Got it, thanks all, so it is mix of production cost, low power consumption and easy implementation.
A meter with a pixel display and open source firmware could lead to an interesting variety of new features once the experimenters got their hands on it.
That's what I exactly thought, seems like it isn't huge problem to install such a display but the bigger is to implement a firmware, if indeed DMMs are single chip solutions.
Anyway nowadays I think it's not a problem to put powerful lot-power consumption ARM processor.
Thanks everyone for answering
