By Charles Pitts
The global copper market is undergoing a structural transformation. The energy transition and rapid growth of artificial intelligence (AI) now drive demand more than traditional industrial cycles. As of mid-2026, LME benchmark copper has moved above $13,000 per tonne. That move has forced analysts to reassess the long-term ceiling for the metal. While Goldman Sachs and J.P. Morgan continue to forecast prices closer to $11,000 per tonne, analysts increasingly view a move toward $15,000 copper as possible.
Supply shortages are not the only catalyst. A major shift in demand elasticity is also driving the market higher. AI data center infrastructure has become a primary source of copper demand. Unlike many industrial sectors, hyperscale data centers remain largely insensitive to rising copper prices. In a market already facing supply deficits and operational disruptions, that behavior could push prices significantly higher.
The AI Catalyst: Why Data Center Demand is Different
The narrative surrounding copper has long focused on electric vehicles (EVs) and renewable energy grids. However, the 2026 market landscape is dominated by the build-out of hyperscale data centers.
S&P Global expects AI training infrastructure to account for nearly 60% of data center copper demand by 2030.
The AI sector operates under different economic rules than the automotive industry. Microsoft, Google and Amazon spend billions on data center construction. Copper represents only a small portion of total project costs. Land, AI chips and cooling systems account for most of the capital expenditure. As a result, higher copper prices rarely change investment decisions.
This is the definition of inelastic demand. Because these hyperscalers must build capacity to maintain competitive advantages in AI, they will pay whatever the market demands to secure physical delivery. As the copper demand forecast for 2026 suggests, this price-insensitive buying floor removes the usual “demand destruction” that historically caps commodity spikes.
The Supply Wall: A Deficit of “Hundreds of Thousands of Tonnes”
While demand is accelerating, the supply side of the equation is hitting a literal and figurative wall. The mining industry is currently grappling with a projected global copper market deficit of hundreds of thousands of tonnes in 2026. This is not a temporary blip but the result of a decade of underinvestment in new “greenfield” projects.
The current supply constraints are multi-faceted:
- Declining Ore Grades: Major producers in Chile and Peru are reporting lower copper concentrations in their existing pits, requiring more energy and more earth-moving to produce the same ton of finished metal.
- Operational Disruptions: As seen in recent updates from Zijin Mining’s Xinjiang operations, even with the adoption of electric fleets and autonomous technology, geopolitical risks and permitting delays remain a constant drag on output.
- Long Lead Times: The journey from discovery to the first ton of copper typically spans 12 to 15 years. The projects being approved today will not provide relief to the 2026 or 2027 markets.
When a supply shock occurs: such as a mine closure or a major labor strike: the market has no buffer. In a high-demand, low-inventory environment, the price must move exponentially higher to clear the market.
Price Inelasticity and the “Squeeze” Mechanism
In economics, price elasticity of demand measures how much the quantity demanded changes in response to a change in price. In a standard market, as prices rise, buyers find substitutes or reduce their consumption. However, copper is uniquely difficult to substitute in high-performance computing. While aluminum can be used in some grid applications, copper’s superior conductivity and thermal properties make it the only viable choice for the high-density power requirements of AI chips.
This creates a “squeeze” mechanism. If the market is short 200,000 tonnes of copper, and the biggest buyers (AI hyperscalers and national grid operators) refuse to reduce their orders regardless of price, the remaining small-scale industrial buyers are forced out of the market. The price must rise until it hits a level where marginal industrial users: who cannot afford $15,000 copper: are forced to shut down.
This scenario is exactly what happened during the 2021-2022 lithium spike, and the copper market is currently showing similar technical indicators. Speculative and financial flows are already starting to chase this “bottleneck” narrative, further accelerating the move toward the $15,000 target.
2026 Outlook: Base Case vs. The $15,000 Bull Scenario
To understand where we are headed, we must weigh the consensus against the outlier risks.
- The Base Case ($10,500 – $11,500/t): This scenario assumes that while the market is tight, a slowdown in the global manufacturing sector or a faster-than-expected ramp-up in secondary (recycled) copper supply will prevent a runaway spike. Goldman Sachs maintains a view that prices may “decline somewhat from record highs” as new supply eventually responds to the 2025 price signals.
- The Bull Case ($15,000/t): This scenario requires a “triple threat” of factors: sustained AI infrastructure spending, further supply disruptions in Latin America, and a weakening U.S. dollar. If the copper deficit for 2026 deepens due to resource nationalism or technical failures at major mines, the $15,000 level becomes the next technical resistance point.
For operators and investors, the key takeaway is that the “floor” for copper has fundamentally shifted. The days of $6,000 copper are gone, likely forever. Even in a bearish macro environment, the cost of production and the intensity of new-age demand provide a support level that was once considered a peak.
Future-Proofing the Supply Chain
For the mining industry, the path to $15,000 copper is both an opportunity and a call to action. Higher prices provide the necessary margin to invest in deeper underground mines and more complex processing facilities. We are already seeing a shift toward advanced technology, as captured in the Skillings media kit, where the focus is increasingly on “mining intelligence”: using data to find efficiencies where traditional methods have reached their limit.
As we look toward the second half of 2026, the question is no longer if copper will remain expensive, but how the global economy will adapt to a world where one of its most essential metals is in a state of permanent scarcity. The AI data center squeeze is merely the first chapter in this new commodity super-cycle.
