Copper's Squeeze Broke but the Price Held: Antofagasta's Read

On Monday 17 August the London Metal Exchange's cash-to-three-month copper backwardation reached close to $550 a tonne, its widest in more than five years, and closed the session at $436. By Wednesday it had collapsed to under $250. In between, roughly 63,000 tonnes of on-warrant copper appeared in LME warehouses in three days — an increase of more than 50% — after a spread that paid over $400 a tonne between the August and September contracts made it worth someone's while to deliver.

That is the industry condition that changed this month, and it is not the one the headlines described. The squeeze broke. What did not break was the price: copper was still trading around $6.56-$6.59 a pound on 27 August, within touching distance of the all-time Comex high of $6.77 set on 7 August. A financial squeeze resolved in 48 hours once the incentive was large enough. The underlying scarcity did not resolve at all. Separating those two things is the whole exercise for anyone holding listed copper equities — and Antofagasta, the only pure-play copper producer in the FTSE 100, is where the distinction is cleanest.

What actually happened in the physical market

The August tightness was real but specific. LME stocks had been drained hard: US refined copper imports exceeded 200,000 tonnes in July 2026, the highest monthly total in twelve years, as metal was pulled towards American warehouses ahead of tariff decisions. That left LME inventories down around 14%, with stocks of roughly 214,550 tonnes on 11 August and cash copper settling at $14,424.50 a tonne, $207.50 above the three-month contract.

Over the following week the backwardation more than doubled again, peaking near $550. A spread that wide is not a forecast about 2027 demand. It is the market paying a very large premium for metal today versus metal in three months — a statement about deliverable inventory in a particular set of warehouses, not about the annual supply balance.

And it behaved accordingly. Market sources indicated one large trading company was behind the bulk of the deliveries, with some metal reportedly sourced from Chinese participants. Once 63,000 tonnes arrived, the spread fell by more than half within two sessions. That is the textbook resolution of a positioning squeeze: the tightness was in location, not in tonnes.

The price is the part that should hold attention. If the entire August move had been a warehouse game, copper would have given back most of it when the warehouse game ended. It gave back very little. That is the signal worth trading around: the flat price is being held up by something other than the spread.

What it means for Antofagasta

Antofagasta is the most direct listed expression of that flat price on the London market — a Chilean copper producer with no iron ore, no energy division and no diversified book to dilute the read. It is also, inconveniently and instructively, the clearest evidence that the sector cannot simply supply into a record price.

On 13 August the company cut its 2026 production guidance to 625,000-655,000 tonnes from 650,000-700,000 tonnes, a reduction of about 5% at the midpoint. The cause was weather: extreme rain and snowfall in July forced an orderly shutdown at Los Pelambres, the group's flagship high-altitude operation, after Chile's government declared a state of catastrophe in the Coquimbo Region. Management described it as the most severe weather event in the mine's history. Mining and processing have resumed, though inspections identified repairs needed to pipeline platforms and water management systems. The shares fell around 5% on the day.

The half-year numbers underneath that cut were strong. For the six months to 30 June 2026, Antofagasta reported:

• Revenue of $4.48bn, up 18% year on year
• EBITDA of $2.84bn, up 27%, with the margin expanding five percentage points to 63.4%
• Profit before tax of $2.0bn, up 72%
• Operating cash flow of $2.77bn, up 53%
• Group copper production of 285,000 tonnes, down 9%
• Net cash costs of $1.22 a pound, down 8% on stronger by-product credits and cost control

Read those two lines together — EBITDA up 27%, production down 9% — and the mechanism is plain. Antofagasta is capturing an extraordinary price on materially fewer tonnes. A 63.4% EBITDA margin on falling volume is what a price-driven earnings year looks like, and it is a genuinely powerful position while the price holds. It is also entirely dependent on the price holding, because there is no volume growth cushioning it in 2026.

This is why the backwardation collapse matters more to Antofagasta's shareholders than to a trader. The company does not earn the spread. It earns the flat price, on tonnes it has just told the market will be fewer than planned. The spread widening in mid-August never flowed to its P&L; the spread collapsing does not take anything out of it either. What would take something out is the flat price following the spread down — and so far it has not.

The Openbook read

Profitability — exceptionally strong, and price-led. A 63.4% EBITDA margin, five points of expansion, and net cash costs down to $1.22 a pound. The cost side is doing real work here: an 8% reduction in unit cash costs during a period of falling volume is unusual, since fixed costs normally spread worse over fewer tonnes. By-product credits from gold and molybdenum are carrying part of that. This is the strongest of the five factors and the least contentious.

Growth — negative on volume, positive on the funded pipeline. Production fell 9% in the half and full-year guidance has been cut. There is no 2026 volume growth story. The 2027 story is different and it is already funded: the Centinela second concentrator, a $4.4bn project revised up from a $3.7bn estimate in 2022, with $2.5bn of project financing secured and first production expected in 2027. It adds roughly 170,000 tonnes of copper-equivalent output, of which about 144,000 tonnes is copper — meaningful against a 625,000-655,000 tonne base. Growth scored on the reported half looks poor; scored on the funded pipeline into a tight market, it looks considerably better. The gap between those two readings is the single most important thing about this company right now.

Solvency — the factor that is quietly doing the most work. Operating cash flow of $2.77bn in a half year against a $4.4bn multi-year project with $2.5bn already financed means Antofagasta is building new supply out of current cash flow rather than out of hope. In a sector where the standard complaint is that nobody is funding the next mine, being able to fund it during the price spike — rather than after it — is a distinguishing characteristic, not a footnote.

Momentum — strong but weather-interrupted. The shares have tracked the copper rally through the summer, then took a roughly 5% hit on the guidance cut in mid-August. Momentum here is essentially a levered copper-price signal, and the July weather event was exactly the kind of idiosyncratic shock that interrupts it.

Reward/Risk — the balance the squeeze story obscures. The reward case is straightforward: near-record realised prices, 63% margins, a funded expansion landing into a market that may still be short. The risk case has three distinct legs and they are not correlated. First, forecast risk: the copper deficit consensus is far from settled. Jefferies models a 442,000-tonne global deficit in 2026, widening to 782,000 tonnes by 2030, on demand of 28.184mt against supply of 27.742mt — but the International Copper Study Group flipped in April 2026 to forecasting a 96,000-tonne surplus for 2026, reversing its own previous 150,000-tonne deficit call, and sees a 377,000-tonne surplus in 2027. Those two houses are not describing the same market. Second, operational risk: Los Pelambres is high-altitude and, as July demonstrated, weather-exposed in a way that translates directly into guidance. Third, single-commodity, single-jurisdiction concentration — the same purity that makes Antofagasta the cleanest read on copper makes it the least hedged when copper or Chile disappoints.

Notably, Jefferies' own 2026 average price forecast of $13,380 a tonne sits below where copper has traded through August — even the constructive side of the sell side is not modelling current prices as the run rate.

Which other listed names sit in this flow

The tightness splits the London-listed copper complex into three groups, and a single "copper is at a record" headline hides the differences entirely.

Producer-traders capture what producers cannot. Glencore is the only major LSE name that monetises the backwardation itself. Its marketing division delivered roughly $3.3bn of adjusted EBIT in the first half alone — already near the top of its $2.3bn-$3.5bn long-term annual guidance range — with full-year marketing EBIT guided to $4.7bn-$5.0bn. Alongside that, mined copper output rose 15% to 397,000 tonnes from 343,900 tonnes, with full-year guidance maintained at 810,000-870,000 tonnes. Physical dislocation of the kind seen in August pays Glencore twice: once through the mine and once through the trading book. It is the structurally different business in this group, not merely a larger version of the others.

Diversified majors dilute the signal. Rio Tinto guides to 800,000-870,000 tonnes of copper for 2026 — volume comparable to Glencore's — and cut its 2026 C1 unit cost guidance to 30-50 US cents a pound from 65-75 cents, helped by higher gold prices and productivity gains. That is a striking cost position. But first-half copper production rose only 1%, and copper sits inside a group whose earnings are still dominated by iron ore. Anglo American produced 344,000 tonnes in the first half against unchanged full-year guidance of 700,000-760,000 tonnes, split between Chile at 390,000-420,000 tonnes and Peru at 310,000-340,000, and lowered unit cost guidance in both — Chile to around 210 cents a pound from 230, Peru to around 65 cents from 100. Copper is doing the heavy lifting in Anglo's earnings mix, but the read-through to the copper price is still filtered through other divisions.

The distinction that matters is not who owns the most copper. It is who earns the spot price on tonnes they can actually ship this year, who earns the spread, and who is funding the tonnes that arrive after the spread has gone. Those are three different exposures wearing the same sector label, and they can be compared side by side on our screener.

What to watch next

LME on-warrant stocks and the cash-to-three-month spread. The single cleanest ongoing test. If inventories keep building and the backwardation stays compressed while the flat price holds near records, the physical deficit thesis is intact and August was a positioning event. If the flat price follows the spread down, August was the whole story and the price was the squeeze.

Antofagasta's Q3 production report, due in October. The company has published third-quarter production in the second half of October in prior years. This is the first hard read on whether Los Pelambres has recovered to a rate consistent with the revised 625,000-655,000 tonne guidance, and whether the identified pipeline and water-system repairs have cost more time than flagged.

Centinela commissioning milestones into 2027. With first production expected next year, any schedule or capital commentary carries outsized weight — this is the volume growth that the 2026 numbers do not contain.

The 2027 balance forecasts. Jefferies and the ICSG currently disagree by roughly 700,000 tonnes on where the market lands. Revisions from either, particularly around Chinese demand and secondary supply, will move the whole complex before they move any individual company's results.