Copper Prices Near Records, but Which Signals Can You Trust?

Copper hit an all-time high of US$14,875/t on 10 September 2026, then dropped roughly 3% in a single session when White House tariff headlines shifted. That whipsaw is the story of the whole quarter compressed into a few hours.

The metal is closing out its third straight monthly gain, settling around US$14,416-14,447/t at the end of September, up approximately 0.9% for the month and 7.9% for the quarter. What makes this rally genuinely contested is the split between two sets of signals.

On one side sit the physical readings: Shanghai Futures Exchange inventories down roughly 70% since June, China’s manufacturing activity back in expansion. On the other sit the speculative distortions: tariff-driven logistics squeezes and reversals sharp enough to erase a record high in an afternoon.

The copper price sitting near record territory forces a specific analytical question. This piece works through which of those signals deserve trust and which risks deserve weighting before you conclude that Q3 momentum carries into Q4.

What the inventory numbers are actually telling you

Start with the drawdown, because it is the cleanest real-time signal the market produces. ShFE copper cathode stocks fell to 38,744 tonnes by the end of September, a 17.8% week-over-week drop and the lowest reading since January 2024.

That figure did not appear out of nowhere. It followed 43,900 tonnes on 23 September (itself the lowest since 2023) and roughly 63,000 tonnes in early September. The direction of travel is one way, and it has been for months.

SMM copper inventory analysis published on 29 September confirms the same directional signal, showing ShFE stocks declining from 359,100 mt at end-March to 47,100 mt by 24 September, a drawdown that corroborates the fabricator-driven consumption interpretation rather than a simple positioning shift.

The full picture: ShFE inventories are down approximately 70% since early June 2026.

Here is the part that changes the interpretation. That metal is not disappearing into a warehouse somewhere. Incoming imported cargoes are moving straight to fabricators rather than entering storage, which means the drawdown is a consumption signal, not a positioning one.

The tightness is not confined to China either. LME copper stocks stood at approximately 254,250 tonnes as of 22 September, but a large share was already marked for withdrawal, meaning the freely available figure is considerably thinner than the headline suggests.

The LME warehouse dynamics behind that 254,250-tonne headline figure carry significant nuance: when a large share of registered stock is already cancelled for withdrawal, the freely available tonnage that can clear genuine end-user demand shrinks to a fraction of what the exchange reports.

Put the two exchanges together and the distinction matters for you. When real end-users are pulling metal directly rather than traders rebuilding stock, the price floor rests on demand you can point to, not on financial flows that can reverse on a headline. That is the difference between a genuine floor and a trap.

Three forces that drove copper’s 7.9% quarterly gain

Not every driver behind copper’s 7.9% quarterly gain carries the same weight. Ranking them by durability is the analytical work, because it tells you what survives if conditions shift.

1. Tariff-driven logistics distortion (most transient). Expectations of US refined-copper tariffs pulled metal toward America, draining availability everywhere else and amplifying price moves well beyond what physical demand alone would justify. The fragility showed instantly: when Reuters reported the White House was hesitating on the tariffs, prices fell roughly 3% from the record to around US$14,330/t.

US copper tariff mechanics, specifically the Section 232 investigation structure and the front-loading behaviour it incentivises, explain why metal flows toward American destinations even before duties are formally imposed, producing the global tightness the Q3 data reflects.

1. China’s manufacturing recovery (medium-term). The official manufacturing PMI climbed to 50.1 in September from 49.8 in August, crossing back above the 50-point line that separates contraction from expansion. That macro turn lines up directly with the fabricator drawdown from the previous section, giving the demand story a second, independent confirmation.
2. Structural electrification and AI demand (most durable). Grid upgrades, renewables, and data-centre build-outs form a demand floor that operates regardless of tariff politics.

The hierarchy carries a specific implication. If the tariff dynamics resolve or reverse, the price does not automatically collapse, because the structural layer underneath holds something up. The open question is how high that floor sits relative to a spot price near US$14,400/t.

The long-duration demand case: grids, EVs, and AI infrastructure

Three pillars support the structural floor. Grid upgrade investment is the first, as ageing networks require heavy copper reconductoring. Renewable energy infrastructure is the second, with solar and wind installations consuming multiples of the copper that fossil-fuel generation requires. Data-centre and AI build-outs are the third, driving a new wave of power and cooling demand that barely registered five years ago.

ING commodities strategist Ewa Manthey has noted that structural demand trends combined with limited supply expansion should sustain price levels through Q4. Critically, this demand is relatively insensitive to China’s property market, which remains a drag on construction-related copper use.

For context on the gap between structural conviction and consensus, ING’s December 2025 forecast put the 2026 average at US$11,500/t, a figure Q3 actuals have already overtaken by a wide margin.

The case against the rally: where bears see cracks

The sceptical case is not a fringe view worth waving away. It rests on the same data the bulls cite, read differently, and it deserves genuine weight.

Bloodstone Research framed the 10 September record bluntly: “the record has no buyer.”

The empirical backing for that line is the speed of the reversal. The move from US$14,875/t down to roughly US$14,312-14,330/t after tariff hesitation reports, a 3.1% drop, looks less like a market pricing in fundamentals and more like a squeeze unwinding once the pressure came off. LME three-month copper then closed its first losing week since June, around US$14,233/t.

The three pillars of the bear case:

• Squeeze fragility. The record set on 10 September lacked sustained buying interest, per Bloodstone Research, and unwound within days once tariff optimism faded.
• China construction drag. ING characterises Chinese copper demand as “mixed,” with property completions lagging and construction-related use depressed, a structural gap that electrification themes have not yet filled.

China’s economic rebalancing away from property-led construction toward domestic consumption and infrastructure investment is the structural force that explains why manufacturing PMI can cross back above 50 while construction-related copper use stays depressed, a divergence that makes aggregate demand figures more difficult to interpret than headline PMI suggests.

• Supply normalisation. Chilean mine disruptions provided part of the price support. If output recovers faster than expected, or new smelter capacity ramps up in response to high prices, the visible tightness narrative weakens.

Kitco has also flagged dollar strength as an additional headwind, the kind that caps rallies even when physical signals stay tight.

The bears are not forecasting a collapse. Their argument is narrower and sharper: current spot sits well above where fundamental balance would place it, which makes the risk-reward on new long positions asymmetric. For you, that asymmetry is a sizing question, not a reason to exit. Knowing which conditions would signal the real floor is lower than the current price is the early-warning system worth building.

What Q3’s pattern suggests about price behaviour from here

The most reliable guide to what copper does next is what it has already done this year. The 2026 pattern is consistent enough to read.

Read down that table and a rhythm emerges: record, rapid retracement, consolidation, partial recovery. Copper has not moved in straight lines from its highs this year; it has tested the demand underneath each new record and settled somewhere below it. The late August consolidation and the mid-September losing week are the same behaviour repeating.

Copper price volatility patterns in 2026 have followed a consistent structure: a speculative-driven spike to a new high, a sharp retracement once the trigger fades, and a consolidation phase that reveals where genuine physical demand sits, a sequence that makes the record-then-reversal behaviour on 10 September legible rather than random.

That pattern carries a warning for how you read the all-time high. The US$14,875/t print is not the new floor. The roughly US$14,400/t month-end level is closer to the market’s current best estimate of fair value once squeeze and tariff dynamics clear, and even that is an estimate, not a guaranteed minimum.

The two variables that will determine whether the floor holds in Q4

The forward view narrows to two things worth watching.

The first is the ShFE inventory trajectory through October. As post-holiday demand resumes after China’s National Day break, continued drawdowns would confirm the floor is real demand-driven; a sharp rebuild would suggest the September tightness was partly logistical.

The second is US refined-copper tariff policy. Progress toward tariffs keeps pulling metal toward America and supports non-US tightness; abandonment would, per ING, push inventory back onto global markets and apply downside pressure.

ING frames the balance of risks accordingly: upside from further supply disruptions and energy-transition plus AI investment momentum, downside from weaker demand or a tariff reversal releasing stock.

What the Q3 evidence tells you, and what it does not

Weigh the two sides and a calibrated position emerges, not a coin toss. The bullish signals are verifiable: the 70% ShFE drawdown, fabricators consuming metal directly, PMI back above 50, and a structural electrification demand base that does not depend on tariff politics. The risks are equally real: squeeze fragility, China’s construction drag, and the potential for supply to normalise.

Here is what the evidence actually confirms, and what it does not:

• Confirmed: A genuine physical demand floor exists somewhere below current spot. Real end-users are pulling metal, and the macro backdrop supports it.
• Contested: That the US$14,875/t all-time high represents a durable new equilibrium. The speed of every reversal this year argues against it.

The single most telling data point is the gap between ING’s US$11,500/t 2026 average forecast and the roughly US$14,400/t spot price. Either consensus was wrong by a wide margin, or current prices carry a premium that partially unwinds. Deciding which is the real analytical task in front of you now.

That makes the two variables from the previous section, October inventory data and tariff policy, the signals worth tracking above all others. Copper’s Q4 trajectory hinges on whether structural demand can hold price in the US$14,000-14,500/t range once the distortions clear.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results, and financial projections are subject to market conditions and various risk factors.