Background <p>Atherosclerotic dyslipidemia is associated with an increased risk of type 2 diabetes (T2D). Although previous studies have demonstrated an association between the atherogenic index of plasma (AIP) and insulin resistance, there remains a scarcity of large cohort studies investigating the association between AIP and the long-term risk of T2D in the general population. This study aims to investigate the potential association between AIP and the long-term risk of T2D in individuals with normal fasting plasma glucose levels.</p> Methods <p>This retrospective cohort study included 15,453 participants. The AIP was calculated using the formula log [triglyceride (mmol/L)/high-density lipoprotein cholesterol (mmol/L)]. Cox proportional hazard regression models were employed to assess the association between AIP and T2D risk. The nonlinear association was examined using a restricted cubic spline (RCS) model.</p> Results <p>During an average follow-up period of 6.05 years, 373 participants developed T2D. After adjusting for confounding factors, elevated AIP was independently associated with an increased risk of developing T2D (HR 1.763, 95%CI 1.210–2.568, <i>P</i> = 0.003). The RCS analysis revealed a J-shaped association between AIP and T2D risk, with a sharp increase in risk when AIP levels exceeded − 0.268. Moreover, time-dependent receiver operating characteristic analysis consistently demonstrated a moderate predictability of AIP for new-onset T2D within 1 to 12 years.</p> Conclusion <p>The AIP exhibits a J-shaped association with the risk of developing T2D. Therefore, maintaining AIP levels below a certain threshold (-0.268) might help prevent the onset of T2D.</p> Graphical abstract <p></p>