Scientific alliances are typically referred to as ‘collaborations’ but in recent times, those with global health or other development goals are increasingly referred to as ‘partnerships’. I observe that one of the features common to this type of partnership is temporality: flagship programs are frequently initiated but less commonly sustained. Thus the pressure that short-term transnational projects place on African health and educational systems that implement them is sometimes hard to justify. I suggest that one reason for the short life spans of partnerships is inadequate attention to the need to build ‘hard’ and leadership capacities: infrastructure, managerial expertise, administrative capabilities, and the capacity to improvise at African partner institutions.