Most of the posts here and on Charlie's blog (http://www.antipope.org/charlie/blog-static/2009/12/21st_century_phone.html) seem to support the original assertion made by Charlie.
"They [Google] intend to turn 3G data service (and subsequently, LTE) into a commodity, like wifi hotspot service only more widespread and cheaper to get at. They want to get consumers to buy unlocked SIM-free handsets and pick cheap data SIMs. They'd love to move everyone to cheap data SIMs rather than the hideously convoluted legacy voice stacks maintained by the telcos; then they could piggyback Google Voice on it, and ultimately do the Google thing to all your voice messages as well as your email and web access."
Tom in comment 37 (Charlie's Blog) makes an economic case to support Charlie's assertion:
"Information is different as a commodity. Sending 1 bit basically has no direct cost associated to it. Nearly everything stems back to the infrastructure costs. Operating costs are pretty minor in comparison. As such, whenever you have a situation where your pricing is primarily based upon fixed costs and amortization of infrastructure capital costs, with no real per unit marginal cost, the price invariably ends up plummeting as performance per price of technology increases, service offerings become standardized, and it results into a race to the bottom."
I do not believe Google will succeed in turning the mobile network operators (MNOs) into cheap data providers by driving the MNOs to commoditization. The service provide by the MNOs is not bits through the air "with no real per unit marginal cost." The core service provided by the MNOs is access to the mobile spectrum. This core service will become more valuable over time and combined with additional services (voice, Internet, video on demand, mobile banking, financial transactions, identity transactions, new advertising models, etc.) will insure the long term success of the MNOs.
Either directly through partnerships or indirectly through data charges, the MNOs will participate in all revenues that flow through their networks.
There is a key insight missed by Charlie and others who have posted on this topic: Unlike cable and fiber which in theory could be laid in infinite amounts, spectrum bandwidth is a finite resource and the dominant MNOs have already been awarded incredibly valuable allocations.
An idea of the complexities of frequency allocation can be gained by viewing frequency allocation charts:
U.S. Frequency Allocations
http://www.ntia.doc.gov/osmhome/allochrt.PDF
U.K Frequency Alocations
http://www.onlineconversion.com/downloads/uk_frequency_allocations_chart.pdf
Additionally, several of the posts here and on Charlie's blog make the mistake of equating higher throughput with greater bandwidth. While each generation of mobile technology has increased throughput, bandwidth (the usable spectrum range) remains a finite and very valuable resource which is leased primarily by the dominant MNOs.
In the United States, bandwidth is usually allocated through a government (FCC) auction process. As more bandwidth is dedicated ("unleashed") for mobile use, the dominant MNOs are in the best position to win the auctions. This is exactly what happened in the 700 MHz auctions held in 2008 (http://en.wikipedia.org/wiki/United_States_2008_wireless_spectrum_auction)
Even with improvements in throughput, consumer demands for new services on intelligent mobile devices will eventually push the limits of allocated bandwidth. What this means is the dominant MNOs have a resource (spectrum allocation) that will become even more valuable over time. What this also means is that consumers will be charged based on their data usage.
