Not sure on the GPU. I do think it's a matter of time before they're competing for top spot, rather than being in the middle of the pack (or wherever they are), but GPUs rely heavily on process tech. If 14nm and 10nm continue to be delayed, they won't be able to steal that spot. What Intel really needs to do is focus less on compute and more on texture and fillrate performance. Once they've done that, they'll have a very competitive product.
But yeah, Intel has plenty of customers already. Bay Trail's doing really well. They still have a lot to go before they begin to make a profit, but it's inevitable at this point. The T100 was a slam dunk, and if they get the new Nexus, they'll be all set. There's basically no question that they'll get a Surface win this coming year as well.
Smartphones are still "iffy," though. It's unclear whether Merrifeild will do well, as much as I wish it would (I want x86 in a phone so badly). It's coming out way too late, unfortunately. Regardless, they should start displacing Qualcomm. Once their cheaper offerings debut later this year, they'll take the smartphone world by storm. But we won't really see that until next year.