You can underwrite the split. Demand on the chart rises through 2032. Operating mines do not. Committed projects do not catch the line. The rest is a decision that has not been made. You can underwrite the warehouse as a separate fact. Visible stocks are tight. London is in backwardation, which means prompt metal is worth more than later metal. Ghali's explanation is that the United States and China are holding a large share of what exists. You can underwrite his target only as his target. $22,050 in the second quarter of 2027 is a forecast with an escape hatch he already named. Buyers can switch to aluminum. Stockpiles can come back.
You can underwrite the strike as small and dated. One mine. A few hundred workers. A company that has not cut guidance. A union that says November is the month to watch. Useful for a trader with a week. Useless as proof of a decade.
You cannot underwrite the four names from this page. A sales team's preferred list is a sentence in a note. It does not price a balance sheet, a permit, a grade, or a country. Freeport, First Quantum, Hudbay, and Teck each have their own risks, and those risks are not in the paragraph that named them. You cannot underwrite a tariff as either the whole bull case or a thing to ignore. Ziegler said look through near-term trade policy. Ghali said the tariff threat is why metal is piling up in America. Both can be true. A tariff can lift the price of trapped metal this year and still be noise beside a mine that takes fifteen years. Do not use one to explain the other.
What would make this reading wrong
The reading is wrong if the non-committed projects get built. The 100 percent case on the chart nearly closes the gap by 2032. If boards approve them, banks fund them, and the tonnes arrive, the scarcity in the picture was a delay, not a fate. Watch commitments, not adjectives. A final investment decision is a fact. A dashed line is a guess.
The reading is wrong if the warehouses open. Ghali's $22,050 needs the stockpile to keep crowding out other buyers, and it needs the price to stay below the level that sends those buyers to aluminum. If London stocks rebuild and the backwardation dies, the prompt squeeze was a location problem, not a missing mine. The mine gap can still be real in 2031. It will not pay you in 2027 for a reason that already reversed.
The reading is wrong if you bought the gap. The unbuilt tonne is the reason a producer might earn a high price. It is not itself a producer. A company that still has to raise the money, win the permit, and survive the build is the supply response the forecast is doubting. Treat it as the doubt. Do not treat it as the theme.
The idea, once
UBS called copper a high-conviction theme and named four miners. The chart next to that call, drawn from Wood Mackenzie and Morgan Stanley, says operating mines and committed projects fall short of demand into the early 2030s. The tonnes that would close the hole are not committed. Deutsche Bank's $22,050 target is a different claim. It is about metal already mined and then stored in the United States and China. A strike at Centinela is a third claim, measured in weeks, and the company says its guidance stands.
Three clocks. One rule. The conviction is the unbuilt tonne. Do not buy it as if it were a mine.
A note on sources and limits
Théa Ziegler's comments and the four names are from her Friday note, as carried on October 10, 2026. The supply chart is sourced on its face to Wood Mackenzie and Morgan Stanley Research, not to UBS. Daniel Ghali's inventory comments and the $22,050 target were reported by The Wall Street Journal in late September 2026 and repeated in later notes. The October 9 London prices, the $119 backwardation, and warehouse stocks near 233,000 tonnes are from LME-based reports that day. Centinela's 2025 output, 2026 half-year output, the walkout, the company's guidance, and the union's November claim are from Reuters and company statements in late September and early October 2026. Nothing here is investment advice or a solicitation. Prices, stocks, and strike status change. Readers should read the primary notes and filings and should speak with a licensed adviser before any decision.