Laser scanning is the de-facto system to collect large amounts of geometrical information, especially indoors. Total stations are usually used for accuracy and information-rich results. However, in some cases the target area is not suitable for using a terrestrial laser scanner. For instance, too cramped or difficult-to-access space can be a limiting factor. In complex cases, standard scanning methods are usually time-consuming, leading to unexpected costs and sometimes scanning can be omitted for time and cost reasons. Simultaneous location and mapping (SLAM)-based rapid laser scanning could be considered as a practical method to perform scanning in poorly accessible spaces. In this paper, we present results obtained using a ZEB1 handheld SLAM-based laser scanner to ~1440 m depth in a mine shaft in Finland while standing on the roof of the moving elevator, for very fast data collection in an extreme environment. In a scanning time of approximately one hour, we collected and archived 120 million data points. Compared with standard tripod lasers, the accuracy was low, but could be improved using positioned targets. The method also gave an imperfect model which required manual correction. Considering the rapid maturation of SLAM-based systems, the method can be seen a promising tool to collect laser point data in difficult-to-access areas in the near future for maintenance model applications. These systems already work well enough when speed is required over accuracy and in cases where the environment is too difficult to use standard total stations, as in the case presented in this paper.