AI Supply Chain Threatened by Critical Minerals and Model Dependence

Artificial intelligence is driving a massive wave of investment in data centres, advanced semiconductors and cloud infrastructure. While attention has focused on chips, the next shortage could involve the strategic minerals required to build them. Copper, essential for wiring, printed‑circuit boards and power grids, is already outpacing new mine development, and gallium – dominated by China – is a chokepoint for high‑frequency and power‑electronics components. These materials, sourced from a handful of countries, are becoming a geopolitical bottleneck for AI‑driven growth.

At the same time, a new form of supply‑chain risk is emerging from the AI models themselves. Open‑source models such as the Chinese‑developed Qwen are gaining rapid adoption, with their share of fine‑tuned derivatives projected to rise dramatically. While open models reduce reliance on a few closed‑model providers, they create ecosystem gravity, locking enterprises into specific model families and raising concerns over future licensing, governance and geopolitical availability. Together, critical minerals and model dependencies could constrain the expansion of AI technologies worldwide.