A system design game is a model of a situation in which agents’ actions determine the structure of a system, which in turn affects the system’s value and the share of value that each agent may capture through bargaining or market competition. This paper describes a class of games in which agents design interdependent products, for example software programs, which may be complements or substitutes for each other. These relationships are epresented by an object called a design structure network (DSN). Depending on the modeler’s choice of allocation rules, agents may benefit from owning critical nodes in the DSN, corresponding to the type of architectural control enjoyed by firms like Microsoft and Intel. Preliminary experiments confirm that boundedly rational agents tend to capture more value than a agents that take random actions, and that different allocation rules yield different patterns of agent behavior and network structure. Future work will use this framework to explore such phenomena as the emergence of technology platforms and the incentives to innovate when market power is diffuse.