Andy Abramson discusses the cost of GPRS for data (see:VoIP Watch: Do The Math) on his blog. His conclusion, the high cost of GPRS for data will lead to a greater push for GSM/WiFi phones, which can use VoIP over WiFi to save money.
I'm not sure I see Andy's point here. I pay $24.95 a month for unlimited data transfer from Cingular on their GPRS/Edge network. Any time I've been in a Starbucks, the t-Mobile rate for WiFi has been $9.95 a day, or something like $3.00 an hour.
Given that I'm already paying for GPRS/Edge so I can get mobile e-mail & web access via my Treo, what would I need with a Hotspot? If anything, the growing availability of GPRS, Edge, and EVDO are making public hotspot services far less attractive.
We've been talkiing a lot among our analysts about the future of mobile access. Given all the variability of coverage of public hotspots (multiple carriers requiring multiple accounts, service not available everywhere), the idea of simply equipping employees with a GPRS/Edge or EVDO card is getting increasingly difficult to beat. The key benefit from a network management standpoint is that you can provide easy access for your employees, rather than needing to support multiple accounts on the various wireless services (though it should be stated that services such as iPass and Fiberlink do offer greater ability to use a single account for multiple services).
However, you still have the issue of lack of coverage. I was in the Miami airport a few weeks ago, which didn't have WiFi outside of Laptop Lane ($10.00 an hour or so if I recall correctly). Now that I've got my Treo with Bluetooth DUN, the days of searching for a WiFi hot spot are over.