The idea to split the sun spectrum in light beams and to drive theem toward different cells of selected material is not new. As well as the idea of concentrating the light, it can be realized in a number of different configurations; the constrains for its implementations are mainly related to the costs of these assemblies, considering that additional complexities are introduced; indeed, to split the solar spectrum, two physical ways are possible: dispersion through a transparent prism or reflection/transmission through dichroic filter working for light interference. The light is concentrated too, in order to reduce the costs of the cells dedicated to defined wavelengths. In these configurations the theoretical efficiency can achieve its maximal level, because each cell produces power in the best conditions of irradiation, without constrains of series electrical connections as happen for multijunction monolithic structures. Multi-cells arrays with a record efficiency of 43% have been fabricated (Green & Ho-Baillie, 2010) to demonstrate the feasibility of this approach.