Cobalt is the canonical example of a material where mining geography (DRC 76%) is decoupled from refining geography (China leading, with processing concentration that USGS acknowledges qualitatively but does not quantify in MCS 2025). Chinese firms own the majority of DRC mine stakes and the refining stage, so the February 2025 DRC export ban addresses the wrong chokepoint. LFP battery adoption is the most important demand-side dynamic — USGS explicitly notes that cobalt-free iron-phosphate chemistries already hold significant market share in China. 2024 was a record-high production year for both mine and refinery output, driving prices down and idling the US Idaho operation.

Within-country analytical limitation (documented, intentionally not modelled): the DRC row at 220,000 t aggregates two very different supply modes. Large-scale industrial operations — principally Chinese-owned, including CMOC's Tenke Fungurume and Kisanfu, plus Glencore's Katanga and Mutanda — account for the bulk of the figure. Artisanal and small-scale mining (ASM), estimated at roughly 15–30% of DRC cobalt output, is the other mode: it raises distinct concerns around child labour, human-rights violations, and opaque routing through Chinese trading houses, and the Entreprise Générale du Cobalt (EGC) monopoly set up to formalize the ASM stream has struggled with volume uptake. Country-level production figures here do not split output by mode, so the DRC row aggregates both the industrial stream and ASM; readers should treat the DRC cobalt figure as a combined total rather than interpreting it as either mode alone.