Archos is talking about deploying a lot of LORA mini-basestations in europe (200K) , and for client , subscription fees would be $2/year for the low end.
http://www.telecompaper.com/news/archos-plans-picowan-smartplug-based-lora-network--1106399
And with others entering the market, how long before the ISM band is saturated to the point that it becomes useless for everyone? There are duty cycle limits which aren't so much a problem for clients with low data throughput but because of the long range, the basestations will potentially see a lot of clients of their own, their competitors clients and basestations and other private networks - eg. smart meter reading, pet and child monitoring systems etc. And there are other existing users of the ISM bands including fire alarms, security systems etc. to add to the problem.
Duty cycle limits vary but 1% is typical meaning potentially that with say 6 channels, 600 radios could saturate the band. In practice the actual number would likely be far less due to frequent collisions. Perhaps these public networks will place very low duty cycles restrictions on clients to reduce the problem but there is nothing to stop private networks being installed which want to maximize their throuhputs.
In the European 868MHz ISM band the duty cycle limits can be avoided by the use of 'listen before talk' and 'adaptive frequency agility', but I don't know how well they would work with potentially hundreds of nodes competing for access. Badly I would expect.
There aren't many channels available in the 868MHz band in Europe for the wideband Lora modulation and the very low data rate means that even short messages are on air for a relatively long time.
Wifi congestion is a serious problem in many places but WIFI has very short range in comparison. My guess is that it could all fall apart very quickly when they start to get rolled out in earnest, unless rather more spectrum, is released - unlikely.