By Charles Pitts
Copper demand is entering a transformative and potentially volatile phase as two of the world’s most capital-intensive technological shifts: Artificial Intelligence (AI) and Electric Vehicles (EVs): collide. While the global energy transition has long been the primary driver for copper’s bullish long-term outlook, the sudden and explosive growth of generative AI and its associated data center infrastructure has introduced a “lumpy,” front-loaded demand profile that analysts believe will trigger a significant supply shock by 2026.
As mining operators and investors evaluate the 2026 landscape, the focus is shifting from gradual growth to an immediate scramble for “red gold.” According to recent forecasts from J.P. Morgan and UBS, the structural deficit in the copper market is expected to widen significantly within the next 24 months, driven by the dual pressure of high-intensity AI hardware and the maturing EV market.
The AI Surge: Why Bits Require Bricks (and Copper)
While AI is often discussed in terms of software and silicon, its physical reality is grounded in heavy infrastructure. Data centers are essentially massive heat exchangers and power distributors, both of which require immense amounts of copper. Unlike traditional data centers, AI-ready facilities require significantly more power density to handle the thermal loads and electrical requirements of advanced GPU clusters.
According to research from S&P Global, each megawatt (MW) of AI-grade data center capacity requires between 30 and 47 tonnes of copper. When including the external power generation and grid upgrades necessary to feed these facilities, the figure jumps to approximately 65,000 tonnes of copper per gigawatt (GW) of power.
J.P. Morgan estimates that AI data centers could add 110,000 tonnes of incremental copper demand by 2026. While this is a small percentage of the total 28 million-tonne global market, it represents a significant portion of the marginal supply: the “extra” copper that is typically available to buffer against shocks. BloombergNEF (BNEF) suggests that cumulative copper “locked into” data centers could exceed 4.3 million tonnes by 2035, highlighting that this is not a transient spike but a permanent architectural requirement of the modern world.
The EV Baseline: The Structural Giant
If AI is the “new player” causing a stir, the EV sector remains the structural giant. Despite shifts in consumer adoption rates and the rise of hybrids, the copper intensity of electrified transport remains a primary pillar of demand.
A standard internal combustion engine (ICE) vehicle uses roughly 22 kg of copper. In contrast, a battery electric vehicle (BEV) requires approximately 83 kg. This four-fold increase is found in the battery, the motor windings, and the extensive wiring harnesses. Even hybrid vehicles, often viewed as a “middle ground,” use nearly double the copper (40 kg) of their ICE counterparts.
Currently, EV-related demand stands at approximately 2.6 million tonnes per year. By 2040, this is projected to rise to 6.3 million tonnes. For 2026, the EV sector is expected to remain the much larger absolute consumer compared to AI, but its growth is more predictable. The “shock” factor in 2026 arises because AI demand is hitting the market at exactly the same time that EV infrastructure and vehicle sales are reaching critical mass in major markets like China and the EU.
2026: The Intersection of Scarcity
The year 2026 is increasingly identified as the “crunch year” because of the widening gap between these two demand drivers and a supply pipeline that is struggling to keep pace. UBS forecasts a global copper deficit of 230,000 tonnes in 2025, widening to over 400,000 tonnes in 2026.
To put this in perspective, the incremental demand from AI data centers alone (110,000 tonnes) could account for roughly one-quarter of that forecast deficit.
Data Comparison: Copper Consumption Profiles (2026 Projections)
| Metric | AI Data Centers (AI-Heavy) | Electric Vehicle Sector |
|---|---|---|
| Copper Intensity (Unit) | 30–47 t / MW (DC) | ~83 kg / BEV |
| Est. 2026 Annual Demand | ~0.2–0.3 Mt | ~2.7–3.0 Mt |
| Share of Total Demand | ~1% | ~10–11% |
| Growth Character | Lumpy, Front-loaded | Steady, Structural |
| Role in 2026 Deficit | Significant Marginal Driver | Primary Volume Driver |
Source: Compiled from BNEF, J.P. Morgan, and S&P Global data.
Supply Pipeline Realities and Operational Risks
The primary challenge for the mining industry is that copper supply cannot be “switched on” to meet sudden demand spikes. S&P Global data indicates that it takes an average of 17 years for a new copper mine to move from discovery to first production. This long lead time means that the supply available in 2026 is essentially fixed by decisions made a decade ago.
The current supply environment is also plagued by operational disruptions. Political and logistical issues in key regions: including the Democratic Republic of Congo (DRC), Latin America, and Mongolia: have recently disrupted 3% to 5% of global supply. For instance, the closure of First Quantum’s Cobre Panama mine and declining ore grades at established sites like Escondida in Chile have removed hundreds of thousands of tonnes from the market precisely when they are needed most.
As discussed in our previous report on copper demand and AI outlook, the industry is also facing “grade inflation,” where miners must process more rock to extract the same amount of metal, increasing energy costs and operational complexity.
What it Means for Operators and Investors
For mining professionals and decision-makers, the 2026 supply shock signals a period of heightened market tightness. While AI demand is smaller in volume, its “lumpy” nature: where a single 500 MW data center project can suddenly require 20,000 tonnes of copper: creates price volatility.
Operators are responding by prioritizing brownfield expansions and technology that improves recovery rates. In control rooms across the globe, the focus has shifted to maximizing throughput and reducing downtime to capitalize on expected price peaks, which BNEF suggests could reach $13,500 per tonne by 2028.
Furthermore, the 2026 outlook is accelerating M&A activity. As the cost of exploration rises, major producers are looking to acquire mid-tier companies with “shovel-ready” or active projects. This trend, as noted in our analysis of 2026 mining M&A targets, is a direct consequence of the realization that organic growth cannot meet the AI-EV demand curve.
Conclusion
The 2026 copper supply shock is not the result of a single factor but the synchronization of two massive technological tailwinds. While EVs provide the steady, high-volume floor for demand, AI data centers provide the sharp, marginal ceiling that threatens to push the market into a deep structural deficit. For the mining industry, the next 24 months will be defined by a race to squeeze more efficiency out of existing operations while navigating the geopolitical complexities of a world where copper has become as strategic as oil.
As the industry moves toward this 2026 milestone, the message is clear: the digital and green revolutions are built on a physical foundation of copper, and that foundation is showing signs of significant strain.
