Governments worldwide increasingly partner with diverse sets of stakeholders in delivering public policy. This spreads commercial risk, but also, particularly where agendas diverge, introduces new risks related to trust in relationships. Urban development is an example of a sector where partnerships are used to spread risk. Urban development is also a domain where already divergent agendas are spread further by requirements to build for climate change. The ‘risk hypothesis’ distinguishes between networks for ‘cooperation’ problems, where partners have high individual payoffs for uncooperative behaviours, and ‘coordination’ problems, where partners subscribe to a common goal and uncooperative behaviours are less rewarding. We use mixed-methods to study networks of local and state government, developers, and consultants that centre on joint-venture partnerships for developing new urban, residential projects. Our quantitative indicators show how state government interactions are most exposed to risky relationships. In contrast to the state governments’ apparent exposure to risk, qualitative data showed they are not only well trusted, but overall the partnership networks reported very low levels of conflict. By exploring the distribution of ‘cooperation’ and ‘coordination’, we identified which stakeholders perceived the most risk. Our approach helps inform efforts to manage risk in urban development networks, the functioning of which will be critical in delivering key climate adaptations in human settlements.