That's the thing though. They are going to (or have already started with Medfield) attack mobile computing with x86. It will take some time to be successful, but it will happen.
We have been hearing that for years. Just wait until next gen, they will suck less! It gets a bit stale after the first four or five years of hearing about it. The Razer I does well in SunSpider, and that's about it. Intel can show up late to the game and lose against mid range devices in ~90% of benches while offering no real benefit? Intel still sucks, badly, in the ultra portable space. Medfield, Cedar Trail, Cloverfield- Intel's failures when they try to compete against ARM just keep growing.
That is problem one, Intel just plain sucks at making an ultra portable processor. The greatest achievement they have managed to obtain is not sucking horribly in every way.
Problem two, ~50% of the UP market is controlled by people who control their own SoC IP and have the means to get it made themselves. Apple and Samsung own their own SoC designs, Apple can get it fabbed themselves while Samsung can actually fab it themselves. You could also throw Sony into this group when looking at people capable of keeping everything in house(although as of right now Sony isn't doing it for their UP devices yet).
Problem three- Intel expects ~60% margins. While people can expect pie in the sky type philanthropy from Intel the reality is that Intel, their board and their share holders expect them to keep in the range of 60% margins. That isn't a sustainable goal in the UP market. Do you spend R&D time, resources and fab space to go after a market where the margins are half of your norm? What benefit does this have when moving the company forward? x86 ceased to be the top computing platform some time ago, ARM has had the mantle for a while now and the rate is accelerating. At some point Intel may be in a position where they are forced to accept significantly lesser margins, but everything to date indicates that if they are going to enter this segment, they are going to do it on their terms which indicates they will continue to demand high margins for parts that aren't very competitive at all. This isn't going to work to gain them any sort of traction.
Problem four- graphics IP. Intel has nothing to offer for competitive GPUs, they are going to be forced to rely on licensing from PowerVR- Mali isn't an option and neither nV nor Qualcomm are going to help them out there. The reason this creates a major problem is because Intel's CPUs in the UP space are already *huge* compared to ARM offerings, PowerVR's GPUs are even bigger. In order for Intel to bring a part to the market that has any hope of being realistic from the battery life perspective they are going to be forced to use long outdated PVR graphics paired with their monstrous CPUs. This isn't something they can really do much about as their graphics IP and development of it is *very* far behind what nVidia and Qualcomm have, not to mention ARM's own Mali division. ARM has the advantage of designing to be power effective from the ground up, Intel's entire GPU goal has been to catch AMD/nVidia as much as possible and it shows in terms of their designs and the amount of die space it utilizes.
Problem Five- Developers are already going ARM native. On Apple it has been this way for a while now, Android devs are also starting to push native ARM development. Intel is already dragging around transistor baggage in terms of x86 to uOps translation hardware, having to deal with another abstraction layer converting ARM code to x86 is going to hamper both their performance and their battery life. The only way around this is to give yourself enough marketshare to stop this from happening, but it is already too late on that front. The Windows mobile platform is their best bet, but to say that has been been a failure to date would be a bit of an understatement. While some people on these forums like to ignore it, BlackBerry is still doing better then Windows Phone. Trying to make decent headway before this trend becomes concrete is paramount for Intel, unfortunately it may already be too late on that front.
Besides that though, Intel should be fine