We revisit MSSM scenarios with light neutralino as a dark matter candidate in view of the latest LHC and dark matter direct and indirect detection experiments. We show that scenarios with a very light neutralino (~ 10 GeV) and a scalar bottom quark close in mass, can satisfy all the available constraints from LEP, Tevatron, LHC, flavour and low energy experiments and provide solutions in agreement with the bulk of dark matter direct detection experiments, and in particular with the recent CDMS results.