Sat-nav systems relies strongly on very stable clocks with minimal drift. An error of nano second gives about 30 cm error in the measured distance.
The frequency standards on the satellites are made up of multiple atomic clock like rubidium resonators or passive hydrogen masers. The clock technology is improving,and thus the coming series of satellites from both GPS and Galileo will have more stable clock sources, it seems by the writing in magazines and articles!
E.g. this article from 2011 http://www.weblab.dlr.de/rbrt/pdf/ION_11B56.pdf that discusses measurements of the new clock types and models that are used in Galileo and GPS block IIF satellites.
Galileo uses hydrogen masers as clock references in the satellites.
ESA has this short introduction to the atomic clocks used on the IOV satellites (and onwards). The particular maser design selected for Galileo should be ten times more precise that the atomic clocks currently used in the deployed GPS satellites (before January 2010, before block IIF). Web reference at http://news.bbc.co.uk/2/hi/science/nature/4555276.stm .
The new GPS IIF satellites that are now launched and put into orbit and becoming usable, will probably also have atomic clocks that are a class better than what was used on the earlier launched satellites. It is a part of the general "GPS modernization". (Oh, I do not have the measured numbers of the actual clock performance like drift)
For a nice introduction to the accuracy of atomic clocks in space, see this article at http://www.insidegnss.com/auto/Sept06GNSS_Solutions_secure.pdf .
When the number of satellites in the sky with better clocks increases, the position accuracy goes up as well (for a given system). Another use for a GNSS is a very precise time reference, e.g. for cellular base stations, and these applications will benefit as well.
General site about atomic clocks: http://www.atomic-clock.galleon.eu.com/atomic-clock/atomic-clock.htm