Underground freight transport is a promising means for the regrouping of freight and for the application of Physical Internet concepts to urban freight supply. Such systems were put in order as early as the mid-19th century, and have functioned for several decades until they were eventually shut down. Yet since 1980, despite many promising projects, only two were in fact implemented. We wish to understand the circumstances leading to the discard of urban underground freight transport schemes, with a view to provide explanations to the high failure rate of underground freight. Our work will draw lessons that can be applied to ongoing projects to support their implementation.