The recent enthusiasm for 'big data' (Data for Development, D4D) highlights the increasing interpenetration of public social policy and data from private technology firms. This article examines the process in reverse, whereby public information systems become the basis for increasing financial intermediation. Drawing on ethnographic research on Paraguayan microinsurance programmes, this article describes the broader mandate of 'financial inclusion' embedded in the suite of banking services that support the state's conditional cash transfer welfare initiative, Tekoporã. The article analyses the process by which inclusive insurance infrastructures are built and maintained as an opportunity to critique the political economy of data production that supports financialization. The concept of 'risky data' describes not only the way that the social lives and relations of the poor are quantified and converted into financial knowledge, but also the way that the urgent new mandate to produce such data further exposes public policy and state institutions to new risks to their legitimacy. The Paraguayan case exemplifies the key mechanism of 'risky data' for inclusive infrastructures: the extraction of financial value and speculative profits from state-led investment in a public resource, which amplifies the risks of both service denial and data surveillance. This research was funded by the Australian Research Council Discovery Early Career Researcher Award grant number DE170101406. I would like to thank the research team in Paraguay consisting of Lic. Betania Ayala and Ing. Analia Melo. I received generous feedback on this material from the London School of Economics Financialization seminar series hosted by Kate Meagher and the Department of Anthropology seminar series at the Chinese University of Hong Kong. I would also like to thank Sverre Molland for his review of an early draft, and the editorial board of Development and Change for their very helpful suggestions.