"Currently, T-Mobile's 3G service (UMTS/HSPA) supports 1700 MHz and 2100 MHz frequencies while AT&T supports 850 MHz and 1900 MHz frequencies. Interestingly, both the new iPhone 4 and iPhone 3GS support 3G at the 2100 MHz frequency and, from our understanding, the technical hurdle to support T-Mobile is minor compared to supporting CDMA technology at Verizon and Sprint."
Let's stop right there. Wu is completely wrong about the spectrum. T-Mobile's 3G network runs in the 1700MHz AWS band. The AWS spectrum is actually a paired band. It uses one frequency to transmit data and the other to receive. It is generally just referred to at "1700MHz" without any reference to 2100MHz. The iPhone supports 2100MHz spectrum so that it may be used on European 3G networks. That band is different, and is used to both transmit and receive.
To be clear on this: The new iPhone's 2100MHz spectrum support will only work with European 3G networks. It will not work with T-Mobile's 3G network at all.
While Wu's basic theory that the technical hurdle to support T-Mobile - rather than Verizon or Sprint -- is valid, it's not as simple as Wu presents. Apple would have to engineer a phone to support the paired 1700MHz/2100MHz AWS spectrum. This would largely preclude the iPhone from working on both European and U.S. 3G networks. I just don't see that happening. If Apple is going to reengineer the iPhone, it's not going to aim for the fourth-largest U.S. carrier (T-Mobile), it is going to aim for the largest, which is Verizon Wireless.
Wu, like many analysts, believes that AT&T's exclusive rights to the iPhone is coming to an end, and he is probably right on that. I think his logic, in this case however, is slightly flawed.