Abstract: High-resolution simulations are essential to resolve fine-scale air pollution patterns due to localized emissions, nonlinear chemical feedbacks, and complex meteorology. However, high-resolution global simulations of air quality remain rare, especially of the Global South. Here, we exploit recent developments to the GEOS-Chem community model in its high performance implementation (GCHP) to conduct one-year simulations in 2015 at cubed-sphere C360 (~ 25km) and C48 (~ 200km) resolutions. We investigate the resolution dependence of population exposure and sectoral contributions to surface PM_2.5 and NO₂ focusing on understudied regions. Our results indicate pronounced spatial heterogeneity with global mean population-weighted normalized root mean square error (PW-NRMSE) at C48 of for primary (50% - 105%) and secondary (26% - 36%) PM_2.5 species. Under-represented regions are more sensitive to spatial resolution resulting from sparse pollution hotspots, with PW-NRMSE for PM_2.5 in the Global South (34%) 1.3 times higher than globally (25%). The spatial heterogeneity in southern cities (50%) is substantially higher than the more typically clustered northern cities (27%). High-resolution simulations also change the relative importance of emission sectors for both black carbon and NO₂ in the Global South. Overall, spatial gradients of population exposure and sectoral contributions are artificially reduced with coarse simulations, especially in the Global South.