Abstract Aims/hypothesis People with diabetes have an increased cardiovascular risk with an accelerated development of atherosclerosis and an elevated mortality rate after myocardial infarction. Therefore, cardioprotective effects of glucose-lowering therapies are of major importance for the pharmacotherapy of individuals with type 2 diabetes. For sodium–glucose cotransporter 2 inhibitors (SGLT2is), in addition to a reduction in blood glucose, beneficial effects on atherosclerosis, obesity, renal function and blood pressure have been observed. Recent results showed a reduced risk of worsening heart failure and cardiovascular deaths under dapagliflozin treatment irrespective of the diabetic state. However, the underlying mechanisms are yet unknown. Platelets are known drivers of atherosclerosis and atherothrombosis and disturbed platelet activation has also been suggested to occur in type 2 diabetes. Therefore, the present study investigates the impact of the SGLT2i dapagliflozin on the interplay between platelets and inflammation in atherogenesis. Methods Male, 8-week-old LDL-receptor-deficient (Ldlr−/−) mice received a high-fat, high-sucrose diabetogenic diet supplemented without (control) or with dapagliflozin (5 mg/kg body weight per day) for two time periods: 8 and 25 weeks. In a first translational approach, eight healthy volunteers received 10 mg dapagliflozin/day for 4 weeks. Results Dapagliflozin treatment ameliorated atherosclerotic lesion development, reduced circulating platelet–leucocyte aggregates (glycoprotein [GP]Ib+CD45+: 29.40 ± 5.94 vs 17.00 ± 5.69 cells, p < 0.01; GPIb+lymphocyte antigen 6 complex, locus G+ (Ly6G): 8.00 ± 2.45 vs 4.33 ± 1.75 cells, p < 0.05) and decreased aortic macrophage infiltration (1.31 ± 0.62 vs 0.70 ± 0.58 ×103 cells/aorta, p < 0.01). Deeper analysis revealed that dapagliflozin decreased activated CD62P-positive platelets in Ldlr−/− mice fed a diabetogenic diet (3.78 ± 1.20% vs 2.83 ± 1.06%, p < 0.01) without affecting bleeding time (85.29 ± 37.27 vs 89.25 ± 16.26 s, p = 0.78). While blood glucose was only moderately affected, dapagliflozin further reduced endogenous thrombin generation (581.4 ± 194.6 nmol/l × min) × 10−9 thrombin vs 254.1 ± 106.4 (nmol/l × min) × 10−9 thrombin), thereby decreasing one of the most important platelet activators. We observed a direct inhibitory effect of dapagliflozin on isolated platelets. In addition, dapagliflozin increased HDL-cholesterol levels. Importantly, higher HDL-cholesterol levels (1.70 ± 0.58 vs 3.15 ± 1.67 mmol/l, p < 0.01) likely contribute to dapagliflozin-mediated inhibition of platelet activation and thrombin generation. Accordingly, in line with the results in mice, treatment with dapagliflozin lowered CD62P-positive platelet counts in humans after stimulation by collagen-related peptide (CRP; 88.13 ± 5.37% of platelets vs 77.59 ± 10.70%, p < 0.05) or thrombin receptor activator peptide-6 (TRAP-6; 44.23 ± 15.54% vs 28.96 ± 11.41%, p < 0.01) without affecting haemostasis. Conclusions/interpretation We demonstrate that dapagliflozin-mediated atheroprotection in mice is driven by elevated HDL-cholesterol and ameliorated thrombin–platelet-mediated inflammation without interfering with haemostasis. This glucose-independent mechanism likely contributes to dapagliflozin’s beneficial cardiovascular risk profile. Graphical abstract