Critical Minerals in AI Infrastructure
Building and operating AI at scale requires enormous quantities of copper, silicon, gallium, germanium, silver, and tin — concentrated in power delivery, cooling, chip fabrication, and high-bandwidth networking.
AI compute is a mineral-intensive enterprise.
A single hyperscale AI data center campus can consume more power than a small city. At gigawatt scale, these facilities require massive copper busbars, thousands of power transformers, kilometers of fiber optic cabling, and millions of solder joints — all underpinned by critical mineral inputs.
The shift from air cooling to liquid cooling for AI accelerators increases copper demand per rack. GaN power semiconductors enable higher efficiency at the power conversion stages, but depend on gallium — a mineral subject to export controls since August 2023.
AI Infrastructure Minerals
AI supply chains intersect with trade restrictions.
Gallium and germanium — critical to GaN power semiconductors and fiber optic networking — have been subject to Chinese export licensing requirements since August 2023. These controls directly affect the procurement of components used in every AI data center.
Diversifying gallium supply is technically feasible but capital-intensive. Primary gallium is a byproduct of aluminum refining, and new extraction circuits require significant modification of existing alumina processing facilities.