Objective: Despite current diagnostic systems distinguishing schizophrenia (SZ) and bipolar disorder (BD) as separate diseases, emerging evidence suggests they share a number of clinical and epidemiological features, such as increased cardiovascular disease (CVD) risk. It is not well understood if poor cardiac autonomic nervous system regulation, which can be indexed non‐invasively by the calculation of heart rate variability (HRV), contributes to these common CVD risk factors in both diseases. Method: We calculated HRV in 47 patients with SZ, 33 patients with BD and 212 healthy controls. Measures of symptom severity were also collected from the patient groups. Results: Heart rate variability was significantly reduced in both these disorders in comparison with the healthy participants; however, there were no HRV differences between disorders. Importantly, these reductions were independent of the medication, age or body mass index effects. There was also preliminary evidence that patients with reduced HRV had increased overall and negative psychosis symptom severity regardless of SZ or BD diagnosis. Conclusion: We suggest that HRV may provide a possible biomarker of CVD risk and symptom severity in severe mental illness. Thus, our results highlight the importance of cardiometabolic screening across SZ and bipolar spectrum disorders.