Purpose: To compare the metabolic cost of paddling on different commercially available kayak ergometers using a standardized kayak incremental exercise protocol. Methods: Six male sprint kayak athletes undertook an incremental exercise protocol on 3 different kayak ergometers utilizing a randomized counterbalanced pair-matched design. Results: Mean maximal aerobic power on the WEBA ergometer (265 [14] W) was significantly higher than on the Dansprint (238 [9] W) and KayakPro® (247 [21] W, P < .01, effect size [ES] = 0.80). At the fifth stage, absolute oxygen consumption on the WEBA (3.82 [0.25] L·min−1) was significantly lower (P < 0.05, ES = 0.20) than KayakPro and Dansprint (4.10 [0.28] and 4.08 [0.27] L·min−1, respectively). Blood lactate concentration response at the sixth stage was significantly lower for the WEBA (3.5 [0.8] mmol·L−1), compared with KayakPro and Dansprint (5.4 [1.2] and 5.6 [1.5] mmol·L−1, P = .012, ES = 0.20). Stroke rate was significantly higher, without any effect of pacing during the submaximal stages for the Dansprint, compared with the WEBA (P < .001, ES = 0.28) and KayakPro (P < .001, ES = 0.38). A pacing effect was present at the maximal stage for all ergometers. Conclusions: This study demonstrated that paddling on different kayak ergometers when controlling power output elicits different metabolic and work outputs. It is recommended that scientists and coaches avoid testing on different ergometers and regularly calibrate these devices. Moreover, when an ergometer has been calibrated against a first principle device, it is necessary to consider calibration of various drag settings, due to their impact on stroke rate. Further research should explore the relationship between drag settings and stroke rate.