Extraction / Resources & Sovereignty
In May 2026, copper prices on the London Metal Exchange reached $14,153 per metric ton, within 2.5% of the all-time record above $14,500. The price had crossed $12,000 in December 2025, then $14,000 in January 2026. Behind the price move is a structural shift: the refined copper market is projected to swing from a slight surplus in 2025 to a deficit of 150,000 to 330,000 tonnes in 2026, depending on the forecast. The International Copper Study Group puts the shortfall at approximately 150,000 tonnes. JPMorgan’s more conservative supply scenario projects up to 330,000 tonnes. Morgan Stanley forecasts a 590,000-tonne deficit in 2026, widening to 1.1 million tonnes by 2029, the most significant shortage in more than two decades. The numbers diverge, but the direction does not: every major forecast points to deficit.
The demand side of the equation has a new structural driver that did not exist at scale three years ago. AI data centres require copper throughout their physical architecture: in power distribution busbars, cooling water systems, high-density server interconnects and facility-wide electrical wiring. JPMorgan estimates that AI-related data centre construction could generate approximately 110,000 additional tonnes of copper demand in 2026 alone. BloombergNEF projects that data centre infrastructure globally could contain more than 4.3 million tonnes of copper by 2035. The IEA forecasts that data centres, which currently consume about 1.5% of global electricity supply (roughly equivalent to the entire United Kingdom), will more than double their energy consumption by 2030, with AI responsible for the bulk of the increase. By the end of the decade, data centres could be consuming more than half a million metric tonnes of copper annually.
What makes AI copper demand structurally different from traditional demand categories is its price inelasticity. Hyperscaler technology companies (Google, Microsoft, Meta, Amazon) have committed a combined capex of approximately $725 billion in 2026, up 77% from 2025. When a company is spending tens of billions on a single data centre campus, the copper cost per facility, while significant in tonnage, represents a small fraction of total project cost. These buyers do not reduce copper orders when the price rises from $10,000 to $14,000 per tonne. They secure supply and build. This inelasticity changes the dynamics of the copper market. It means that a portion of demand is effectively insensitive to price signals that would traditionally moderate consumption.
The supply side is not keeping pace. Chile and Peru, which together supply nearly 40% of global copper, face operational delays, declining ore grades and logistics constraints. Approximately 6% of global supply is currently offline due to labour strikes, adverse weather and infrastructure bottlenecks. PT Freeport Indonesia cut its 2026 production target at the Grasberg mine to 700 million pounds, down from earlier guidance. Treatment and refining charges have compressed, signalling a shortage of copper concentrate feedstock even as smelters expand capacity. A new constraint has emerged: China’s restrictions on sulphuric acid exports are tightening supply of the chemical essential for oxide ore processing, directly affecting producers in the DRC and Zambia that rely on acid leaching for roughly 20% of their output.
The African supply dimension is where the global deficit narrative intersects with regional dynamics already covered in this series. The DRC produced approximately 3.3 million tonnes of copper in 2024, making it the world’s fourth-largest producer. Zambia produced a record 890,346 tonnes in 2025, an 8% increase from 2024. Together, the two countries account for roughly one-sixth of global output. DRC production is forecast to grow 6% in 2026 to reach approximately 3.4 million tonnes, driven by normalisation at Kamoa-Kakula following the May 2025 seismic event, continued output from CMOC’s operations (650,200 tonnes in 2024), and steady performance at Tenke Fungurume, Mutanda and Kinsevere. Zambia targets over 1 million tonnes in 2026, supported by brownfield expansions at Kansanshi, Lumwana and Konkola, with a long-term ambition of 3 million tonnes annually by 2031.
But Q1 2026 data introduces a complication. Zambian copper output fell 4.27% in the quarter. DRC copper exports declined by nearly 15% in the same period. The simultaneous contraction across both Central African producers is a notable combined supply signal. Kamoa-Kakula’s revised guidance of 290,000 to 330,000 tonnes for 2026 (down from an earlier 380,000 to 420,000 tonnes) means the operation that was expected to be the primary buffer against the global deficit is itself operating below potential. The DRC government’s cobalt export ban adds another layer of supply-side complexity to an already strained regional production corridor.
The infrastructure dimension connects the supply challenge to the corridor competition documented elsewhere in this series. The Lobito Corridor and the TAZARA rehabilitation create competing export routes for Central African copper. Kamoa-Kakula’s first anode shipment via Lobito in Q1 2026, with a seven-day transit to the Atlantic, demonstrated the potential for cost and time savings. For a global market where marginal tonnes matter, the efficiency of the export route directly affects how quickly African copper reaches consuming markets. The sulfuric acid constraint adds a further bottleneck: Kamoa-Kakula’s smelter produces 1,350 tonnes of sulphuric acid per day as a by-product, sold to regional oxide copper producers. As Chinese acid exports tighten, regional acid supply becomes a competitive advantage for producers operating near the Kamoa-Kakula smelter and a constraint for those who do not.
The copper price environment reshapes project economics across the continent. At $14,000 per tonne, deposits that were marginal at $8,500 become attractive. KoBold Metals plans to begin shaft construction at its Mingomba project in Zambia in H1 2026, targeting annual production exceeding 300,000 tonnes. The project, backed by Bill Gates, represents a $2.3 billion investment. In Botswana, MMG acquired the Khoemacau mine for $1.875 billion. Angola’s Tetelo project represents new capacity from previously undeveloped resources. The exploration and development pipeline is responding to the price signal. But the timeline mismatch remains: in developed jurisdictions, opening a new copper mine takes 20 to 30 years on average. Even in copper-rich countries like Chile, where $14.8 billion in new projects are planned, most will not boost output before 2028 or 2029.
The World Bank expects copper prices to rise by an average of 21% in 2026, marking a yearly record, before declining approximately 8% in 2027 as disruptions ease. Citigroup sees copper potentially exceeding $13,000 and approaching $15,000 per tonne in 2026 if supply shortages and low inventories persist. S&P Global projects that global copper demand will grow by 50% by 2040, reaching approximately 42 million tonnes annually. The IEA’s latest critical minerals outlook places copper on a path where existing and planned mines meet only about 70% of projected 2035 demand.
The structural question is whether copper is transitioning from a cyclical commodity to a strategic metal. The convergence of AI infrastructure, electrification, renewable energy and electric vehicle production creates demand that is not correlated with traditional business cycles. Each of these demand drivers operates on its own investment logic, and none is expected to contract in the 2026 to 2035 window. On the supply side, the concentration of growth in the DRC and Zambia means that African operational and institutional conditions, from seismic events at Kamoa-Kakula to mining code revisions in Lusaka to export corridor competition, have a direct and measurable impact on the global copper balance. The deficit is not a projection. At current consumption and production rates, it is the baseline. The question is how wide it gets, and which producers are positioned to fill it.