Re: 3.5GHz
it's ONLY LOS and very much less than fibre capacity.
On top of which as the frequency increases the attenuation due to rain and snow becomes more of a problem, necessitating a rather large fade margin being built in to the links, which can be difficult with the small parabolic reflectors commonly encountered.
From the original article (AT&T) wants to conduct these tests in suburban and rural environments in the California Central Valley including locations around Fresno and Bakersfield. The aim is to collect continuous wave (CW) data for a propagation modelling study, evaluating path loss characteristics in outdoor environments in these midrange and high bands.
All well and good (if it didn't rather point towards the cart being in front of the horse regarding the use of ever higher frequencies) but results obtained in the California Central Valley may well not apply in areas of greater precipitation. IIRC California is a pretty dry State (to the point where some parts are positively arid) so any conclusions drawn may not be universally applicable.
I wonder how many technical staff the MNOs actually employ, or if it's all down to marketeers...