In this paper we assess how production costs and capital accumulation patterns in agriculture have evolved over time, by paying special attention to the influence of risk. A dynamic state-contingent cost minimization approach is applied to assess production decisions in US agriculture over the last century. Results suggest the relevance of allowing for the stochastic nature of the production function which permits to capture both the differences in the costs of producing under different states of nature, the differences in the evolution of these costs over time, as well as the differential impacts of different states of nature on investment decisions.