The printing of physical journals was extremely expensive. So this was the main concern.
Also - this MIT is interesting (and good) but I doubt it would be applied everywhere. Will everyone at MIT stop publishing in IEEE journals (very restrictive about copyright)?