This paper embeds the canonical model of endogenous growth with product proliferation developed by Romer [Romer, P.M., 1990. Endogenous technical change. Journal of Political Economy 98, S71–S102] into a simple urban framework. This yields a reduced form isomorphic to the popular statistical device developed by Simon [Simon, H., 1955. On a class of skew distribution functions. Biometrika 42, 425–440], which in turn can yield Zipf's law for cities. The stochastic outcomes of purposeful innovation and local spillovers can thus serve as foundations for random growth models.