The recent Copenhagen summit highlighted the inability of national governments to agree to binding targets to reduce the emissions of greenhouse gases. Notwithstanding such policy failure at the national level, cities around the world have come together to mitigate global warming. Given that cities account for 80 percent of greenhouse emissions, this is an important development that state centric accounts of global environmental politics tend to overlook. I probe deeper into an important network that has emerged to facilitate such collaboration. The C40 network seeks to bring together world’s key cities which have displayed a commitment to tackle climate change. Cities collaborate by sharing best climate change practices, exchanging personnel, and serving as a pressure group. Thus, the more embedded is a city in this network, higher will be the benefits it can capture by virtue of its participation. I examine the uneven distribution of collaboration ties among cities participating in the C40 climate change network. Specifically, I study why we observe variations among cities in the number of ties they have with other cities, and what factors influence a given city's decision to collaborate with another specific city. Using social network analysis, I focus on how homophily, an attribute of network structure, and policy performance, an attribute of a given node (or city), influence the distribution of collaboration ties. Homophily suggests that that collaboration is more likely among cities with similar structural characteristics. Employing a recently developed network analysis technique, Exponential Random Graph Model, I find that collaboration between cities is more likely when these cities are located in the same continent. Further, I find that cities with higher level of performance in the climate change area tend to attract more potential partners in relation to cities with lower level of policy performance. Important policy implications follow from my analyses. Given that some cities are less likely to find collaborative partners because of their location, policy intervention is required to correct the structural inequities. Second, cities with superior policy performance will serve as magnets for other cities looking for partners, and will therefore corner higher benefits in relation to cities with lower levels of policy performance. While networks are often viewed as equitable structures, the benefits from participating in the network can be unevenly distributed for reasons which are exogenous (at least in the short term) to the nodal actor. Thus, a conscious policy to redirect network benefits is required if network equity is an important policy objective. More broadly, if the benefits from C40 network are to be evenly distributed, important steps need to be taken to encourage ties with cities which are geographically challenged, and which are late comers to this policy (and therefore have low policy performance).