Some would say Intel is already "winning" with Atom, but the real prize for Intel -- as Otellini himself has stated -- is in SoCs.
It's not a simple question of whether Intel can maintain its margins. It's more a question of how Intel will maintain its margins. More specifically, what IP is on that piece of silicon?
In the past, the answer was simple: an Intel microprocessor core, or two or four or eight of them. Going forward, the answer will include a big laundry list of things you're not accustomed to seeing integrated on Intel silicon.
Whether Intel "wins" with Atom design wins or not, it is likely that Intel's profit margins will suffer. CPU power on this end of the market is much more concerned with watts (or preferably milliwatts) than gigaflops per second, and the closer the chips can get to the cost of the silicon they're made of the better. Both of these metrics will make Intel very uncomfortable.
Apple must be watching these numbers and will looking forward to get part of it. Since Apple has successful design of A4 processor, will they be looking forward for new processor to compete directly with Intel? If they can design one better then Intelís, market will change dramatically and consumer will reap benefits. Also, we may get better applications software. Intel must be carefully watching this development.