FWIW, my response to David's post (awaiting moderation):
David, the problem is both you (and possibly Marvell too) are looking at the wrong places. This video will give you a very big hint of who is the real early customer for those ODM designs: http://www.youtube.com/watch?v=j9aE3ubx7h8 - there's a higher-res version of that somewhere on the net where you can see the name of all the carriers, but I couldn't find it.
At the end of the day, there will be three OSes for ARM netbooks: Chrome OS, WinCE, and Ubuntu. Anyone claiming anything else has a chance to penetrate more than a few percent of that market is delusional, IMO. And of course, that's the other problem: on that list, the first (and most interesting) option is far from ready and ARM Ubuntu arguably isn't completely there yet either.
While I am definitely convinced that operators are a good 'go to market' strategy for ARM netbooks for a multitude of reasons (good distribution channel where you've got a more direct contact with your customer for example and interest in solutions that are limited, as long as they are limited in the 'right way'), I am very worried that they might screw up the subscription model.
It's fine to expect the customer to be locked to a 3G contract in exchange for for an even cheaper product ($49?) - but it's not fine for that contract to be a 2 year one when it's essentially a disposable product IMO, and this also moves this even further away from impulsive buy territory.
Frankly, I think companies like NVIDIA/Freescale/Qualcomm should just have sold some self-branded netbooks themselves (low volume is OK, 10-25K?) as ways to excite the market and get OEMs looking at it more seriously. Now the plan is for carrier sales to do that, but the problem is those keep getting delayed. The Tegra-based Mobinnova Elan was supposed to be sold by early carriers this year, but now they claim it'll happen at CES 2010. This better be the last delay...
There's still a chance some other Tegra or Snapdragon-based netbook will start selling in some obscure region of the world this year, but I'm skeptical. The OEMs have clearly dropped the ball on this one, and the carriers are on good track to follow suit. I guess evil scheming could play some part in the OEM reaction, but when it comes to carriers the delays are more a question of them being inherently slow to do things, in addition to often being incompetent.
Let's hope 2010 will be a better year. Maybe carriers will deliver. Maybe NVIDIA will even manage to get Tegra2 netbooks (2xCortex-A9) out in 2H10 as they were and still are hoping. Maybe others might also deliver on their goals and maybe OEMs will finally get it. Or maybe not.