The biggest mystery to me is if and when good CMOS sensors (say with on-sensor ADC as used by Sony, Nikon and in some Panasonic sensors) will become available for formats larger than 36x24mm. Maybe if Pentax develops a somewhat higher volume, lower priced sector of the DMF market than other brands have achieved, this can offer the economies of scale needed to get a sensor maker to produce such a product.
I am not holding my breath though.
The way I see it Pentax already developed "a somewhat higher volume, lower priced sector of the DMF market than other brands have achieved" with the 645D and when production will halt the total will most likely pass 15K or even 20K units. That is without Pentax being a complete system by any means. I suspect that an updated camera launched in let's say 2013 could aim for at least the same numbers if not (a lot) more. I know it costs millions to develop such a sensor and most likely thousand(s) to make it but it is one thing o divide that to let's say 5K sensors and another thing to 20K.
I don't know about holding one's breath or not
but what I know for sure is that Ricoh will rain down money over Pentax (monsoon style) the deluge started some time ago and the first results will appear soon. In all honesty I wouldn't be surprised if they order (or even design themselves and hire a manufacturer) a modern CMOS sensor if they feel like there are benefits in long time.
P.S. Don't forget that Pentax could wipe out one of their biggest drawback in the industry with one move if they will come out with a CMOS global shutter sensor. Then all their less expensive, less bulky lenses will get an instant update to "leaf shutter flash sync" greatness. That reason alone is worth the money put in the project IMO. Plus Live View, CDAF, movie mode, improved battery life as equally hefty bonuses for other people.