My advisor is a specialist in Grothendieck algebraic geometry and Noncommutative algebraic geometry. I ever asked him about reading EGA and SGA. He strongly suggested to finish all the exercises in Hartshorne but not read Hartshorne. And he said it was good to take a look at EGA when you need something.
For SGA, fortunately, it is not necessary for us students to read these materials because he gave us lecture course on his own work which heavily depends on Grothendieck Machine(especially Tohoku lecture,SGA1 SGA 4 SGA6 SGA 4.5 and other book, say J.Giraud's cohomologie non abelienn, Gabriel-Zisman). In fact, we can learn most part of these materials from his lecture course and even more categorical(since he is framework maker for NCAG which is based on category), It is really a happy experience However, he said it was better to listen not to read.