In response to broader political and corporate tendencies towards ‘techno-solutionism’, critical studies of security technology highlight the threat that security technologies pose to civil rights and democratic accountability. This article argues for a slightly different perspective: rather than taking claims of technological efficacy at face value, it explores the multiple ways in which security-related technology so frequently fails to deliver its – confidently anticipated or feared – effects. A focus on sociotechnical failure can offer more comprehensive, on-the-ground understanding of the technopolitics of security. We suggest that these politics may lie precisely in the blurring of concepts of failure and success, as ‘prototyping’ and experimentation become an increasingly powerful logic of urban governance. This argument is developed through an analysis of security interventions in Jamaica, a context characterized by high levels of violent crime. The article focuses on three technologies that have been adapted to security-related purposes: a communication channel connecting police and private security guards, a public–private CCTV network, and a smart electricity grid. Drawing on approaches from science and technology studies, the article adopts a process-oriented approach, attending to both the discourses surrounding the introduction of these technologies and their everyday interactions with their social and built environments.