In the wake of the recent proliferation of the securitization of crowded places, there has been a growth in the development of technologies of crowd behaviour analysis. However, despite the emerging prominence of crowd surveillance in contemporary emergency planning, little has been discussed about its impacts on our understanding of security and surveillance. Drawing from the case of crowd surveillance in Tokyo, this article examines the ways in which crowds are simulated, monitored, and secured through the technology of crowd behaviour analysis, and discusses its implications to the politics of security. It argues that crowd surveillance constitutes a unique form of the biopolitics of security that targets, not the individual body or the social body of population, but the urban body of crowd. The power of normalization in crowd surveillance operates in a preemptive manner through the codification of crowd behaviours that is spatially and temporarily specific. The article also critically interrogates the introduction of crowd surveillance in relation to racialized logics of suspicion and argues that, despite its appearance as a non-discriminatory and 'aracial' technology, crowd surveillance entail the racial coding of crowd behaviour and urban space. The article concludes with the introduction of crowd surveillance as a technology of border control, which reorients existing modalities of (in)securitization at airports.