Rio Tinto’s decision to supply copper produced through leaching operations in Arizona to Amazon for use in artificial-intelligence data centres highlights how rapidly AI infrastructure is colliding with physical resource limits. What might otherwise look like a routine supply agreement instead reflects a deeper shift: the digital economy is becoming increasingly constrained by access to metals, energy, and long-cycle industrial capacity.
The agreement links Amazon’s AI-focused data-centre buildout with copper produced via Rio Tinto’s Nuton leaching technology at a site operated by Gunnison Copper. While financial terms and volumes were not disclosed, the structure itself is revealing. At YourNewsClub, we see this as a form of upstream risk management rather than simple procurement. For hyperscale operators, copper is no longer a generic input purchased opportunistically; it is a strategic dependency tied directly to grid connections, power redundancy, cooling systems, and network density.
The broader backdrop is a sharp acceleration in copper demand expectations driven by AI. Forecasts now suggest that global copper consumption could rise by roughly 50% by 2040, with data centres emerging as a persistent new demand category alongside electrification and renewable energy. From our perspective, this matters because AI infrastructure differs from previous technology waves: it requires continuous expansion, frequent upgrades, and dense electrical architectures, locking in long-term metal intensity rather than one-off build cycles.
Jessica Larn, who focuses on macro-level technology policy and infrastructure dynamics, views the Amazon–Rio Tinto link as part of a wider reclassification of raw materials. In her assessment, copper is moving from a cyclical industrial commodity toward an infrastructure asset, where security of supply increasingly outweighs spot-price optimisation. That shift explains why technology companies are beginning to align directly with producers rather than relying solely on intermediated markets.
Rio Tinto’s Nuton process itself adds another layer to the story. By using naturally occurring bacteria to extract copper from lower-grade ore, the company is targeting resources that were previously uneconomic or environmentally contentious. At YourNewsClub, we see leaching not as a silver bullet for supply shortages, but as a way to stretch existing reserves in a world where new large-scale copper discoveries are becoming rarer and slower to develop.
The experience of other U.S. producers reinforces this view. Leaching has already become a material contributor to domestic copper output, with volumes expected to rise significantly over the next decade. Yet even optimistic expansion scenarios fall short of closing the gap implied by AI growth and the energy transition. The implication is structural tension rather than a temporary imbalance. Copper prices reflect this reality. Trading above $13,000 per metric tonne after a steep rise over the past year, the market is pricing in sustained scarcity rather than short-term disruption. At YourNewsClub, we interpret this as an early signal of a longer re-rating cycle, where metals essential to digital and energy infrastructure command persistent premiums due to constrained supply elasticity.
Owen Radner, whose work examines digital infrastructure as energy–information transport systems, frames the issue more starkly. In his view, AI data centres should be understood as physical utilities disguised as software businesses. Their expansion directly translates into pressure on grids, mines, and material flows. When those flows tighten, strategic alliances between technology firms and extractive industries become almost inevitable.
Taken together, the Rio Tinto–Amazon agreement illustrates a broader realignment. AI’s growth trajectory is no longer determined solely by algorithms or chips, but by access to copper, power, and long-dated industrial processes. At Your News Club, we see this as a defining constraint for the next phase of AI infrastructure: companies that secure upstream resources early gain resilience, while those that treat materials as an afterthought face rising costs and operational risk. The race for intelligence, it turns out, is also a race for metal.