“System innovation” is a multi-actor process that entails interactions between firms, consumers, policymakers, universities, supply chain actors, societal groups, media etc. In recent years, policymakers have shown growing interest in the role of innovation for addressing ‘grand challenges, such as climate change, energy security, transport and resource efficiency, food safety, obesity, environmental sustainability. This interest has given rise to a debate about ‘system innovation’, large-scale transitions and socio-economic transformations, due to the realization that addressing grand challenges may require shifts to new systems in energy, food, mobility, and housing. System innovation is difficult to manage and steer, for it is an open, uncertain and complex process, involving multiple social groups and co-evolution between various system elements, many of which are outside the immediate control of policymakers. Furthermore, the state is not one actor, but fragmented across different domains (e.g. public sphere, private sphere, civil organisations, government) and levels (e.g. international, national, local). Policymakers cannot bring about these processes on their own, but need to invite all the aforementioned actors to work together through strategical public-private partnerships, demonstration projects, scenario workshops, vision building, public debates, and network management. So, in early phases of system innovation, policymakers tend to act as facilitator, stimulator, and chain manager. In later phases, when there is more clarity about the best technology, market demand, and infrastructure requirements, other policy instruments (e.g. regulations, standards, taxes, subsidies, financial incentives) tend to become more important, aimed at widespread deployment and uptake. Furthermore, national innovation systems (NIS) (i.e., education and training systems, science base, intellectual property rights, university-industry knowledge exchange networks, venture capital availability) provide important generic contexts in which countries address system innovation. It would be useful if future research would develop more dynamic understandings of NIS and investigate if and how NIS need to change to facilitate system innovation (e.g. through mission-oriented R&D, changes in incentive structures for academic researchers).