The Copper Tariff the Market Is Pricing at a Fifth of Its Face Value

The copper tape in early July 2026 reads like a contradiction. On 30 June, Washington cleared the last procedural gate on a plan to tax imported refined copperrefined copperCopper purified to about 99.9% metal and cast into shapes like cathodes, ready to be drawn into wire or tube. It is the traded, benchmark form of the metal.: a phased Section 232Section 232A US trade law that lets the president impose tariffs on imports deemed a national-security risk. It is the legal basis for the copper, steel and aluminum duties. duty of 15% from 1 January 2027, rising to 30% in 2028. Section 232 is the US trade law that lets a president tarifftariffA tax a government charges on imported goods, paid by the importer when the goods cross the border. imports on national-security grounds; refined copper is the purified, benchmark form of the metal that gets drawn into wire and tube. At the same moment, US warehouse stocks of that copper stood at a record near 652,200 tons, and Goldman Sachs held an end-2026 forecast of $13,735 a ton with a scenario above $14,000. That Goldman figure is a global (London-based) price target, not a New York premium; keep the two apart, because the whole story lives in the gap between them.

Record supply and a rising price target, side by side. A reader scanning headlines could be forgiven for thinking one of them must be wrong: gluts are supposed to push prices down, not sit under a bullish forecast. But there is no contradiction here, and untangling why answers a sharper question the market is quietly posing. If a 15% tariff is really coming, why are US buyers paying New York only about $400 a ton more than London, when 15% of a $13,335 copper price works out to roughly $2,000 a ton?

That $400 is about a fifth of the tariff's face value. The market, in other words, is behaving as though only a sliver of the duty is real. This article works through why, and whether it means the tariff will collect far less than advertised in its opening quarter. The short answer is that the record stockpile and the muted premium are not two facts but one: the same pile of duty-free metalduty-free metalCopper imported before the tariff takes effect, so no import tax was ever charged on it. It can be sold later without the duty. that looks like a glut is exactly what lets buyers ignore the tax for months.

Start with the mechanical benchmark. A tariff is a tax the importer pays at the border, so in the simplest case it should show up dollar-for-dollar in the US price. With copper trading near $13,335 a ton on the London Metal Exchange on 6 July 2026, a 15% duty is worth about $2,000 a ton. If US buyers had to import taxable metal to meet demand, you would expect New York copper to trade roughly $2,000 above London. Economists call that full pass-throughpass-throughHow much of a tax ends up added to the price buyers actually pay. Full pass-through means the entire tariff shows up in the price; partial means only some of it does.: the entire tax lands in the price.

The gap that actually exists, the COMEXCOMEXThe New York metals futures exchange (part of CME Group) where US copper is priced. Metal delivered against COMEX contracts sits in its US warehouses.-over-LMELMEThe London Metal Exchange, the main global benchmark for copper priced outside the United States. premium, is the extra amount copper costs in New York versus London. COMEX is the New York futures exchange; the LME is the London benchmark used for the rest of the world. In late June and early July 2026 that gap was running near $400 a ton, about 3% of the copper price, according to ING. So the market is charging about a fifth of what the tariff is mechanically worth. The question is not whether the market is being irrational. It is: what does a rational buyer know that makes $400, not $2,000, the right number today?

A 15% duty is worth about $2,000 a ton. The market is charging about $400. The difference is not doubt about arithmetic; it is a warehouse full of metal that never has to pay it.

The answer sits in COMEX warehouses. Over roughly sixteen months, US inventories of refined copper multiplied from about 80,000 tons in February 2025 to a record near 652,200 tons by late June 2026 (652,197 tons on 22 June). Over the same stretch, London Metal Exchange stocks fell toward a multi-month low, near 352,100 tons. That pattern is the tell: the metal was not newly mined, it was moved. Traders pulled copper toward the United States ahead of any tax, a maneuver called front-runningfront-runningBuying and importing goods early, before a tariff starts, so the metal already sitting in the country avoids the future tax., so that it would already be sitting inside the border, duty-free, when the tariff arrives.

Now size the pile against demand. The United States uses roughly 1.8 million tons of refined copper a year, about 150,000 tons a month, per USGS figures. Divide 652,200 tons by that monthly draw and you get close to four months of cover, all of it already imported and untaxed. And that is the conservative reading: measured against imports alone, which are the only tons the tariff can actually tax, the buffer is longer, because US mines and scrap supply a large share of demand. A big chunk of what America consumes never crosses a border at all.

There is a subtlety that separates a lazy reading from a rigorous one. The tariff is not live yet, it starts on 1 January 2027, so metal changing hands today owes no duty regardless. A low near-dated (spot) premium is therefore partly mechanical and does not, by itself, prove the market disbelieves the tariff. The number that carries information is the forward premiumforward premiumThe same New-York-over-London gap but for a delivery date in the future, such as 2027. This is the number that should reflect a tariff that starts on 1 January 2027, so it is the meaningful one to test.: the New-York-over-London gap for delivery in 2027, after the tax is supposed to bite. Think of an airline ticket for next January that already prices in a fuel tax starting then, that forward price is where you read the market's real expectation.

Here is where the buffer bites twice. Even the 2027-dated forward premium can stay capped, because a seller can settle a future contract by delivering physical metal from a warehouse, and if that metal is already duty-free, the forward price cannot climb far above today's. Cheap, stored, untaxed copper can always be handed over instead. So the buffer explains not only why spot copper is calm, but why the forward curve can stay well below 15% until the duty-free pile actually thins. That is the single most important thing to watch.

This is not the market's first pass at pricing a copper tariff, and last year's episode is the template. In July 2025, the administration floated a 50% tariff on copper, and the New York premium exploded, reaching on the order of a quarter of the copper price at its peak (estimates run between about 25% and 28%, or roughly $2,600 to $2,900 a ton). Then, on 30 July 2025, the proclamationproclamationA formal presidential order that carries the force of law, used here to set the copper tariff schedule. landed and refined copper was carved out: the 50% duty hit semi-finished products (rod, tube, sheet, wire) and copper-intensive goods, but cathodes, anodes and scrap were exempt. The premium collapsed almost to zero, and COMEX copper futures posted their steepest one-day drop on record.

The lesson is that the premium round-tripped on legal text, not on any physical shortage. Not a single ton was consumed or destroyed; a clause in a proclamation moved the whole basisbasisThe price difference between the same commodity in two markets or contracts; the COMEX-LME premium is a basis between New York and London copper. up and back down. That is the discipline to carry into 2026-27: a copper premium driven by policy is a bet on words on a page, and it can reverse the moment the words change. It is why part of today's muted $400 reading may be simple doubt that the duty is ever signed or survives a court challenge, a separate force from the inventory buffer, and one this piece will insist on distinguishing.

Does a low premium today mean the tariff never really lands in the price? The nearest precedent says no, but it also says do not expect a clean jump on day one. Aluminum has lived under Section 232 since 2018, and its Midwest premiumMidwest premiumThe extra amount US aluminum buyers pay over the world price, largely reflecting the US aluminum tariff. It is the closest historical analogy to the copper premium., the extra US aluminum buyers pay over the world price, is the closest cousin to the copper premium. When a 10% aluminum tariff took effect in 2018, the Tax Foundation found the premium rose by more than the duty itself, an over-100% pass-through. But the full move took years, not days: by February 2026, after tariffs had been ratcheted to 50%, the aluminum Midwest premium hit a record $2,182 a ton.

So the honest reading of a $400 copper premium is not 'the tariff is fake.' It is 'convergence is deferred.' A regional premium under Section 232 tends to reach, and eventually exceed, the size of the tariff, but as a multi-year grind, gated by how fast the duty-free buffer empties and how quickly new taxable metal has to be imported. For copper, the 652,200-ton pile is the throttle on that timeline.

Pull the threads together and there are three distinct reasons the premium could be stuck near $400, and they are worth separating because they predict different futures. First, the buffer thesis: timing, not doubt, holds the premium down, and the forward premium should rise toward 15% as the pile empties. Second, tariff-doubt: the market simply is not sure the duty will be signed or will survive court, so the premium should react to legal and political headlines more than to inventory. Third, term-structure discounting: the forward premium stays capped for the mechanical reason that duty-free metal is deliverable against 2027 contracts, and stays that way until stored metal thins.

The three make different bets, and the next two quarters will separate them. If premium and inventory move in opposite directions, as the pile drains and the premium climbs, the buffer thesis wins. If the premium jumps on a court docket or a signature while warehouses stay full, tariff-doubt was doing the work. If the forward curve stays well below 15% even as the front-month pile drains, the mechanical term-structure story dominates. These are not mutually exclusive, but their relative weight is testable, and that is the frame to hold through the first half of 2027.

The most exposed party is not the manufacturer but whoever financed the 652,200-ton pile. Holding close to four months of copper in a warehouse ties up an enormous amount of working capitalworking capitalThe cash a company ties up to hold inventory and fund day-to-day operations. Storing months of copper duty-free locks up a large amount of it as a bet the tariff arrives., the cash a company locks into inventory, and that spend only pays off if the tariff actually arrives to lift the value of the untaxed metal. If the presidential determinationpresidential determinationThe formal decision a president signs to turn a recommended tariff into a legal one. Until it is signed, the rate is proposed, not final. never comes, the duty-free bet sours, and the same traders who front-ran the tariff could unwind, which would drag the premium back toward zero exactly as it did in July 2025.

For the buyers who ultimately matter, the US grid build-out, data centers, electric-vehicle makers and construction, the buffer decides timing, not existence. Copper is called 'Dr. CopperDr. CopperA nickname for copper as an economic bellwether: because it is used everywhere (wiring, motors, construction), its price is read as a gauge of global demand. A tariff distorts that signal in the US.' because its price reads like a diagnosis of the economy; a tariff distorts that reading inside the US. If the duty holds, the cost shock is real but deferred: it lands when the duty-free pile thins and buyers must import taxable metal, plausibly deeper into 2027 rather than on 1 January. If the duty is watered down or struck down, the shock may never fully land. Either way, the tell will be in the tenor of the premium, not the headline rate. Watch the 2027-dated forward against the inventory line. That single chart, not the face value of the tariff, is where the answer will show up first.