<p>Metabolic reprogramming toward aerobic glycolysis, a phenomenon analogous to the Warburg effect, is increasingly recognized as a hallmark of pulmonary arterial hypertension (PAH). However, how glycolytic activation orchestrates pulmonary vascular remodeling and endothelial–mesenchymal transition (EndMT) remains poorly understood. To delineate the molecular determinants of glycolytic remodeling, we integrated bulk and single-cell transcriptomic datasets from hypoxia-induced (SuHx) PAH rat models. A multi-algorithm machine-learning pipeline combining LASSO, BORUTA, and SVM-RFE was employed to identify glycolysis-related hub genes. Diagnostic performance was validated in independent rat and human datasets, followed by single-cell communication and pseudotime trajectory analyses to infer cellular dynamics. Experimental verification was performed using SuHx/MCT rat models and hypoxia-treated HUVECs. Five glycolysis-related biomarkers—Cxcr4, Pfkfb3, Rgcc, MGC105649, and Mt2A—were identified, exhibiting consistent differential expression across datasets. PFKFB3 and MT2A emerged as core regulators that link glycolytic activation with endothelial–fibroblast crosstalk and EndMT. Single-cell and pseudotime analyses revealed that endothelial cells progressively transition toward mesenchymal-like phenotypes under hypoxia, accompanied by enhanced glycolytic signaling. Functional validation confirmed increased expression of PFKFB3 and MT2A in vivo and demonstrated that hypoxia and lactate synergistically promote endothelial proliferation in vitro. This integrative framework combining multi-omics and machine learning uncovers a glycolysis-driven EndMT axis that underlies pulmonary vascular remodeling. The identified biomarkers provide mechanistic insight into metabolic–structural coupling in PAH and highlight PFKFB3 and MT2A as promising molecular targets for future glycolysis-based therapeutic interventions.</p>