BHP, the world's largest mining company by market cap, warns the rise of data centers and AI could exacerbate a looming shortage of copper by increasing demand
World's largest miner expects global demand for red metal will rise by more than 70% by 2050 …
Context & Ripple Effects
BHP places copper alongside power as a physical input exposed by the build-out of AI infrastructure. Later coverage makes the connection more concrete: AI facilities use materially more copper per megawatt than conventional data centers, according to a later estimate of AI data-center copper intensity.
The warning also foreshadows buyers moving closer to supply: AWS later secured access to Rio Tinto’s Arizona mine through a two-year copper supply agreement. That shifts copper from a broad commodity-cost issue toward a potential project-planning constraint for large data-center operators.
First-order effects
- BHP’s outlook strengthens the case for miners and developers to treat incremental copper supply as strategically valuable as AI-related demand expands.
- Data-center builders face another exposed input cost and availability risk, alongside the grid bottlenecks already associated with AI and crypto data-center demand.
Second-order effects
- Large cloud operators may seek longer-term procurement arrangements or mine-linked supply access, while miners gain a clearer commercial rationale for copper investment and expansion.
- Tighter copper availability can raise the cost or extend the timelines of data-center electrical systems, making site selection and power-network build-outs more interdependent.
Third-order effects
- If AI demand persists at the projected scale, infrastructure planning will increasingly be constrained by upstream materials and grid equipment, not just chips, servers and generation capacity.
- The pattern could favor operators able to secure durable power and materials supply chains, though the severity of any shortage will depend on mine development, recycling and demand growth.
The trend: AI infrastructure is turning commodity and utility inputs into strategic constraints that shape where and how quickly compute capacity can be built.