They did and it failed miserably. I may be mixing the timeline a little bit but as I recall they first partnered with Canonical and did a sort of Ubuntu-based Moblin. This didn't work well (Canonical has almost zero engineering resources, they can't actually write drivers and solve problems, and there were horrible hacks like touchpad interfaces implemented incorrectly, etc). Management changed and they got a Suse guy in charge, he scrapped the Canonical-based stuff and went RPM-based so then Moblin started looking more like SuSe and the like, in the end this was a mess too. So much for the "netbook remix" approach there, even if it did make sense on paper.
Meanwhile in MeeGo land they're trying to create an embedded distribution with the software stacks needed to do devices (ex: BT as a device, SoC support, touch screens, etc). This isn't supported well in "desktop Linux" anyway, they're competing with WinCE variants and Android and the like. The Linux distribution companies are no use here (especially not Canonical), this is work that consulting companies or device makers typically do. They were (are?) targeting POS, in-vehicle, and mobile devices.
In the end this absolutely can be done, I've seen companies do it successfully internally (at least for their own families of devices), it's just that no one has produced a quality usable and well-accepted "industry standard" stack that easily competes with Android, CE, etc. yet and Intel and friends seem to schizophrenic to pull it off.