I know that it is entirely possible to make a great network setup over 2.5ghz. Sprint has proven such. Clear just made such a poor example of it. Reading that white paper on the issue, the propagation decrease to 3.5ghz wasn't as much as I initially thought (I am an engineer, I took an educated guess). It just won't be cheap to deploy it, especially for a company that has tower spacing based on low bandwidth networks (ahem, VZW). That's probably why they're only looking at it for hotspot type capabilities.