I don't normally cover the cellular space, deferring to co-columnist and 3G expert Peter Rysavy, but it's impossible for me to pass by this past week's biggest wireless news story: T-Mobile USA's joint announcement with German parent company Deutsche Telekom to roll out nationwide 3G.

The news was hardly unprecedented in light of T-Mobile's recent success at FCC Auction 66 for spectrum more affectionately called AWS (Advanced Wireless Services). High-speed cellular data services, typically labeled '3G,' require a lot of spectrum. EV-DO, a CDMA-based technology, uses a pair of 1.25 MHz channels. T-Mobile likely will use HSDPA (High-Speed Downlink Pack Access); this GSM-based technology uses a pair of 5 MHz channels.

Before the auction T-Mobile only had an average spectrum depth of 25.9 MHz among its top 100 markets, just two-thirds of that possessed by Verizon Wireless and less than half of leader Cingular (which recently purchased AT&T Wireless). Post-auction this bumped up to a more respectable third-place finish, with an average of 52.2 MHz, just behind Verizon Wireless. T-Mobile's conference call and associated press collateral alluded more than once to the fact that its current spectrum constraints have now been resolved in "New York, Los Angeles, Boston and Miami."

Until two years ago, T-Mobile's cellular data network offered respectable speeds. While GSM providers were trudging along with GPRS rates of 30 kbps and CDMA providers with CDMA-1xRTT rates of 60 to 70 kbps, T-Mobile's EDGE network dominated with speeds of 100 to 130 kbps. That lead has long been overtaken. Cingular is at least a year into its own substantial HSDPA (based on the UMTS framework) deployment, which offers downstream speeds of 400 to 700 kbps. And Verizon has had its EVDO-Rev 0 technology for even longer; it offers 400 to 600 kbps.

In its semi-defensive position, T-Mobile has pointed to the roughly 8,000 Wi-Fi hotspots it provides as well as its cellular data service, which costs only $30 a month--less than half of its competitors' initial price points. Nevertheless, in the months leading up to the AWS auction T-Mobile dropped several hints that it was going to participate. When the auction finished, T-Mobile had bid $4.2 billion to claim 120 licenses that provide at least 10 MHz of spectrum in the continental United States and contribute to a national average of 23.2 MHz.

Although late to the 3G game, T-Mobile has two advantages: mature and more affordable second-generation technology and cheap spectrum. Now that Cingular and the rest of world are well along with their HSDPA deployments, T-Mobile can purchase and implement the same or newer equipment without the initial teething pains. Merrill Lynch, in a research note, says that it expects base station providers Ericsson and Nokia-Siemens to be selected. Nokia circulated a flyer concerning its Flexi Base Station to AWS auction participants, and Huawei announced its own AWS product offering this spring.

On the customer side, T-Mobile claims that handset prices will fall below $230 in the next 12 months, meeting this technology's "adoption curve" for consumers. The company has been working with handset manufacturers over the last several months, so it will be ready once the network has been built out.

Spectrum prices are usually compared using dollars per MHz POP. For example, if $1 million were spent on 20 MHz across 100,000 people, the price would be $0.50 per MHz POP. In the recent AWS auction, T-Mobile was able to acquire spectrum at a mere $0.63 per MHz POP, compared to previous cellular auctions that exceeded a $1 or more. It's interesting to note that the 3G build out--or upgrade, as described by T-Mobile--will cost two-thirds of the spectrum price. Enterprises should have a greater appreciation for the free (albeit unlicensed) spectrum they have in the 2.4 and 5 GHz ranges.

It's not all win-win, though, for T-Mobile. AWS spectrum, found at 1.7 and 2.1 GHz, is not used worldwide. (This demonstrates the lack of a global, coherent spectrum allocation policy, similar to what the United States has with 3.5 GHz and WiMax.) The handset radios will need to be tweaked for the U.S. market, and they will likely include support for 1900 MHz PCS spectrum, too. For the world traveler, the handsets also will need to support the slightly differently configured 2.1 GHz used in Europe and elsewhere. Multi-radio support seems to be the norm in small form-factor devices, with the addition of Bluetooth and Wi-Fi, but any customization only adds costs and reduces handset selection.

If T-Mobile was behind when it announced its plans last week, it will be even further behind as it builds out the network starting in the first quarter of 2007. The company has promised some market availability in mid-2007, but the network will only be complete in 2008. This leaves plenty of time for Cingular and Verizon Wireless to tune their market messages and bulk up their networks.

According to a Computer Business Review column on the announcement, the business marketplace will not be the focus of T-Mobile's efforts. T-Mobile USA CEO Robert Dotson dismisses the business segment as a "smaller marketplace." The company has never dominated the enterprise market, so perhaps it does well to focus on consumers. Verizon Wireless has a strategic advantage with its wireline associate, Verizon, because both firms do very well in the business segment. Cingular is owned by AT&T/SBC, which offers services similar to Verizon, and Sprint-Nextel has been working hard to build up its enterprise mobility arm.

T-Mobile would like to spin this 3G announcement as a planned success. But the reality is that only a few months ago there was speculation that parent company Deutsche Telekom would divest itself of its subsidiary. Through the good fortune of affordable spectrum and amenable hardware vendors, however, T-Mobile may yet have a 3G play.

Frank Bulk is a contributing editor to Network Computing Magazine covering wireless and mobile technologies and works for a telecommunications company based in the Midwest.