Sustainability transitions literature argues that path-breaking innovations 'survive' in protective spaces which (temporarily) render them more or less immune from mainstream selection pressures. But while this effect of protective spaces has been thoroughly discussed, systematic theorization of how these spaces are created and maintained or removed over time has not occurred. Recently, Smith and Raven (2012) have argued that protective spaces shield, nurture and empower innovations. Shielding is about multi-dimensional work aimed at shaping a protective space by exempting an innovation from mainstream selection environments; nurturing is about work aimed at improving its socio-technical performance; empowering is about work aimed at altering mainstream selection environments (e.g. incorporating sustainability criteria) or making innovations competitive within existing mainstream selection environments. This paper first makes an ‘insider analysis’ of how actors have created, maintained and removed protective spaces for solar PV in the Netherlands (which has so far failed to ‘break through’). It does so for five key spaces: research and development, off-grid applications, building-integrated PV, retrofitting, and PV manufacturing. In each space, "actors are driven by ‘a logic of control’ to effectuate through complex processes" and "agency is full blown and located in individuals" (Garud et al., 2010). Second, the paper explores the utility of shielding, nurturing and empowering concepts across these five spaces for constructing a more decontextualized ‘outsider’ analysis’. The paper concludes that 1) shielding, nurturing and empowering are useful concepts to gain a richer understanding of the survival of PV in the Netherlands; 2) they are not temporally successive phases, but analytical dimensions of different types of concrete work by (networks of) champions; 3) more detailed and comparative analysis of work done by these (networks of) champions is necessary for finding patterns across cases; 4) this poses ...