Despite increased attention, investment and research, the security technologies deployed at sovereign borders often fail to ‘properly’ distinguish between safe and dangerous travellers and goods. This paper transposes on-site border malfunctions into the antecedent register of research and development to explore how failure operates in the design, construction and testing of a border security technology. Drawing from extensive ethnographic research of a European-wide project to develop the Handhold portable CBRNE detector, this paper traces how failure is mobilized as an instructive experience in the professional cultures of science, engineering and governance, and retrospectively narrated as a necessary step on the road to eventual success. In exploring the limitations of this understanding, the paper demonstrates how competing experiences of failure constantly trouble the daily routines of scientists and engineers, shape informal de-risking strategies, and fuel the relentless pursuit of innovation. To contest the dominant account of failure as an instructive experience, the paper argues that more creative and critical research into the heterogeneous life-worlds of failure is needed in order to politicize the consequences of border failures and connect them to the antecedent register of scientific research and development.