Introduction In the emergence of new technological areas and industries, ambiguity of technologies and markets are large, posing challenges for the building of innovation systems (ISs) conducive both to the necessary experimentation and variety creation, and to the likewise crucial ability to select and de-select technologies. One particular problem is that one can seldom beforehand predict which technologies will become the most pervasive, but instead, taking an evolutionary approach to industrial dynamics, it is acknowledged that continuous experimentation is necessary, resulting in a multiplicity of competing or complementary solutions to the societal and economic needs at hand. Technological diversity creation and the competition it entails are generally beneficial, both in early phases of development where uncertainty is large and in later phases where market acceptance is key. Indeed, such experimentation and diversity, the literature argues, is often a prerequisite for innovation to come about and thus something that is encouraged in the modern knowledge society. Given a multitude of alternative technologies in a well-functioning IS, selection and retention mechanisms kick in, assuring some alternatives to take the upper hand at the expense of others. In practice, experimentation and technology variety may be difficult to maintain; instead, early selection and thereby lock-in situations often occur. One reason can be that choices between options are called for due to scarce resources. Also, at each point in time, certain alternatives might appear more advantageous and be selected. In particular, this balance becomes pertinent in the case of sustainable technology. In such instances it is often not simply a matter of encouraging variety and leaving it to the market mechanisms to choose which is most suitable. Instead, additional considerations enter the puzzle in the form of environmental priorities, raising questions of which alternatives most efficiently mitigate pressing environmental apprehension, how current lock-ins can be abolished, and if early selection may be beneficial in order to get an alternative - less detrimental for the environment - out on the market quicker. Such decisions of favouring diverse experimentation or, conversely, making early selections are taken through a number of different mechanisms and actors. On the one hand, in situations where several firms compete with substitute technologies, experimentation might thrive. On the other hand, specific firms may through investment in one alternative technology guide others to move along the same path, and variety creation may be stifled. In the same way universities, financiers, policy organizations, etc. can, through their actions - alone or together - influence the way technological experimentation and the IS as a whole evolve. The intentional coordination by one or several actors is what we call the ‘governance’ of the IS. In this chapter we are particularly interested in how one, through various governance arrangements to enhance innovation systems for sustainable technologies, can handle variety creation and technological competition (Sandén, 2004; van den Bergh, 2007).1 While governing actors may in essence be of any kind, in this study we pay attention to when public actors intervene, alone or in combination with other actors. Leaning on evolutionary policy (e.g. Nill and Kemp, 2009; van den Bergh and Kallis, 2009) and a functional approach to ISs (Bergek et al., 2008a), our research question addresses if and how technological variety may be maintained when early market selection is warranted for reasons of sustainability.2 How did the governance arrangements implemented to support biofuel innovation in Sweden handle the balance between technological variety creation versus early market selection? Our empirical case is that of the technological innovation system (TIS) for biofuels in Sweden for the period from 1990 to 2010, including a number of competing ‘sub-technologies’ of first-and second-generation biofuels, i.e. biodiesel, ethanol, biogas and synthetic fuels. The discussion starts with a theoretical foundation for our work, considering the role of variety creation and how it raises a dilemma for governance of innovation processes (next section). Next, our approach to TISs is presented in some detail, along with a description of the research design. A short account of the biofuels’ value chains and the activities in Sweden follows, leading us into the analysis of governance. The concluding section summarizes our findings.