Machine learning analysis of 2,979 NHANES participants identified four distinct exposure clusters based on urinary biomarkers of combustion pollutants—polycyclic aromatic hydrocarbons and volatile organic compounds. The "High combustion" cluster, representing 5.1 million US adults, showed 39.3% hypertension prevalence versus 28.7% in low-exposure groups, translating to 38% higher odds of hypertension after demographic adjustment. This represents a significant departure from traditional single-chemical analyses that miss the complex co-exposure patterns people actually experience. The findings reveal troubling health equity implications, with Non-Hispanic Black participants comprising 40% of the high-risk cluster despite smaller overall representation. This clustering approach offers a more realistic assessment of environmental health burden by capturing how multiple pollutants interact in real-world exposure scenarios. However, this preprint awaits peer review and results may change. The cross-sectional design limits causal inference, and the 0.849 AUC suggests room for model improvement. While confirmatory of combustion pollutants' cardiovascular risks, the machine learning methodology represents an incremental advance in exposome research that could reshape how we assess environmental health disparities.