The paper argues a disequilibrium theoretical model of the expansion of transportation infrastructure, which is concordant with J. Schumpeter’s theory of economic development and M. Hirooka’s application of this theory. The model describes a long-term growth of transport networks as dependent on the lagged rates of utilization of an aggregated resource by the transport infrastructure and satisfying the transportation demand. The arguments in favor of this model, which are available in the earlier literature, are strengthened with testing three new empirical specifications of 1871-2013 data about surface transport networks in the USA. These specifications address the hypotheses about the adequacy of the chosen proxies of the exogenous factors, influence of intensive factors of network expansion, influence of other transport networks on the development of roadway and railroad networks, influence of the varying structure of the aggregated resource. The data support the proposed model. The key theoretical contribution of this study is the revealed regularities in the expansion of transport infrastructure, which are stable during the periods as long as a century; rejection of influence on this expansion of long-term investment lag and support for the Hirooka’s concept of the primacy of infrastructure expansion in relation to Kondratieff-like waves. The methodological contribution is the development and approval of workarounds for obstacles to testing the proposed model. The ability to use the parameter of resource to service conversion, which is proved stable during long periods, for the purpose of assessing the outlook of emerging new and ousting existing transport networks is of practical relevance.