You are extrapolating a projection scenario from a very transitory situation. In 2014 Celeron Broadwell will have roughly the size of Kabini/Temash, and AMD will have to move Kabini production to GLF, plus 22nm Atom will be online with 14nm Atom in the end of the year. All those changes will improve Intel position while worsening AMD position, and significantly changing the rationale behind your AMD assessment.
You explain the reason i wrote h2 2013 and h1 2014 - it takes time to get temash/kabini to market, and it takes time for Intel to get the 14nm Atom to market.
And its not about size,- but cost.
We will se how Atom is positioned power wise because thats the main competitor:
We know The AMD Hyderabad team putting together the IP for bobcat APU, and getting it ready for production, was aprox 80 men. The fixed cost is, by comparion to BD/Core next to nothing. We know the marginal production cost is in the order of 4-5usd. What this means is, that even when Intel will have a far superior product in broadwell, they can in no way compete on price. The job can only be done by Atom if Intel is going to pay for the Capex.
If the economy continues to be in a slump and perhaps be even worse, that could give AMD some comparative advantage here. Not that i can see where they are going to get their profit from on their total busimess.
How do we know AMD have to move Kabini to GF?