Copper Prices Soar to Record Highs Amid AI Boom and Looming Supply Crisis

Copper Prices Soar to Record Highs Amid AI Boom and Looming Supply Crisis
Published on: Dec 8, 2025

Copper prices skyrocketed to unprecedented levels on Monday, with futures on the London Metal Exchange (LME) breaching $11,700 per tonne and marking a year-to-date surge of over 34%. This relentless rally is fueled by an insatiable confluence of demand from artificial intelligence (AI) infrastructure and traditional sectors, starkly juxtaposed against a constrained supply pipeline that analysts warn is nearing a breaking point.

The AI “Copper Beast” Awakens

The most voracious new source of demand stems from the global AI boom. The market is awakening to the staggering copper intensity of AI’s physical backbone: data centers. While a conventional data center uses 5,000 to 15,000 tonnes of copper, a single hyperscale facility built for AI training and operations can devour up to 50,000 tonnes—more than the combined total of three traditional centers. The International Energy Agency (IEA) predicts data center electricity demand will double by 2030, potentially leading to annual copper consumption exceeding 500,000 tonnes from this sector alone.

Critically, this new demand is largely price-inelastic. According to consultancy Wood Mackenzie, copper accounts for less than 0.5% of a data center’s total project cost. “Whether copper is $10,000 or $20,000 per tonne, the AI infrastructure build-out won’t slow down,” noted an industry veteran. “The demand rigidity is formidable.”

Traditional Demand Remains Robust

As the new “beast” feasts, traditional consumers’ appetites remain strong. China, the world’s top metals consumer, has reaffirmed a proactive fiscal stance and prioritized economic growth for the coming year. Nationwide initiatives in grid upgrades, electric vehicles, and renewable energy—all heavily copper-dependent—continue to underpin robust demand. “Copper will benefit from policy support for power grids and computing power. The bullish momentum remains very strong,” said Xu Wanqiu, an analyst at Cofco Futures.

Simultaneously, preemptive stockpiling in the United States, driven by anticipated tariff expansions, is draining immediately available copper from the global market, exacerbating near-term supply anxieties and supporting prices.

A Looming Supply Crisis: The 30% Deficit Warning

While demand accelerates, supply struggles to keep pace, revealing a deepening structural deficit. Citic Securities warns the global refined copper shortfall could reach 450,000 tonnes next year. The long-term outlook is even more alarming. The IEA recently cautioned that copper could face a supply deficit as high as 30% by 2035, potentially making it one of the most vulnerable links in global supply chains.

The issue is not a lack of resources but the protracted and capital-intensive process of bringing new mines online. Wood Mackenzie estimates that meeting 2035 demand will require 8 million tonnes of new mining capacity, needing investments surpassing $210 billion. However, the development timeline is daunting: permitting and building a new copper mine in the U.S. averages 19 years, with global discovery-to-production cycles often spanning decades.

The View Through the Telescope

In response to this powerful demand-supply narrative, major investment banks are revising targets upward. JPMorgan forecasts copper could reach $12,500 per tonne by Q2 2026, while UBS projects a climb to $13,000 by the end of next year.

The current rally represents a historic convergence where all major growth engines—the AI revolution, energy transition, and broad electrification—are simultaneously funneling demand into copper. As a single AI facility can consume as much copper as a small city, and as authoritative bodies sound alarms over catastrophic shortages, the market’s fervor—and anxiety—may just be beginning. The telescope searching for copper’s price ceiling may need to be pointed even higher as the “hunger” shows no signs of abating.

AI Base Metals China News Copper