Anti-microbial resistance (AMR), or the ability of microbes to resist the effects of drugs designed to kill them or inhibit their growth, has taken a central stage within global health institutions, with predictions of resistant microbes spreading ever more widely in the coming years. The making of evidence that forms the basis of such warnings has received little attention. Despite AMR’s inherently context-dependent nature - local environments provide not only varying possibilities of diagnosis and care, but they also mould the very microbes and their resistance-encoding genes -, AMR continues to be framed and evidenced as a coherent condition. Building on scholarship in medical anthropology and science and technology studies on global health research, this thesis explores how evidence on AMR is made. Empirically, the work is based on three years of non-continuous ethnographic research conducted between 2020 and 2023, in-person and remotely in Uganda and the UK. In both places, I engaged with researchers working on AMR as part of a transnational global health research collaboration, here called D-Lab. I also carried out documentary analysis of their published and unpublished materials and conducted interviews with experts on infection prevention in one of the project’s field sites. Paying particular attention to the perspective of Ugandan research staff, I analyse and contextualise the workings of this transnational research collaboration. The evidence on AMR that D-Lab produced, reflects the set-up of the research collaboration, and broader structures within global health research. D-Lab’s research apparatus did not pick up on the local social, historical, and political conditions of the place in which the research unfolded. The hierarchical set-up of D-Lab additionally did not invite reflection of the Ugandan research staff on the project’s methods, practices, and tools, and their relation to the project's field sites. Overall, I argue that by attempting to produce a global, coherent object of AMR that can be managed through existing global health policy frameworks, D-Lab, and global health research more generally, obfuscates the very conditions that shape AMR.