A considerable body of literature has demonstrated – empirically as well as analytically – that information systems need to be situated to the local context of use. Yet, for infrastructural information systems that span numerous contexts spread out globally, this is literally prohibitive. For these systems to work, it is necessary to strike a balance between sensitiveness to local contexts and a need to standardize across contexts. We analyze a key element in this, namely spelling out the (largely invisible) “costs” that the different actors pay to achieve working solutions. Empirically, we draw from an ongoing case study. We analyze the attempts of a maritime classification company with 5500 employees located in 300 sites in 100 countries to develop an infrastructural information system to support the surveying of ships globally. We elaborate design implications and concepts relevant to developing information infrastructures that also apply to the context of developing countries.