Cardiac implanted electronic devices (CIEDs), including pacemakers and implantable defibrillators that perform atrial sensing typically using an atrial electrode, frequently detect subclinical atrial high-rate episodes (AHREs). When the intracardiac electrograms are carefully examined, the majority of AHREs are atrial fibrillation (AF) or other atrial tachyarrhythmias, which have been shown to be associated with both an increased risk of stroke, and subsequent development of clinical AF. However, the absolute risk of stroke among patients with AHREs is less than might be expected for clinically diagnosed paroxysmal AF. In addition, a close temporal relationship between AHREs and stroke is seen in only 15% of strokes in patients with a CIED: the majority have either no AHREs before the stroke, or AHREs very distant from incident stroke, suggesting that AHREs might be more of a risk marker than a risk factor for stroke. Management of AHREs should not be the same as for clinical AF, and a degree of uncertainty underpins the rationale for much-needed, ongoing, randomized trials of oral anticoagulation in patients with CIED-detected AHREs. We propose a management algorithm that takes into account both the stroke risk and the AHRE burden, but highlights the current uncertainty and evidence gaps for this condition.