Knowledge is an integral part of verification regimes. Knowledge cannot be seen as pure factual truth that arises directly from data, but has to be analysed as aggregated interpretations that stem from technical data and political judgement as well. These interpretations are related to specific, verification regime-related epistemic communities; however, such interpretations are typically unquestioned if the overall regime works well. As the produced knowledge is at the very basis of decision-making, its quality and the production process beyond the technical realm must be of interest: technologies used for knowledge production are embedded in wide-ranging production chains – so-called knowledge infrastructures. These infrastructures are institutions that regulate and organise the production, distribution and use of knowledge by setting gathering and processing standards. Within these infrastructures, practitioners interpret the incoming flow of information through constant interaction. As the practitioners shape not only the knowledge but also the infrastructure itself, a systematic epistemic analysis of knowledge infrastructures and key actors therein, such as the IAEA, can provide a reflexive element to uncover biases and blind spots (also referred to as non-knowledge). Our contribution aims at elucidating the challenges of knowing and deciding within knowledge infrastructures, especially concerning the influx of new kinds of information and constantly evolving verification technologies.