Yeah, Charlie's sources are horribly bad in the handheld market.
But unlike some of you, I can see the appeal of trying to sell the OMAP division. TI is trying to focus on their analog business. And yes, smartphones are a very high growth business, but I always liked Bob Swanson's memorable quote when Linear Tech divested its handheld division: "We recognized there was no pot of gold at the end of the cellphone rainbow, just red ink."
The wireless division at TI had revenue of $658M in Q1 and $558M in Q2. That decline is mostly a reduction in the legacy baseband business (i.e. Nokia) and a bit of seasonality, but the fact remains that eventually the entire legacy baseband revenue will be gone. That's $228M in Q2, which brings us down to only $330M revenue - and the $82M operating profit would become slightly negative. Excluding legacy business, TI is already losing money in the wireless space (or at least not noticeably above break-even). They'll need that strong smartphone growth just to get back to where they are today.
I'm sure OMAP and the associated connectivity division can be a good and profitable business despite the strong competition and TI's lack of a 3G/4G baseband, but I'm not convinced it can be so profitable that it wouldn't be more attractive to sell it to the right buyer at the right time. That buyer nearly certainly isn't Apple or NVIDIA, but I can see very good justifications for Samsung, AMD, ATIC, and even Intel. However there's also the complication that the embedded division reuses OMAP tech and even OMAP chips and I'm not sure how that would be handled (would TI be willing to sell the whole thing? I'm skeptical).
Then again this probably won't happen, both because it's not very likely in itself, and because Charlie said it will - which is a pretty reliable sign that nobody's actually considering it at this point