The AI revolution runs on copper, and the United States is scrambling to secure enough of it before supply shortages choke the buildout of next-generation data centers.
When we talk about artificial intelligence, the conversation usually centers on algorithms, compute clusters, and which startup just raised a billion-dollar round. Rarely do we discuss the physical infrastructure that makes any of it possible. That blind spot is becoming a strategic liability. The surge in AI workloads demands enormous power, and delivering that power requires copper: for transformers, for busbars, for the high-voltage cabling connecting data centers to the grid. We are looking at a scenario where copper supply could become the single greatest constraint on AI growth.
The numbers put this into sharp focus. Market analysts project that AI data centers alone could consume between 500,000 and 1.1 million tonnes of copper annually by 2030. That accounts for nearly three percent of total global demand, a staggering figure for a single sector that barely registered on commodity radars a few years ago. S&P Global forecasts broader copper demand rising 50 percent by 2040, driven by electrification and AI combined. Meanwhile, more than half of all new data center projects currently face potential delays because transformers and electrical equipment simply cannot be manufactured fast enough.
As Bloomberg Technology recently reported, US copper production has stagnated for decades even as demand has accelerated. The consequences are showing up in commodity markets. In March 2026, copper prices surged past $13,000 per tonne, a record high fueled by aggressive purchasing from the US and Europe. The International Copper Association estimates the industry must invest $2.1 trillion over the next 25 years to prevent a chronic shortage from crippling both the AI and green energy sectors.
Arizona sits at the center of this supply crisis. The Resolution Copper project, a joint venture that could become one of the largest copper mines in North America, finally cleared a critical land exchange with the US Forest Service and Rio Tinto in March 2026 after years of regulatory delays. It is the kind of project that Washington needs to treat as urgent. Industry leaders like Freeport-McMoRan are pushing to have copper formally designated a critical mineral, which would unlock federal funding and fast-track permitting under the Defense Production Act. Right now, copper does not receive the same strategic priority as rare earth elements or lithium, despite being arguably more immediately consequential for AI infrastructure.
A Geopolitical Vulnerability
Securing the raw material is only half the battle. The US mines a respectable amount of copper, but it sends significant volumes of raw ore to China for smelting and refining. That dependency is a strategic liability that current administration officials have explicitly acknowledged through new executive orders tying minerals security to reduced reliance on Chinese processing. The $1.8 billion Critical Minerals Fund established in February 2026 includes allocations specifically for building domestic refining capabilities and friendshoring supply chains through partnerships with Brazil and Chile.
What makes this moment different from previous commodity squeezes is the speed at which demand is scaling. AI infrastructure is being built on timelines measured in months, not decades. Mining and refining operations work on the opposite end of that spectrum. You cannot bootstrap a copper mine the way you can spin up cloud compute. The mismatch between AI deployment velocity and minerals production reality is the core tension, and it shows no sign of resolving cleanly.
Environmental opposition complicates the picture further. Projects like Resolution Copper face fierce resistance from groups concerned about ecological damage, including ongoing disputes at Oak Flat in Arizona. These are legitimate debates, but they collide with an equally legitimate urgency. Every month of delay in bringing new supply online tightens the bottleneck.
For startups building AI infrastructure, investors allocating capital to data center plays, and policymakers weighing mineral security, the message is straightforward. Compute is not the constraint anymore. Power delivery is. And power delivery runs on copper. The companies and countries that solve the supply chain equation fastest will dictate the pace of AI development for the next decade. Watch for copper's critical mineral designation to become a decisive policy fight in the months ahead.
