In the age of Big Open Linked Data (BOLD), we inhabit a landscape where future scenarios are imagined, modeled, planned for and embedded in policy. Between the euphoric techno-utopian rhetoric of the boundless potential of BOLD innovations and the dystopian view of the dangers of such innovations (e.g. ubiquitous surveillance etc.), this paper offers a critical understanding of the boundaries that are traversed by the implementation of BOLD within policy modeling. We examine BOLD as a tool for imagining futures, for reducing uncertainties, for providing legitimacy and for concentrating power. In doing so we further develop the LIMITs community’s conceptualization of the societal limitations on computing, with specific reference to the assumptions, interpretations and trust that we place in these models when making socio-environmental policy decisions. We use an illustrative case of policy modeling, which provides a much-needed critical discussion of the inherent limitations and risks as well as the promises that are offered by BOLD.