One of the biggest challenge for the SED models is to make them applicable to the large statistical samples of panchromatic data that are now becoming available for the local universe galaxies, like for example the GAMA survey (Driver et al. (2009)), the Herschel ATLAS survey (Eales et al. (2009)) and GALEX (Martin et al. (2010)). In the future we will also need to re-calibrate our SED models to make them applicable to the distant universe, using data from the next generation observatories, like SPICA (Swinyard et al. (2009)), ALMA, JWST and VLT. At the same time we will need to include dust physics in the simulations for galaxy formations within a fully cosmological context. Realistic simulations of disk galaxies are now being for the first time produced (see Brook et al. (2010)) and, within a few years, we expect this to be routinely done. It is therefore crucial to incorporate all known physics related to sources and sinks of grains in galaxies and in the surrounding IGM, as well as the cooling and heating mechanisms involving grains. Only then we will be able to make reliable predictions, which will directly constrain theories for the formation and evolution of galaxies.