Over at GigaOM there is an
article about T-Mobile's planned launch of a converged Wifi/GSM service in September:
T-Mobile plans to launch a Wi-Fi-cellular converged phone service in Seattle and potentially one other market next month on September 12th, sources say. More markets will follow soon after. The city of Chicago and the San Francisco Bay Area are the likely candidates for a possible rollout.
As the article notes the solution is based on UMA - a concept that has the backing of many of the big names (Nokia, Alcatel...) and seems to have been created by Kineto Wireless. There are alternate solutions around such as one developed by Irish startup Cicero and the VoIP pico-cell route taken by ip.access.
The attraction for T Mobile to offer this sort of service is clear: unlike the other large operators it has little or no fixed line business in the US and plenty of WiFi hotspots, so not only is it able to roll this out quite widely, it won't see any cannibalization of fixed line revenues. The question is whether it will actually catch on.
I can see a couple of positives, firstly the "single device" aspect, which I think is key. That is to say the possibility to have one device to make calls which automatically uses the cheapest method.
Secondly it may also help drive usage of mobile data services because the sorts of phone that support UMA are also the ones that tend to have good data capabilities. If the phones can also move the data services between T-mobile hotspots and EDGE then T-mobile has a lot of potential in the business market. For mobile data this sort of capability is a win-win because WiFi will tend to be deployed in those locations where demand is highest, thus alleviating demand for cellular data, which is more costly to provide.