They were never in mobile per se, they only licensed mobile GPU designs through their Imageon (Adreno = Radeon) unit. However, Dirk Meyer sold it off to Qualcomm for an absolute song. Both the value of the deal (tiny $65m) and the strategic decision were heavily questioned at the time. Ultimately it went down as one of the worst tech asset disposals of all time.

I don't see any likelihood of AMD entering that section of the ARM market anytime soon. There's heavy competition and very low nominal margins. Moreover they have no expertise in the field, and need to concentrate on K12, Zen, custom business and desktop and laptop graphics.

Licensing their GPU designs, adapted for mobile, helps them push HSA, which gives them far more leverage vs Intel and NVIDIA.

Apple bought Freescale, and also received massive help from Samsung. They also have an essentially unlimited budget. Samsung have a long history of involvement in telephony, gigantic budgets, and both their own fabs to utilise (AMD don't anymore) and their own screens, RAM and solid state memory. Both use their own product, and AMD have no consumer devices.

Meego was never really in, since Elop ensured that it was stillborn, and the old Nokia guard can be blamed just as much for stalling and under-investing in Maemo, then the largely disastrous decision to merge it with Moblin.

Palm I have little sympathy for. They were a disastrously badly run company that rested on their laurels. WebOS was way ahead of its time, but the hardware was terribly low quality and it was too little too late.

Sharp failed to develop more affordable, mass market mobile products, and more generally had appalling financial mismanagement. The JDI joint venture was massively more prudent and provided more affordable offerings. However, they're still a big player in displays.

Next year will be a watershed year for AMD ... if they get K12 and Zen right, they could find themselves in a better spot than ever before.