Background: Relatively little is known about physiological cerebrovascular haemodynamics during physical stress in elderly healthy individuals. The aim of this study was to determine the effect of ergometer stress on cerebrovascular haemodynamics in elderly healthy individuals in comparison with young healthy individuals, using non-invasive methods. Methods: Continuous middle cerebral artery blood flow velocity (CBFV; transcranial Doppler ultrasound), beat-to-beat blood pressure, heart rate and transcutaneous pCO2 were measured in response to 3 min ergometer exercise stress in 18 elderly healthy subjects (mean age ± SD 66.5 ± 5.8 years) and 18 healthy young subjects (mean age ± SD 29.4 ± 4.7 years). Pulsatility index (PI) was used as a parameter for cerebrovascular resistance. The subjects were in a supine position with an elevated trunk and performed exercise by pedalling on an ergometer, generating 75–100 W. Statistical analysis was carried out using MANOVA, a general linear model with repeated measures. Results: In both groups, blood pressure increased significantly (p < 0.001) with time during exercise, with no significant differences between the groups or regarding interaction (time sequence/group factor). Heart rate increased significantly with time during exercise (p < 0.001) and was significantly more prominent (p = 0.002) and prolonged (p < 0.001) in the young group. pCO2 did not differ with time or between the groups and with regard to interaction. Mean CBFV (MFV) increased significantly during time (p < 0.001). Between the groups, there was no significant difference (p = 0.836), but with regard to interaction (time sequence/group factor), there was a significant delay in MFV increase in the group of young subjects (p = 0.002). The PI, a measure of cerebrovascular resistance, increased significantly with time without significant differences between the groups (p = 0.061), but was significantly delayed in the elderly regarding the interaction time sequence/group factor (p < 0.001). Conclusion: The cerebrovascular changes during ergometer exercise may reflect the combined activation of the cerebrovascular autoregulative mechanisms (predominantly neurogenic and myogenic). In healthy normotensive elderly subjects, cerebral autoregulatory capacity is retained but delayed in response to ergometer stress compared with young healthy subjects. We speculate that these findings may contribute to a higher risk of cerebral hypoperfusion in the elderly.