T-Mobile could get spectrum the same way AT&T Wireless got spectrum to overlay GSM on TDMA. Better filters at the base-stations to allow tighter frequency reuse and better codecs on the mobiles to tolerate higher system noise. They could also free up spectrum simply by building more base-stations, which would allow better frequency reuse without increasing noise levels. All of these things would have cost, money, but so does buying new spectrum.

KDDI wasn’t selling EVDO in 2003, as it didn’t exist. They were selling 1xRTT which is much slower than DoCoMo’s proprietary WCDMA, but they were selling it much cheaper. Also, DoCoMo had a huge problem with overpriced and defective handsets at this time.

Europe’s 3G problems stem from the fact that carriers paid 3X too much for their licenses and as such couldn’t properly build out the networks in a timely fashion. Flawed business model or not, it wasn’t until 2005 that there was much of a network to run it on anyway.

Secondly, in 2002-3 — KDDI was killing the all mighty DoCoMo at 10:1 ratio for 3G subscriber net adds. Verizon would have beat TMO by the same ratio.

Thirdly, even if TMO decided to launch 3G in the US 3 years ago — they would have launched the wrong “Euro-centric” 3G business model. Video calling turned out to be a dud worldwide.

Note, for example, that KDDI and SK Telcom also both show up on that global top 10 list, and they use EV-DO. China Unicom also has an EV-DO network, though I think that they also have some GSM. Which means that not even on the global list does GSM have an obvious head start.

What makes you think that GSM has/had an early lead on delivery of 3G services? GPRS and EDGE are not really faster than 1xRTT, and HSDPA and EV-DO are comparable, while EV-DO generally beat HSDPA to the marketplace worldwide. (SK Telecom in South Korea with a EV-DO network in January 2002, KDDI in Japan in 2003 had a smoother introduction than the initial DoCoMo UMTS rollout.) Throw in the backwards compatability of cdma2000 gear with the cdmaone stuff, and it’s really not clear that T-Mobile and GSM in general should have been leading.

With the early GSM lead, they should have been the ones setting the pace for early delivery and capitalization of 3G services.