The numbers indicate that it's a small shift at best. People switch carriers every 3-4 years on average (usually citing employer discounts, pricing plans, family networks, etc,) meaning that for any given phone, you'd expect 25-33% of purchasers to be new customers. The iPhone is somewhere between 30-40%, meaning that it's probably outperforming the average by 5-10%, which is great but not indicative of a sea change in consumer habits.

It's probably more effective at keeping people on AT&T who would otherwise have left than in bringing new people on.

Hmm, aren't you forgetting the mathematics of compound interest? A 5 to 10% conversion advantage over your competitors, combined with a higher retention rate -- this is going to really hurt competitors after several years. After 10 years? Ouch!

It's a good thing that AT&T sucks so much. Otherwise, the iPhone would be unstoppable.

Oh it's great, and if I'm a shareholder in AT&T or Apple I'm pretty damn happy about it. (The retention rate is a total unknown for a device that's 2 years old. Even a shitty free-after-contract clamshell has a 2 year retention.)

But it's not a massive change in the power balance between carriers and phones. Carriers aren't going to offer massive concessions for something that brings in a few % more users than the other 50 phones in their lineup average. Don't forget that overall, the iPhone isn't even in the top 10.

Similar to the sibling, I see 5-10% as being a fairly solid impact.