GPS in
phones is a)
100% unnecessary b) Provably
unprofitable.
GPS in mobile phones is unnecessary as they can triangulate their position, as long as they can contact two or more base stations. This works in buildings, in urban canyons, anywhere else where you have reception. GPS needs a pretty good look at the sky to function.
People who want navigation systems buy navigation systems. People who want to communicate buy telephones. People like saving money. Oh, and you can track someone's position based on their mobile phone signal without putting anything special in the phone, so you could get real-time position updates over WAP. Cheap to manufacture, and you charge by the minute. That's what I call a business model!
In defence of GPS in mobile phones, the proposed business model is that the phone companies would sell location-based advertising. The way this would work is that as you passed a shop/church/morgue, you would be subjected to advertising trying to entice you inside. To do this you need accurate location sensing.
It is true that when it works, GPS is much more accurate than base-station triangulation, but it does require data to be sent from the phone to the base-station concerning current location. The problems with triangulation is that distance from the base-station needs to be inferred by signal strength, which can be affected by barriers. This introduces inaccuracies which can be quite large. It is good enough for most people who just what to know their general locality, but if you want to know which shop someone's standing outside, perhaps you ought to install some base stations in the town centre?