Purpose The three-minute all-out test (3MT), when performed on a laboratory ergometer in a linear mode, can be used to estimate the heavy–severe-intensity transition, or maximum metabolic steady state (MMSS), using the end-test power output. As the 3MT only requires accurate measurement of power output and time, it is possible the 3MT could be used in remote settings using personal equipment without supervision for quantification of MMSS. Methods The aim of the present investigation was to determine the reliability and validity of remotely performed 3MTs (3MTR) for estimation of MMSS. Accordingly, 53 trained cyclists and triathletes were recruited to perform one familiarisation and two experimental 3MTR trials to determine its reliability. A sub-group (N = 10) was recruited to perform three-to-five 30 min laboratory-based constant-work rate trials following completion of one familiarisation and two experimental 3MTR trials. Expired gases were collected throughout constant-work rate trials and blood lactate concentration was measured at 10 and 30 min to determine the highest power output at which steady-state VO2 (MMSS-VO2) and blood lactate (MMSS-[La−]) were achieved. Results The 3MTR end-test power (EPremote) was reliable (coefficient of variation, 4.5% [95% confidence limits, 3.7, 5.5%]), but overestimated MMSS (EPremote, 283 ± 51 W; MMSS-VO2, 241 ± 46 W, P = 0.0003; MMSS-[La−], 237 ± 47 W, P = 0.0003). This may have been due to failure to deplete the finite work capacity above MMSS during the 3MTR. Conclusion These results suggest that the 3MTR should not be used to estimate MMSS in endurance-trained cyclists.