Why Copper Prices Are Falling Despite a Record-High Quarter
Key Takeaways
- LME three-month copper traded at $14,616.50 per metric ton on 24 September 2026, roughly 1.7% below the record high of $14,875, with the pullback driven primarily by dollar strength and Fed rate expectations rather than a breakdown in the structural demand case.
- BHP suspended all operations at Escondida, the world's largest copper mine, on 23 September 2026 following a fatal accident, with no restart timeline provided, creating genuine near-term supply uncertainty at the single largest source of mined copper.
- A strike authorisation vote at Antofagasta Minerals' Centinela mine is scheduled for 26 to 28 September 2026, representing a binary near-term supply risk that could tighten the market further if unions vote to proceed.
- The LME cash-to-three-month spread widened to $107.50 per ton on 24 September 2026, the highest since 1 September and up from near zero a week earlier, signalling physical tightness but requiring caution given the 2026 pattern of flash squeezes and rapid reversals.
- Wood Mackenzie's base case projects copper demand growing 24% to 42.7 Mt per year by 2035, anchored by EV and AI data centre demand, a structural gap that today's macro-driven price retreat does not close by a single tonne.
Copper just set a record high of $14,875 per metric ton roughly two weeks ago and has climbed about 9% this quarter. Today it is falling.
That is the tension worth sitting with. A metal in the middle of one of its strongest structural bull runs in years is selling off on the very day the world’s largest copper mine sits idle after a fatal accident.
Three forces collided in a single session on 24 September 2026. The US dollar is hovering near a two-month peak, Chinese markets are shut for the Mid-Autumn Festival, and scepticism about a US-China trade breakthrough helped drive Chinese equities to their steepest single-session drop in a month.
The collision is the story, not any one of its parts.
After reading this, you will be able to separate which of today’s pressures are transient noise and which reflect a structural reality worth taking seriously before the quarter closes. The distinction matters because it determines whether today’s level looks like a re-entry point or the start of something longer.
What is actually driving today’s copper pullback
The surface explanation is China’s absence. The deeper one sits with the dollar and the Federal Reserve, and one strategist made the hierarchy explicit.
Alastair Munro, senior base metals strategist at Marex, characterised macroeconomic headwinds, rather than China’s holiday absence alone, as the dominant force weighing on the market.
That framing is the analytical anchor for everything that follows. LME three-month copper traded at $14,616.50 per metric ton as of roughly 1610 GMT on 24 September, after dipping to an intraday low of $14,526. That leaves it about 1.7% below the record high, a modest retreat by the standards of a metal up 9% on the quarter.
The dollar is doing the heavy lifting. Copper is priced in US dollars on global exchanges, so when the dollar strengthens, buyers holding other currencies face higher local-currency costs for the same tonne of metal. That pressure works through three channels.
The relationship between dollar strength and copper prices operates through multiple transmission channels simultaneously, and the 2026 cycle has made each of them visible in compressed timeframes that earlier cycles spread across months.
- Pricing currency effect: A stronger dollar raises the effective cost of copper for non-US buyers, from Chinese importers to European manufacturers, which trims price-sensitive demand and speculative long positioning.
- Interest-rate and growth expectations channel: Dollar strength here reflects expectations of further Fed rate increases, signalling tighter financial conditions and a weaker outlook for the industrial activity that consumes copper.
- China amplification: With the largest consumer of refined copper offline for the holiday, there is no domestic buying to absorb the macro pressure, leaving the sell-off room to run.
The read you should take from this is precise. The pullback is not evidence that copper’s bull case has broken. It is a demonstration of how quickly currency and rate expectations can overwhelm near-term fundamentals in a market where sentiment resets daily.
That matters for how you calibrate the current level. If you misread a dollar-driven dip as a structural reversal, you risk exiting a position over noise. The forces above China’s holiday, the dollar and the Fed, are the ones to watch for a genuine trend shift.
When big ASX news breaks, our subscribers know first
Two Chilean mines, one suspension and one looming strike vote
While the macro engine drove the price lower, the physical market told a different story. Two Chilean supply risks landed almost simultaneously, and their combined timing against already-elevated LME spreads is what makes them structurally significant rather than episodic.
| Mine | Operator | Event type | Status as of 24 September 2026 | Key variable to watch |
|---|---|---|---|---|
| Escondida | BHP | Fatal accident, full suspension | All operations suspended, no restart date | Restart timeline and contract talks |
| Centinela | Antofagasta Minerals | Strike authorisation vote | Vote called, scheduled 26-28 September | Vote outcome |
Escondida: suspended with no timeline
BHP confirmed the full suspension of all operational activities at Escondida, the world’s largest copper mine, on 23 September 2026, after a worker was killed during maintenance. The company stated that “all operational activities at the mine have been suspended” and gave no timeline for restarting.
Contract talks at the mine were already under way, which compounds the uncertainty. A stoppage of unknown length at the single largest source of mined copper is precisely the kind of event that supports nearby spreads.
Centinela: strike vote imminent
At Antofagasta Minerals’ Centinela mine, two unions, Trabajadores de Minera Esperanza and Trabajadores Distrito Centinela, declared that negotiations had reached a dead end and called a strike authorisation vote for 26 to 28 September 2026.
The outcome is a near-term binary risk. A vote to strike would tighten supply further, while a negotiated settlement would remove the threat.
Escondida and Centinela are the most visible examples of a broader pattern: copper supply chain disruption risks in 2026 span labour disputes, ageing infrastructure, and permitting bottlenecks across multiple jurisdictions, making single-mine events harder to read in isolation.
Here is the analytical frame that ties both together. Supply shocks of this kind exert their strongest pull on time-spreads and regional premiums, while macro factors set the direction of the three-month benchmark price. The LME cash-to-three-month spread widened to $107.50 per ton on 24 September, the widest since 1 September and up from essentially zero a week earlier.
For anyone with exposure to copper equities or futures, the Escondida suspension reinforces the case for near-term spread support. But the absence of a restart timeline means the supply-shock premium is genuinely uncertain in both size and duration, and the coming days will decide whether it holds or accelerates.
Why the structural demand case survives a bad day for prices
Step back from the daily tape and the picture changes. A single macro-driven retreat from record highs says nothing about where copper demand is heading over the next decade, and the longer map is unambiguous.
Wood Mackenzie’s base case projects total copper demand rising 24% to 42.7 Mt per year by 2035, driven by four disruptors.
- Energy transition: EVs, renewables, and grid infrastructure.
- Data centres: AI-driven electricity demand.
- Defence spending: including a large European military build-up.
- Economic development: rapidly industrialising economies, particularly India and Southeast Asia.
Two of these verticals matter most for the reader trying to size the demand pull, and neither is speculative.
EVs and data centres as the two biggest copper pulls
EV copper demand stood at 1.7 Mt per year in 2025 and is projected to reach 4.3 Mt per year by 2035, roughly 10% annual growth that requires an additional 2.6 Mt per year to be brought to market. That trajectory rests on regulatory mandates and contracted capacity already in motion, not hopeful forecasts.
AI data centres add a second pull. Their electricity demand is projected to consume an additional 2,200 TWh by 2035, lifting copper demand for grid infrastructure alone to roughly 1.1 Mt per year by 2030. Supporting the broader renewable build-out requires about 2 Mt per year of additional supply over the next decade.
Wood Mackenzie is honest about the risk in this, and its framing captures the whole tension in a phrase.
Wood Mackenzie describes the outlook as a “high-wire act”: soaring copper demand could itself become an obstacle to future economic growth if supply fails to keep pace.
The gap between what this demand case requires and what current mining investment and permitting timelines can realistically deliver is exactly where copper’s long-term bull case lives. Today’s pullback does not close that gap by a single tonne.
The lesson for your positioning is straightforward. A macro-driven retreat is not evidence that the structural thesis has changed. Conflating short-term volatility with a trend shift is how investors exit too early or miss the significance of a supply-demand imbalance building over a decade-level horizon.
Reading the LME spread: physical tightness, squeeze mechanics, or structural distortion?
That $107.50 per ton cash premium looks like a clean signal of shortage. It is not clean at all. The 2026 record shows at least three competing frameworks for interpreting copper backwardation, and each has evidential support this year.
| Date | Cash premium | Context or interpretation |
|---|---|---|
| June 2025 | ~$90/ton | Widest since August 2023, tied to copper flowing toward the US |
| Earlier 2026 | $93/ton peak, then $75/ton | Highest in ~2.5 years, linked to declining LME warehouse stocks |
| July 2026 | ~$33/ton | Tariff-driven trade-flow changes flipped the curve to backwardation |
| 20 August 2026 | $545/ton peak, then $176/ton | Flash squeeze, tom-next spread hit $110/ton before easing to $71/ton |
| 24 September 2026 | $107.50/ton | Highest since 1 September, up from ~zero a week earlier |
The three frameworks compete rather than form a hierarchy.
- Physical tightness driven by low inventory. Reuters linked earlier 2026 premiums directly to falling LME-registered warehouse stocks. The Escondida suspension fits this reading of genuine near-term scarcity.
- Speculative squeeze mechanics. The 20 August flash squeeze drove the cash premium to $545 per ton, the widest backwardation since 2021, with tom-next spreads exploding to $110 per ton. That was positioning and a squeeze on shorts, not a smooth reflection of physical scarcity, and even at the retreated $71 per ton it was painful for anyone rolling shorts.
- Tariff or structural regional dislocation. In July, trade-policy changes flipped the curve into backwardation at a cash premium near $33 per ton, with the risk that the split between the LME and other venues becomes structural rather than reflecting a global shortage.
LME warehouse stock dynamics in 2026 have been shaped by more than physical demand: Trafigura’s large-scale withdrawals and COMEX arbitrage flows created inventory dislocations that made cash premiums harder to read as clean physical-scarcity signals.
So what does today’s $107.50 per ton actually tell you? It is meaningful but inherently ambiguous. It is consistent with real supply tightness from the Chilean disruptions, but the 2026 pattern of rapid spikes and retreats means you should treat it as a signal to watch rather than a conclusion to act on. Read every backwardation spike as a physical shortage and you risk being caught in squeeze mechanics or dislocation effects that reverse hard.
The next major ASX story will hit our subscribers first
What the next four weeks will tell you about copper’s direction
Three variables will decide whether this pullback extends or reverses, and each resolves a different piece of the current ambiguity.
- The Centinela strike vote (26 to 28 September). A vote to strike tightens supply and supports spreads; a settlement removes the threat.
- Any Escondida restart announcement. A quick restart releases the supply-shock premium; continued silence keeps nearby tightness in place.
- The dollar’s trajectory against Fed signals. With the dollar near a two-month peak, further strength caps rallies, while a retreat lets fundamentals reassert.
The interplay is what matters, not any single reading.
The most bullish near-term combination is a Centinela strike alongside a retreating dollar. The most bearish is a quick Escondida restart alongside further dollar strength.
Read those combinations against the trend backdrop. Copper sits about 1.7% below its $14,875 record and up roughly 9% on the quarter, so the bull trend stays intact until the macro forces structurally shift rather than merely spike.
Monitoring these three variables over the next four weeks puts you in a better position to tell a temporary macro correction from a genuine trend break than daily price moves ever could. That analytical clarity is more useful than any directional call on where copper closes the quarter.
Investors wanting to model the macro and supply variables beyond the current session will find our full explainer on Q4 copper price dynamics, which examines why the majority of exchange-registered stocks are effectively unavailable to the physical market and what that means for spread behaviour into year-end.
Where the balance of evidence sits on 24 September 2026
Pull the four threads together and the picture resolves into something honest rather than tidy. The bull case pillars remain intact: the $107.50 per ton spread points to physical tightness, two Chilean supply risks are live, and Wood Mackenzie’s 24% demand growth to 42.7 Mt by 2035 anchors the decade ahead.
Against that stand the forces with the most near-term price influence, the dollar near its two-month peak and expectations of further Fed rate increases. These have the capacity to extend the correction if they do not moderate.
The result is a market where the near-term and the long-term point in different directions at once. Today’s pullback is not a bull-case breaker, but the macro environment can prolong it.
That leaves your time horizon as the single most important variable. Be honest about which horizon you are positioned for before you decide what today’s price action means, because the answer differs completely depending on the answer.
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.
Frequently Asked Questions
Why are copper prices falling today despite being near record highs?
Copper prices are pulling back on 24 September 2026 primarily because of a stronger US dollar near a two-month peak and expectations of further Federal Reserve rate increases, not because the structural bull case has changed. The absence of Chinese buyers during the Mid-Autumn Festival removed a key source of demand support, giving the macro pressure room to push prices lower.
What is the LME cash-to-three-month copper spread and what does it signal?
The LME cash-to-three-month spread measures the premium of immediate copper delivery over delivery in three months; a positive spread (backwardation) indicates near-term physical tightness. On 24 September 2026, the spread widened to $107.50 per ton, the highest since 1 September and up from near zero a week earlier, consistent with supply disruptions at Escondida and Centinela but also subject to squeeze mechanics that can reverse sharply.
What happened at Escondida and Centinela in September 2026?
BHP suspended all operations at Escondida, the world's largest copper mine, on 23 September 2026 after a worker was killed during maintenance, with no restart timeline given. At Antofagasta Minerals' Centinela mine, two unions declared negotiations at a dead end and called a strike authorisation vote for 26 to 28 September 2026, creating a near-term binary supply risk.
How much is copper demand expected to grow by 2035?
Wood Mackenzie projects total copper demand rising 24% to 42.7 million metric tons per year by 2035, driven by electric vehicles, AI data centres, defence spending, and industrialisation in markets such as India and Southeast Asia. EV demand alone is projected to grow from 1.7 Mt per year in 2025 to 4.3 Mt per year by 2035, requiring an additional 2.6 Mt per year to be brought to market.
How does a stronger US dollar affect copper prices?
A stronger US dollar raises the effective cost of copper for buyers holding other currencies, reducing price-sensitive demand and speculative long positioning. Dollar strength in 2026 also signals tighter financial conditions from Federal Reserve rate expectations, which weakens the industrial activity outlook that underpins copper consumption.

