When infrastructural technology is functioning correctly, it is often considered to recede from view and become invisible. According to this perspective, visibility is restored in cases of breakdown and malfunction, which for this reason, are often understood to represent important epistemic opportunities for grasping previously hidden aspects of infrastructure. This article seeks to outline the limitations of the idea that infrastructural failure has a positive epistemic function by distinguishing between two fundamentally different ways in which the nature of technological function can be conceptualised; the first understands function as stemming primarily from the way a technology was designed and produced, the second as a process sustained by constant human activities of maintenance and repair. After illustrating how a number of recent studies of infrastructure build upon insights granted by the latter process perspective, this article aims to demonstrate how the experience of breakdown itself does little to facilitate a shift towards understanding infrastructure in this way. For this reason, this article ends with a suggestion for a shift in emphasis in studies of infrastructure, away from breakdown and towards a sustained examination of everyday practices of maintenance, repair and modification of infrastructural technologies.