Canada's siting process for spent nuclear fuel, led by the Nuclear Waste Management Organization (NWMO), is frequently held within nuclear industry spheres as an exemplary siting process, designed to be inclusive, participatory, and “community-driven.” Drawing from ethnographic observations of the process as it unfolded in Southern Ontario, Canada, this paper focuses on the epistemic issues of how diverse knowledges are treated in the process, whose knowledge is valued, how such knowledges are understood, and whose knowledges are excluded. In particular, I make sense of how epistemic tensions in the process are produced by being situated within a nuclear landscape, informed by local nuclear-dominant socio-technical relations and epistemic regimes, which exceptionalize pro-nuclear Western scientific knowledges. This socio-technical constellation, I suggest, leads to careful but sometimes paradoxical negotiations of the expert/lay divide that subsequently reveals cracks in the policy foundation for inclusion of diverse forms of knowledge. While the NWMO policy framework discursively values diverse knowledges, critical lay community knowledges are often delegitimized and dismissed. Similarly, there are scalar issues in the ways Indigenous knowledges are homogenized and devalued through discursive separation. These epistemic tensions, between how knowledges should be treated in policy, and how knowledges are actually treated in practice, demonstrate clear issues of recognition justice, participatory fairness, and inclusion of diverse knowledges. The implications of this work shed light on understanding the complexities of landscape-based knowledge politics and how they might inform siting practices and technological decision-making more broadly.