Competitive cyclists typically sprint out of the saddle and alternately lean their bikes from side-to-side, away from the downstroke pedal. Yet, there is no direct evidence as to whether leaning the bicycle, or conversely, attempting to minimize lean, affects maximal power output during sprint cycling. Here, we modified a cycling ergometer so that it can lean from side-to-side but can also be locked to prevent lean. This modified ergometer made it possible to compare maximal 1-s crank power during non-seated, sprint cycling under three different conditions: locked (no lean), ad libitum lean, and minimal lean. We found that leaning the ergometer ad libitum did not enhance maximal 1-s crank power compared to a locked condition. However, trying to minimize ergometer lean decreased maximal 1-s crank power by an average of 5% compared to leaning ad libitum. IMU-derived measures of ergometer lean provided evidence that, on average, during the ad-lib condition, subjects leaned the ergometer away from the downstroke pedal as in overground cycling. This suggests that our ergometer provides a suitable emulation of bicycle lean dynamics. Overall, we find that leaning a cycle ergometer ad libitum does not enhance maximal power output, and conversely, trying to minimize lean impairs maximal power output.