The aim of this paper is to investigate the notion of need, in this case an entire city's global need for water. This was a notion invented by Paris technicians between 1760 and 1804 in the context of several water supply projects, notably two river diversion schemes, those of the Yvette and the Ourcq, where the concept was much discussed. Different ways of considering the question of need – such as water resources, consumption and use, whether present or future – were strongly related to engineers' or scientists' conceptions of their own work. State engineers claimed they could make objective estimates of future needs with no reference to either value judgements or political intentions, a position which made it possible to keep strong control over the decision in the name of the state. In contrast, a practically trained engineer working outside the state corps claimed that in the case of expertise about the future, estimates would depend strongly on political intentions, norms and ideals, so the government should first give an outline of its intended actions. The paper studies the differences between these two approaches to the concept of need, especially how they articulated knowledge about what is and knowledge about what ought to be, present and future. The paper ends by linking these differences to conceptions of what was supposed to be technical or political in such projects and of what role engineers intended to play in the decision-making process.