Innovation is one of the most used words in current policy documents within the sectors that give or receive aid. The use of the word is indeed multifaceted. At the recipient side, the meaning found in policy documents is more analogue to introduction of new technologies in developing countries.At the donor side, the concept tends to refer to innovations in the aid methodology rather than to changes in the socio-economic or socio-technologic aspects in the aided developing countries. Aid innovations during the last twenty years have been mostly concerned with policy development and economic reforms. At the nearby end of this trajectory, we now see how recent calls for aid effectiveness and results has led to a situation where control and compliance are the main demands from the political and financial authorities. This modus operandi of aid, focused on political planning, control and compliance, contrast starkly to the “g(e)ist” of innovation, where spontaneous, incremental and evolving are keywords alongside with serendipity. This paper aims to give a background to the trap that Aid donors have constructed for themselves. One of the responses to such challenges was the Innovation Systems and Cluster Development program driven by the faculties of engineering at University of Dar es Salaam, Tanzania, Makerere University, Uganda and the Eduardo Mondlane University in Mozambique. The way this program has developed shows that there might be ways out of the donor’s trap. It also shows that innovation and learning together in clusters can matter for development.