Objectives: The risk of diabetes mellitus (DM) is increased in people with HIV. Chronic inflammation contributes to DM risk. High leukocytes are associated with DM in the general population, but in people with HIV, evidence of a leukocyte-DM association is limited. Methods: We included participants of the Swiss HIV Cohort Study with new-onset DM (2000-2023) and matched controls. We obtained uni- and multivariable odds ratios (ORs) for DM, based on traditional and HIV-related DM risk factors, leukocyte count and potential confounders for leukocyte count. Results: Among 732 DM cases (median age 54 years, 79% male at birth, 86% with suppressed HIV RNA) and 2032 DM-free controls, the latest leukocyte count prior to DM diagnosis was higher in cases than in controls (median [interquartile range], 6200 [5115-7800] vs. 5900/mu L [4800-7180]; p < 0.001), but leucocytosis (>11 000/mu L) was uncommon (3.8% vs. 2.3%; p = 0.032). DM-OR in the highest vs. lowest leukocyte quintile was 1.91 (95% confidence interval, 1.45-2.52) in univariable analysis and 2.47 (1.71-3.57) in multivariable analysis. For comparison, multivariable DM-OR for dyslipidaemia, overweight, >= 12-month stavudine exposure and >= 12-month integrase inhibitor exposure were 2.23 (1.80-2.75), 2.83 (2.21-3.62), 1.54 (1.16-2.04) and 2.29 (1.13-4.62), respectively. We found no relevant confounders for the leukocyte-DM association. Leukocytes were significantly associated with DM up to 10 years before diagnosis (all p < 0.02). Conclusions: High leukocyte count is an independent DM risk factor in people with HIV in Switzerland and increases the risk of DM to a degree similar to traditional and HIV-related risk factors, up to 10 years before DM diagnosis.