Over the past 30 years, there has been a marked proliferation of the use of public–private partnerships (P3s) for the management of agricultural innovation systems. This is part of a larger worldwide trend of using P3s in the provision of public goods and services. Despite the large number of agricultural P3s in operation, a literature review demonstrated paucity of both case studies and of theory, meaning that the study of these emerging business models has not kept pace with practice. Over the last 30 years, only 38 peer-reviewed articles have been published. The objective of this dissertation is to advance the theory, analysis, and policy review of agricultural P3s. There are four independent investigations in this dissertation that advance the knowledge of agricultural P3s in seven specific ways. First, these investigations introduce two quantitative methodologies to empirically demonstrate the critical role P3s occupy in research and development (R&D) innovation networks and in the development, dissemination and commercialization of new technologies that enhance global food security. Second, this analysis suggests that the key variable influencing the formation of these organizations is people, rather than public policy or market incentives. Third, agricultural P3s require large up-front investments and they have extended gestation periods; therefore, they are dependent upon public support. Fourth, P3s are not a means of privatizing public functions; rather, they represent a new and emerging process of collaboration that transcends the public–private dichotomy. Fifth, agricultural P3s appear to operate in “orphan spaces,” sectors that, for a variety of reasons, are ignored by the public and private sectors. Sixth, there is evidence to suggest that many P3s require the services of P3 experts of which there is a shortage, particularly in the developing world. Seventh, each agricultural P3 is novel because each is the result of sector-specific challenges and has a structure that is dependent upon the types and number of partners and their objectives, limiting the ability to transfer explicit lessons from existing models to new P3s.