The current study investigates the predictive ability of critical minerals of price returns of precious metals (gold, silver, platinum, and palladium) in presence of different degrees of climate policy uncertainty (CPU). Using a novel Multivariate Quantile-on-Quantile Causality (MQQC) model, we test the predictive dynamics, unconditional and CPU-conditional, in the entire joint return distribution continuum. Predictability, unconditionally, is localized in the tails, i.e. under extreme market conditions, mineral shocks have strong impact but under normal regimes, they have little impact. The tail dependence is indicative of co-production and industrial-demand relationships of silver, platinum, and palladium, but gold mostly maintains its safe-haven property. After the addition of CPU, the predictive effects are stretched out further to the middle of the distribution, indicating wider and more enduring spillovers. In the case of gold, CPU enhances the safety haven demand by augmenting the crucial mineral precious metal co-movements between regimes. In the case of silver, platinum and palladium, CPU increases industrial sensitivities relating to clean-energy use. These findings highlight the twofold contribution of the geological factors in conjunction with policy uncertainty towards price fluctuations, and significance of the findings on resource planning, governance and risk management.