In this work, we present a game theoretic framework to give insights into these questions. Our model is based on a framework for Cournot competition with congestion effects that has been used in a number of papers to study competition in wireless markets and other setting where users are sensitive to congestion effects. In these models, wireless service providers compete by determining a quantity of customers to serve. The price that they can then charge each customer is determined by a market clearing price as well as a congestion cost that depends on their spectrum holdings and the network infrastructure that they have deployed. This captures the fact the spectrum is a congestible resource and that customers care about quality of service so that they are not willing to pay as much for a network that offers a poorer quality of service. Our model captures several different scale effects. We account for the investment incentives of a service provider as a function of their bandwidth holdings, where a provider with more spectrum may see a larger return from investing in infrastructure. We also consider a congestion model that can capture “multiplexing gains” achieved by serving more users with proportionally more bandwidth. Finally, we also consider how operational cost may scale with an increasing number of users. We use this model to compare the economic welfare obtained under with different number of competing firms and different distributions of bandwidth. Our results show that the “optimal” number of firms can vary depending on these effects suggesting that policy makers need to carefully consider these effects when assessing competition.