Indonesia Smelter Failure Extends Copper Record Rally as Global Stockpiles Near Crisis

Copper futures rose for a second consecutive session on Tuesday, reaching $6.61 per pound on COMEX — five days after the metal set an all-time intraday record of $6.7045 per pound on August 6 and still within 1.8% of that peak. The price has not retreated. The reason is not momentum: it is a three-front supply collapse that has accelerated since the record was set, with a fourth leg confirmed just this morning. Exchange inventory on the London Metal Exchange, now at roughly 223,000 metric tons (492 million lbs) as of August 7 — down from 307,000 metric tons (677 million lbs) at the start of July — has been draining at a historically rapid pace, a physical signal that buyers are paying urgency premiums to secure metal that increasingly does not exist in deliverable form, per COMEX copper price data.

For any organization building an AI data center, an electric vehicle plant, or a renewable energy grid, the takeaway is not that copper had an exceptional week last week. It is that the price floor under the commodity powering every one of those projects has moved — structurally, not speculatively — and it shows no sign of returning to the levels that capital expenditure models written before 2026 assumed.

Gresik Smelter: The Supply Shock Nobody Saw Coming

The newest disruption is also the least discussed. On August 8, PT Smelting's Gresik copper smelter in East Java, Indonesia — a joint facility owned by Freeport Indonesia and Japan's Mitsubishi Materials — went offline after a water leak from a boiler. Buyers were told the plant would remain closed for checks and repairs over the coming weeks. The Gresik facility has a designed copper cathode production capacity of approximately 342,000 metric tons (754 million lbs) per year and processes ore exclusively from Freeport's Grasberg mine, the world's second-largest copper operation, according to SP Angel's August 11 market analysis.

The closure comes at the worst possible moment. Freeport's second smelter at the Manyar site, which had been ramping up as Gresik's replacement capacity, is not expected to restart until September 2026. Indonesia's government is expected to route Grasberg concentrates to Chinese smelters in the interim — meaning the copper will still be processed, but the timeline for cathode delivery lengthens, the supply temporarily disappears from near-term exchange-deliverable inventory, and the already-thin LME stock position tightens further.

Chile's Worst Multi-Mine Winter in Years

The Chile disruption has deeper roots and longer consequences than any single smelter outage. Beginning around July 18, an unusually severe winter storm swept across Chile's copper belt — a stretch of central and northern mine country that accounts for roughly a quarter of global mined copper output, according to reporting on Chile storm mining disruptions.

What made this storm structurally different from prior weather events was the breadth of its simultaneous impact. Codelco suspended surface activities at its Andina mine northeast of Santiago while snow forced the state miner to halt ore shipments from the open-pit operations at El Teniente — the world's largest underground copper mine. Anglo American assessed conditions at its Los Bronces operation and adjusted accordingly. Antofagasta reduced non-essential activities at its flagship Los Pelambres mine; the operation experienced an orderly shutdown of approximately five days before resuming production on July 23.

Lundin Mining's Caserones operation, situated at 4,600 meters (15,093 feet) above sea level in the Atacama region, bore the most severe impact. A combination of heavy snowfall and resulting road impassability cut site access while simultaneously disrupting the mine's primary grid-connected electricity supply. Post-storm infrastructure assessments confirmed damage to two power transmission towers, and Caserones remained without power from July 18 onward, with full power restoration and a gradual restart expected to take approximately two to three weeks, per Lundin Mining's Chile statement. Bloomberg Intelligence estimated the Caserones outage alone would result in a production loss of approximately 10,000 metric tons (22 million lbs) of copper.

Analysts estimated the simultaneous curtailments temporarily placed approximately 1.6 million metric tons (3.5 billion lbs) of annual production capacity under suspension risk at once. That figure does not represent permanent loss — most operations have since resumed — but it contributed directly to a supply signal the market has not forgotten, per Chile copper disruption analysis.

The deeper problem is structural and predates the storm. Chile's Q2 2026 national copper output came in at 1.27 million metric tons — down 7.7% year-over-year and the weakest April-to-June quarterly result in 19 years. Antofagasta reported that its own first-half 2026 production fell 9.5% year-over-year, per the Antofagasta H1 production report. Codelco — the world's largest copper producer — produced 292,300 metric tons (644 million lbs) in Q2 2026, down 14% year-over-year, placing the company's quarterly output near a 28-year low, per Chile Q2 copper output data citing Cochilco.

Codelco's debt has hit a record $25 billion, and its chairman Bernardo Fontaine confirmed there is no realistic path to the company's long-standing 1.7-million-metric-ton annual production target within the next four to five years. Chile's newly elected government has responded by allowing Codelco to retain all of its 2025 profit — approximately $2.42 billion — for reinvestment for the first time in the company's history, reversing a longstanding policy of transferring profits to the state. That policy shift matters for long-term supply recovery; it does nothing for near-term output.

The underlying mechanism that makes Chile increasingly weather-vulnerable is ore grade decline at its mature porphyry copper deposits. As ore grades fall at mines like El Teniente and Andina, miners must process larger volumes of material to extract the same quantity of copper — which means greater dependence on electricity, water, and transport infrastructure per unit of output. When any of those inputs is interrupted by weather, the production loss per unit of disruption is larger than it was a decade ago.

Why the Market Refuses to Believe the Global Balance Sheet

The copper market's refusal to pull back from record territory is creating a sharp institutional disagreement about what 2026's supply-demand balance actually looks like.

J.P. Morgan's analysts forecast a 330,000-metric-ton (728-million-lb) refined copper shortfall in 2026 — the largest deficit in years, per JPMorgan copper deficit forecast. Goldman Sachs raised its end-2026 LME copper target in June to $13,735 per metric ton (approximately $6.23 per pound), up from a prior forecast of $12,465 per metric ton, after slashing its global mine supply estimate by 350,000 metric tons to account for ongoing disruptions at the Grasberg complex and Ivanhoe's Kamoa-Kakula operation in the DRC, per the Goldman copper price target. Citigroup has put a $15,000-per-metric-ton target within a year, per the Citi copper price target.

The International Copper Study Group, a more conservative institutional voice, projects a modest 96,000-metric-ton (212-million-lb) refined surplus for 2026, noting that unreported stock changes can materially alter the apparent balance. The disagreement between a 330,000-tonne deficit and a 96,000-tonne surplus is not a rounding error — it reflects a genuinely contested question about how much of the tariff-front-running inventory in US COMEX warehouses represents real accessible supply versus metal that is functionally unavailable to the rest of the global market.

How Tariff Arbitrage Is Making the Global Shortage Worse

This is the mechanism the price action is expressing that the aggregate balance-sheet numbers obscure. Since the Trump administration's 50% tariff on semi-finished copper products and the anticipated — but still unannounced — tariff on refined copper imports (potentially 15% starting January 2027, rising to 30% by 2028), importers have been shipping copper into US warehouses at an accelerating rate. More than 200,000 metric tons (441 million lbs) entered US ports in July 2026 alone, the largest monthly inflow in over a decade. COMEX inventories reached record levels exceeding 650,000 metric tons (1.43 billion lbs) by late June 2026 — while LME inventories outside the US continued draining, as detailed in the prior TechTimes copper record DRC ban article.

The Commerce Department's June 30, 2026, deadline to recommend whether to impose refined copper tariffs passed without an announcement. Goldman Sachs' base case assumes the US will continue to delay the decision. But the incentive to ship copper into US storage persists regardless of when the tariff lands — because the tariff has not been ruled out, and the arbitrage opportunity remains live.

The result is a bifurcated global copper market: record stockpiles in the US sitting alongside critically depleted global exchange inventory. LME warehouse stocks at 223,000 metric tons (492 million lbs) represent roughly six weeks of drawdown from a July starting point of 307,000 metric tons (677 million lbs). Cash copper has been trading at a premium to the three-month contract — a condition known as backwardation — that widened to as much as $148 per metric ton on August 7. The cash-to-three-month spread reached $150 per metric ton in August, its highest reading since October 2025.

Backwardation is not a speculative signal — it is the opposite. It indicates that buyers are willing to pay a meaningful cash premium to secure copper for immediate delivery rather than accept a contract for future delivery. When the cash premium reaches $148–$150 per metric ton on a $14,000-per-metric-ton commodity, it means the market is expressing that the copper in exchange warehouses right now is acutely scarce relative to what buyers need today. Goldman now estimates the copper deficit outside the United States could exceed 640,000 metric tons in 2026 — a figure that looks nothing like the more moderate global-balance number, because the US stockpile is functionally unavailable to buyers in Europe, Asia, and the rest of the developing world.

How SX-EW Production Works — and Why Acid Still Matters

Supplementing the Chile and Indonesia supply disruptions, a third chemical shock continues to compress a separate portion of global copper output: the collapse of sulfuric acid availability, which directly threatens the solvent extraction-electrowinning method (SX-EW) that accounts for 15 to 20 percent of world refined copper production.

SX-EW is the hydrometallurgical process that recovers copper from oxide ore deposits — the surface-weathered rock that sits above the sulfide ore bodies most people associate with copper mining — without a smelting furnace. The process runs in three stages: crushed oxide ore is irrigated with dilute sulfuric acid, which dissolves copper ions from the rock into a solution called pregnant leach solution; that copper-bearing solution is then contacted with an organic extractant that selectively pulls copper ions out of the aqueous phase; and finally, copper is plated from the electrolyte onto stainless-steel cathode blanks using electrical current. The output is 99.99%-pure copper cathode. The critical requirement is sulfuric acid, consumed at more than one metric ton (2,205 lbs) of acid per metric ton of copper produced — and no commercial substitute exists at scale.

The Strait of Hormuz closed to dry bulk traffic on February 28, 2026, cutting roughly half of seaborne sulfur exports from the Persian Gulf. China's April 2026 ban on sulfuric acid exports, in effect through August 2026, eliminated approximately 25% of global acid supply simultaneously. The DRC's copper belt — the most SX-EW-dependent major copper region in the world — imports roughly 90% of its sulfur from the Middle East and has been running on constrained acid supply throughout the second quarter. The acid ban's August expiration date is approaching, which could relieve one pressure point — but analysts note that sulfur logistics pipelines will take weeks to rebuild, and Hormuz remains closed as of today.

Demand Has Changed Categories

What distinguishes the current copper cycle from every prior one is not the price level — it is the character of the demand that is sustaining it.

Previous copper supercycles were driven primarily by China's construction and infrastructure boom — demand that was cyclical, geographically concentrated, and subject to policy correction from Beijing. The current demand stack is different in kind: AI data centers, electric vehicles, and power grid modernization are proceeding simultaneously across multiple economies, with limited cross-sector substitution and little sensitivity to the price signals that slow down speculative industrial demand.

J.P. Morgan has estimated that a single large-scale AI data center can require up to 50,000 metric tons (110 million lbs) of copper, with total data center demand for the metal projected to reach 475,000 metric tons (1.05 billion lbs) annually in 2026, per JPMorgan data center copper. Wood Mackenzie has projected that AI-driven electricity requirements alone will push copper demand for grid infrastructure to 1.1 million metric tons (2.43 billion lbs) per year by 2030, per Wood Mackenzie copper demand. Trafigura has estimated that AI and data center-related uses could add an additional 1 million metric tons (2.2 billion lbs) of copper demand by 2030, per Trafigura AI copper demand.

Electric vehicles require roughly three to four times more copper than internal combustion vehicles. Grid modernization programs — the transmission lines, substations, and transformers needed to connect renewable generation to urban demand centers — consume copper-intensive infrastructure at accelerating rates across dozens of countries at once.

S&P Global has forecast that global copper consumption will rise from approximately 28 million metric tons (61.7 billion lbs) in 2025 to more than 42 million metric tons (92.6 billion lbs) by 2040 — a roughly 50% increase, per S&P Global copper demand. The International Energy Agency has estimated that achieving global climate targets will require doubling copper demand for clean energy technologies by 2030 compared to 2020 levels.

CME Group's markets team has described the current environment as "inelastic demand across multiple sectors simultaneously," meaning price increases that would historically slow demand in one sector are now being absorbed because the structural drivers — AI build-out, EV mandates, grid modernization — are not price-elastic in the short term. A data center company cannot substitute aluminum for copper in its power distribution infrastructure because copper physics doesn't have a budget-sensitive alternative at the performance levels required.

What Happens When the Record Breaks — or Doesn't

The specific question for the next 60 days is whether the conditions that set the all-time record on August 6 have deteriorated enough to push copper through $6.87–$6.90 per pound — or whether the combined weight of the Gresik outage, still-partial Caserones restoration, continued DRC waiver uncertainty, and LME drawdown is sufficient to hold the current price floor near $6.61.

The DRC ban's ministerial waiver mechanism remains the largest variable. Mines Minister Louis Kabamba Watum retains the authority to issue one-year export waivers under "strategic" circumstances — a provision that operators including CMOC, Glencore, Ivanhoe, and Zijin Mining will be pressing aggressively. The history of the 2013 nominal ban (routinely waived) versus the June 29 order (three ministries, an additional by-product tax, and explicit policy signaling against default waivers) suggests the DRC is serious — but the waiver timeline will determine whether global smelters face a genuine concentrate famine or a negotiated transition.

Goldman's base case has copper sustaining near record levels through the second half of 2026 assuming Chilean mine recovery timelines remain uncertain and the DRC ban holds with limited waiver uptake. The floor the market appears to be pricing in is not the record itself — it is the level at which the structural cost of building the AI economy and the clean energy transition is now denominated.

The metal that spent most of the past decade as a cyclical barometer of Chinese industrial activity has been repriced. The question is not whether the record will be broken — it is whether the infrastructure decisions being made today have fully reflected what a $6.61-per-pound copper floor means for every project scheduled to break ground between now and 2030.

Copper futures data sourced from Trading Economics and Investing.com. This article is for informational purposes only and does not constitute investment advice.

Frequently Asked Questions

Why hasn't copper retreated from its all-time high five days later?

Three separate supply disruptions arrived in rapid succession. Chile's worst multi-mine winter storm in years simultaneously suspended or curtailed output at Codelco's Andina mine, El Teniente's open-pit operations, Anglo American's Los Bronces, Antofagasta's Los Pelambres, and Lundin Mining's Caserones — placing 1.6 million metric tons (3.5 billion lbs) of annual production capacity at simultaneous risk. The DRC's concentrate export ban remains in effect, removing feedstock from global custom smelters. And on August 8, PT Smelting's Gresik smelter in Indonesia — which processes ore from the world's second-largest copper mine at Grasberg — went offline after a boiler failure and is expected to remain closed for weeks. LME exchange inventory has been draining at a historically rapid rate, per copper supply disruption analysis. When physical supply is this constrained, prices don't retreat because the buyers who need copper right now are competing for what little exists in deliverable form.

Why does "backwardation" matter, and what does the $148-per-ton premium tell you?

Backwardation is a futures market condition in which the cash (spot) price of a commodity exceeds the futures price for delivery at a later date — the opposite of normal market structure, in which future delivery typically costs more to account for storage and financing. When copper cash is trading at $148 per metric ton above the three-month futures contract, it means buyers are willing to pay a material premium to receive physical metal today rather than accept a contract for metal in three months. That is not a speculative signal — it is a physical scarcity signal. Speculators drive up futures prices. Only genuine physical buyers competing for limited immediately deliverable supply drive up the cash-to-futures premium. The $148–$150 per metric ton backwardation recorded in early August 2026 was the widest spread since October 2025, per LME copper backwardation analysis, and it confirms that the copper market's tightness is real and physical, not a function of leveraged positioning in futures markets.

How does the tariff arbitrage dynamic make the global shortage worse even if the US never imposes a copper tariff?

The mere credible possibility of a refined copper tariff — starting at 15% in January 2027, rising to 30% by 2028, as one scenario under the still-pending Section 232 review — is sufficient to maintain the economic incentive to ship copper into US warehouses. More than 200,000 metric tons (441 million lbs) entered US ports in July 2026 alone, the largest monthly inflow in over a decade, pushing COMEX inventories past 650,000 metric tons (1.43 billion lbs). That copper is effectively "trapped" from the perspective of the rest of the global market: it is physically present in the US, but it is not available to European, Asian, or African buyers through LME channels. Goldman Sachs estimates the copper deficit outside the United States could exceed 640,000 metric tons in 2026 — a figure that looks very different from the more moderate global-balance numbers, because the global balance counts US COMEX stockpiles as supply when they are functionally inaccessible to non-US buyers.

What does a $6.61-per-pound copper floor mean for AI data center and EV construction costs?

J.P. Morgan estimates a single large-scale AI data center requires up to 50,000 metric tons (110 million lbs) of copper for wiring, power busbars, cooling infrastructure, and high-voltage distribution. At $6.61 per pound, that represents a copper-alone material cost of approximately $728 million for a large hyperscale facility — compared to approximately $590 million at the $5.35-per-pound price that prevailed a year ago, per JPMorgan data center copper. Electric vehicles require three to four times more copper than combustion-engine cars, meaning every dollar increase in the structural copper price baseline flows directly into vehicle production costs. Wood Mackenzie projects that AI-driven electricity infrastructure alone will consume 1.1 million metric tons of copper annually by 2030. Capital expenditure models for any of these projects that were built before 2026 need to be revisited.