Models of endogenous growth have not been able to account for the variety of empirically observed distributional properties of the returns to innovation, in part, because of the limitations necessarily imposed on competition to cope with increasing returns to scale. Exponential growth, fat tails, Pareto–Levy distributed upper tails, and upper value outliers, are associated with increasing returns to scale and innovation. At the same time, properties such as bifurcated research investment strategies, bimodal returns to innovation, and Laplace distributed firm growth rates are products of competition. We build an agent-based model of endogenous technical change in which heterogeneous investments in patented knowledge and increasing returns to scale emerge these distributional properties within a competitive market. The combination of heterogeneous agents, costly information, and patents allow for a competitive landscape to persist amidst increasing returns. The ability of model to foster a coexistence of competition and increasing returns underlies the observed distributional properties.