Whether this translates to actual take-up of a service is a different matter - even if it should be launched at the right price-point and executed well. Why? Firstly, there's a precedent for people being poor predictors of how they'll use new technology. For two mobile-specific examples of this, look at the unexpected rise of SMS and the research DoCoMo did before launching I-mode. The former is well documented, the latter showed that the public expected to focus usage on news and financial services, whereas entertainment-based content has shown itself to be more popular.
But perhaps we're missing the point here... maybe what's attractive about TV on mobile phones is that the idea brings together two things which are loved and understood by consumers. And whilst the marrying of these may not deliver a service that obvious relates to either parent (what works for mobile TV may be different to what works on conventional TV; pricing for delivery via mobile may lead to different patterns of usage), it's a concept that, when they're asked, consumers say they can, or think they can, understand.
(For an idea on how it might work, imagine a service that distributed video clips to handsets "in the background" - putting otherwise-unused 3G bandwidth to work in delivering low-priority content over minutes hours, not seconds. So that, for example, at 5:30pm when your office bod leaves work, he has a daily news bulletin which is personalised to his needs, maybe ad-supported or maybe not, has made minimal impact on network usage for revenue-generating events like calls that are key for operator revenues, and is charged on a subscription basis rather than usage.)