Securitization theory has long aimed to explain how security problems come into existence. Yet, limited by the notion that security is only articulated in an institutional voice by the elites (Wæver 1995:57), the modern processes of securitization encouraged by the proliferation of social media have been understudied. Contending that securitization develops in non-institutional online spheres like Twitter, this article develops a processual refinement that reconceptualizes the production of (in)security as an activity of connecting (Mützel 2009) and depicts securitization as the process of constructing networked structures of meaning. The study explores this puzzle using a text-as-data approach to analyse over 10 million Twitter messages shared by five actor groups in the months leading up to the 45th G7 summit: politicians and governmental institutions, the media, advocates, politicians' Twitter friends, and citizens on Twitter. After developing a securitization-desecuritization dictionary, the study employs a semi-supervised semantic scaling model to position each tweet on a unidimensional scale and identify the securitizing moves and agents. The study then traces the chain of associations between securitizing actors and performs social network analyses to discover the evolutional steps of the formation process of (in)security. The results demonstrate that securitization develops on Twitter as networks of repeated, mundane, routinized, and banal securitizing acts produced by elite and non-elite actors. Moreover, the study reveals that quantitative text analysis and social network analysis tools are instrumental in producing reliable indicators of securitization.