1. A GPS satellite is in a very high orbit compared to LEO, and so all stages were requested to burn to completion, to ensure the satellite has enough margin to reach that orbit.
2. The second stage ends at a ~800 km orbit, while the satellite itself needs to increase this to ~12,000 km (I'm rounding these numbers). The sat has to carry enough on-board propellant to raise it's orbit that much which necessarily increases it's wet-weight due to the rocket equation.
Looks like they use a liquid fuel thruster lift it to its final orbit[0]. Some other missions have used ion engines for orbit raising[1]. I wonder why that wasn't incorporated into the GPS design.
[0] https://en.wikipedia.org/wiki/Lockheed_Martin_A2100 [1] https://space.skyrocket.de/doc_sdat/ses-12.htm
The only information I can find on why this one was so energy intensive has to do with the high inclination the GPS satellite has to be in (55 degrees), which is pretty far off from inclination of the launch pad in Florida.
Wikipedia says this launch was originally planned on a Delta IV. That would make sense because it has higher specific impulse hydrogen engines. The Falcon does well in LEO, but improved Isp really helps as you go farther out.
Here's a video comparing ULA's rockets to SpaceX: https://www.youtube.com/watch?v=QoUtgWQk-Y0
https://www.google.com/search?q=plane+change+delta+v&ie=utf-...
That said, remember they could potentially be using a rocket that's flown twice before (and AFAIK three launches is the most they've done so far), and as such may view the rocket as largely expendable anyway (they lose the ability to use it for re-entry experiments and to pull it down after the launch, but nothing else).