U.S.-China Relations
NextFin News - Copper traders are bracing for a price crash toward $10,000 a ton if President Trump does not follow through on long-promised import tariffs, as the premium that built a million-ton US stockpile collapses with no levy in sight. The London Metal Exchange's three-month copper contract fell $90.75 to $14,316.50 a ton by 8:48 a.m. London time on October 19, and the spread between New York's Comex and London futures has all but vanished after peaking at $783 a ton in late August.
The metal's record run - a settlement high of $6.89 a pound on September 9 and $6.83 a pound, or $15,057 a ton, on September 22 - was never driven by a tariff. It was driven by the threat of one. Now, with the Commerce Department's June 30 report deadline passed, a July 30 decision that imposed a 50% levy on certain copper products but excluded refined copper, and a September 28 decision date that slipped by without announcement, the market is repricing that threat back toward zero. "I've never seen anything like this, where the price outlook ultimately hinges on one man's decision," said Alice Fox, a commodities strategist at Macquarie.
The Arbitrage That Built a Million-Ton Stockpile - and Why It Is Unraveling
The mechanism is straightforward, and it worked without a single dollar of public money. When the White House signaled it might impose a Section 232 tariff on refined copper, Comex futures in New York began trading at a premium to London Metal Exchange prices. That gap - the arbitrage - paid for freight, insurance, and financing to ship copper into the United States, so traders moved the metal themselves. President Trump ordered a study on potential tariffs in February 2025; since then, copper has flowed into Comex-approved warehouses, which by mid-October held roughly 70% of the world's visible copper stocks, more than 714,000 tonnes after rising nearly 750% since February 2025.
The premium was, in effect, a traded policy option. Societe Generale analysts modeled the full cost of moving LME-grade copper from European warehouses to the US East Coast and compared it with Comex futures; on August 14, the premium implied a 14.6% probability of a 15% tariff taking effect in January 2027 and a 37% probability of a 30% duty by January 2028. The Commerce Department has recommended exactly that phased approach: 15% from January 1, 2027, rising to 30% from January 1, 2028. The entire trade was a wager that a recommendation would become a regulation, and that a regulation would carry a shipping window tight enough to strain non-US supply.
"The COMEX-LME spread has increasingly become a gauge of U.S. tariff expectations, with a wider premium signaling greater perceived tariff risk and continuing to pull metal into the U.S.," said Ewa Manthey, a commodities strategist at ING.
That gauge is now reading close to zero. The spread fell from $783 a ton on August 25 to $62 a ton on September 15, and by mid-October the month-to-date average had flipped to an 85-point LME premium from a 17-cent Comex premium in June, according to calculations by John Gross of The Copper Journal. By October 14 the arbitrage stood at about $310 a ton - down sharply from the levels that made transatlantic shipments profitable. US imports, which topped 200,000 metric tons in July, the highest in 12 years, have slowed as the incentive to keep moving metal evaporates.
Natalie Scott-Gray, senior metals demand strategist at StoneX, has called the overdue Section 232 decision the "single biggest catalyst" facing the copper market, adding that while comprehensive tariffs would squeeze supply outside the U.S., no tariffs would unwind the COMEX-LME arbitrage.
The unwinding is a repricing of policy risk, not a change in the physical metal. That distinction matters because it means the crash risk is concentrated in the policy premium, not in the underlying demand for copper. But it also means the repricing can run fast: a spread built on expectations can collapse on silence. The July 30 decision illustrates how easily the premium can deflate - when the White House announced a 50% levy on certain copper products but left refined cathode exempt, the market immediately stripped the tariff risk back out of the metal that was actually excluded.
A Surplus Disguised as a Shortage - the Fundamentals Under the Bubble
Beneath the tariff trade sits a market that is long on drama and short on actual scarcity. On October 14, the International Copper Study Group raised its 2026 refined-copper surplus forecast to 104,000 tonnes and kept its 2027 forecast at 368,000 tonnes - a reversal from the 150,000-tonne deficit it had projected a year earlier. Global refined copper usage growth for 2026 was cut to 1.6% from 2.1%, with Chinese demand now expected to rise only 1.9% and consumption in the European Union and Japan remaining subdued. Production is forecast to expand just 0.4%, constrained by scarce concentrate.
Yet mine supply remains genuinely tight. The study group cut 2026 mine-production growth to 1.3% from 1.6%, and spot copper concentrate treatment charges fell to minus $272.20 a tonne on October 9 as smelters competed for feed - a negative charge means smelters are paying miners for the right to process their ore, one of the clearest signals of a raw-material squeeze. Macquarie counts a 223,000-tonne surplus this year only after allowing 531,000 tonnes for disruptions; strip out the disruption allowance and the market is much closer to balance. The point is not that copper is drowning in supply. The point is that the price spike was never about a shortage - it was about a million tonnes of metal parked in the wrong place.
That misplacement is what made the bubble look like a shortage. With 70% of visible stocks stranded in US warehouses, the rest of the world tightened: LME inventories fell for seven straight sessions to 229,650 tonnes, and cash copper settled at a record $14,822 a tonne earlier in October. So the same stockpile that signals glut in the United States signals scarcity in London. The premium did not create metal; it relocated it, and the price reflected the relocation rather than the total.
Macquarie's verdict is blunt: prices "have disconnected from the fundamentals and could arguably move higher or lower depending on the outcome of the Section 232." The bank, which in July called the rally "running ahead of reality," now sees copper averaging $13,573 a tonne this year and $12,750 in 2027, with a cycle low of $12,000 in the fourth quarter of next year. Analysts at Natixis have warned that the tariff trade has inflated a bubble in one of the world's most important industrial metals, and that the bubble could burst quickly once tariffs are ruled out.
Here is the cyclical-versus-structural call that steers the whole read. The price spike is cyclical and policy-driven: it is a premium for a tariff that does not exist, and it will mean-revert once the policy question is settled either way. Copper has lived through policy shocks before - the metal fell more than 25% in the second half of 2008 as the financial crisis crushed demand, and it took until 2021 to reclaim those highs on the back of a genuine supply-and-electrification story, not a tariff headline. What is different this time is that the spike was financed entirely by positioning rather than by consumption: end users did not bid the price up, warehouses did. The supply tightness underneath - scarce concentrate, disruption-prone mines in Chile and Peru, rising grid and data-center demand - is structural, and it is why the floor is not the 2020s average. Goldman Sachs Research expects LME copper to trade in a $10,000-$11,000 range next year, with $10,000 supported by demand from the power grid, AI infrastructure, and defense spending. The bubble is the $4,000-$5,000 of policy premium above that floor, not the floor itself.
The Binary Gamble - $10,000 or $18,000
Macquarie's scenario analysis frames the stakes as a near-symmetric binary: copper could slide toward $10,000 a ton if the tariffs fail to materialize, or climb to $18,000 a ton if they are announced with a shipping window attached. That is an $8,000 range resting on a single White House decision. The market's current pricing leans heavily toward the lower end: on mid-October prices, the December Comex premium over LME three-month implied roughly an 8% chance of a 15% duty, down from Macquarie's 10% read at the end of September and far below the 14.6% implied in mid-August.
The asymmetry, however, is not symmetric in timing. A rule-out can be priced instantly - a headline, a statement, a missed deadline - and it would release US stocks back into a tight ex-US market, which Morgan Stanley flags as the key downside risk. A tariff, by contrast, requires a decision, a Federal Register notice, and a shipping window before the metal actually moves. Silence favors the downside; only action sustains the premium. This is the second-order effect that the headline crash narrative misses: the damage is done not by a tariff decision either way, but by the delay itself, which bleeds the premium down day by day without any announcement at all.
The strongest case against the crash thesis is that the tariffs are still coming. The Commerce Department's recommendation stands: 15% from January 2027, rising to 30% from January 2028. BMO Capital Markets expects no decision before the US midterms in November, and says investors it talks to are coming around to the view that the US stockpile "could simply stay in situ for many more years." If the metal stays stranded, the argument goes, the ex-US market stays tight and the price stays elevated without any tariff at all. That is a plausible path for the floor, and it is why the structural call does not require a crash to be right.
That counter-thesis holds for the floor, not for the bubble. A stockpile that "stays in situ" is a stockpile that never reaches US consumers - which is exactly the outcome that caps the upside while leaving the downside headline risk intact. Even if tariffs eventually arrive, the premium over delivered-cost fundamentals that defined the bubble would deflate into the actual duty; the structural floor would hold, but the policy option would no longer trade at crisis pricing. And the consumers who would bear the duty are already signaling their limit: US fabricators and wiring manufacturers are expected to absorb the cost increase by compressing margins rather than passing higher copper prices on to consumers, which dulls the demand impulse that a tariff-driven rally would need. The bubble is not a bet that tariffs are possible. It is a bet that they are imminent, and imminence is what silence erodes.
What Comes Next - and the Signal That Would Break the Call
The beneficiaries of deflation are clear: LME-linked producers and non-US miners in Chile and Peru, whose metal is priced against a tight ex-US market; and consumers outside the United States, who have been paying for a scarcity created by warehousing rather than by geology. The exposed are equally clear: US fabricators and wiring manufacturers holding high-cost inventory bought at the premium; arbitrage traders long Comex and short LME whose convergence bet has already run; and US copper-related firms that would absorb tariff costs by compressing margins rather than passing them on.
The forward path splits by horizon. In the short term, volatility will track policy headlines - every statement, deadline, and leak will move the spread, and the November midterm timeline gives that volatility a calendar. In the medium term, the surplus and the stranded stockpile cap prices even without a tariff decision; BMO's "stay in situ" view is a range-bound view, not a bullish one. In the long term, the structural story - electrification, grid buildout, data centers, defense - supports a floor around the $10,000-$11,000 range that Goldman Sachs projects, well above the pre-rally decade average but far below the record.
Three scenarios frame the range. The base case: no decision before the November midterms, the spread stays narrow, and copper trades range-bound as the stockpile sits idle. The downside case: an explicit White House rule-out of refined-copper tariffs, which would release US stocks and drive prices toward Macquarie's $10,000 scenario. The upside case: a tariff announced with a defined shipping window, which would restart the scramble for non-US metal and reopen the path toward $18,000.
The falsifying signal for the deflation thesis is specific and observable: an explicit rule-out of refined-copper tariffs from the White House, or a sustained re-widening of the Comex-LME premium accompanied by Comex warehouse outflows. Either would prove that the policy option is still alive. Without one of those two prints, the market is pricing a tariff that keeps not arriving.
The copper market built a million-ton stockpile on a threat, and it is now learning that a threat deferred is a premium erased. The metal's real demand story survives the bubble; the traders who bet on imminence may not.
Prices as of 8:48 a.m. London time, October 19, 2026. Data sourced from exchange indications, the International Copper Study Group, and broker scenario analysis.
Explore more exclusive insights at nextfin.ai.
