Background: Type 2 diabetes (T2D) has been associated with reduced abdominal aortic aneurysm (AAA) risk and slower growth in observational studies, but the mechanism for this paradoxical association remains unclear. We hypothesized that T2D's mechanistic heterogeneity obscures opposing, pathway-specific effects on AAA. Methods: We performed two-sample Mendelian Randomization (MR) to investigate the effect of T2D (T2DGGI, 242,283 individuals with T2D and 1,569,730 without) on AAA (AAAgen, 37,214 individuals with AAA and 997,456 without) using 449 genetic instruments. We also performed MR within previously defined clusters representing different mechanisms of T2D. We extended the analysis to individual hyperglycemic, insulin-resistance, and insulin-sensitivity traits. Our hypothesis was tested in a porcine pancreatic elastase murine model of abdominal aortic aneurysm formation. Results: Overall T2D liability showed no net association with AAA liability (OR 1.00, 95% CI 0.96-1.04, P = 0.90), albeit with marked instrument heterogeneity (Cochran's Q = 1356, P = 5 x 10-92) and inconsistency across sensitivity analyses, with a notably protective MR-Egger result (OR 0.87, 95% CI 0.80-0.94, P = 0.0011) and non-zero intercept (P = 2 x 10-4). MR of T2D on AAA within mechanistic clusters showed opposing effects: lipodystrophy (OR 1.47, 95% CI 1.26-1.72) and obesity (OR 1.27, 95% CI 1.18-1.36) cluster membership associated with increased AAA risk, while beta-cell negative proinsulin (OR 0.83, 95% CI 0.76-0.90) and beta-cell positive proinsulin (OR 0.86, 95% CI 0.80-0.93) clusters were protective. This pattern was robust to clustering method. Traits indexing hyperglycemia (fasting glucose, 2-hour glucose, hemoglobin A1c) were associated with reduced AAA risk, while traits indexing insulin resistance (fasting insulin, BMI, waist-hip ratio, triglyceride-to-HDL ratio) were associated with increased AAA risk. In mice, those treated with a high fructose diet as a model for insulin resistance showed increased growth of the abdominal aorta compared with controls, whereas those treated with pancreatoxin streptozotocin as a model for isolated hyperglycemia showed blunted growth. Conclusion: Distinct components of T2D physiology may have differential effects on AAA development. Whereas insulin resistance may promote AAA, hyperglycemia appears to protect against it. Our findings support insulin sensitization as a rational direction for AAA pharmacotherapy and suggest that the pathways mediating the protective effect of hyperglycemia may harbor additional therapeutic targets.