I'm not sure how much faster they can go with current Intel and AMD designs, 5ghz, sure at some point, too much beyond that I'm less sure about. Higher speeds in them are FAAAAR too depending on die and trace size, you can only divide so many times before you reach zero. Sure you can eek it ever and ever closer down to an atom, but at some point reliability will also be an issue. When your voltages get that low and the traces become more and more fragile even the slightest power fluctuation would fry it. And that's only half the problem, you also have Moore's Law, each jump in performance requires far more computing power, because it multiplies. It's not 1,2,3,4,5 it's 1,2,4,8,16. Like spinning drives, I think we're nearing the end of the line for copper on silicon, it's certainly peaked if nothing else.
ARM was where myself and many others put our hopes, however if Nvidia buys them then that hope is dead and it seems others are starting to realize that as well. We've all seen how nvvidia acts when it has no competition and ARM has almost none. There's a few here and there but nothing on that level, and you can forget it on desktop. Sure Nvidia will push it but they'll also push the price once it gains traction. I consider them one of the least trustworthy companies, right up there with Comcast and big oil, I never trust them to do the right thing.
The greatest jump right now is in going to be in software design or a whole new architecture. We're starting to see the cracks in software as it was all designed to offset slow hard drives, now the drives are fast and the tricks that allowed spinners to run well are coming back to haunt us. I'm not sure how hard it will be for Windows to deal with this but I know Linux has been working on it for a while.
Mr. Squishy, can that board supply enough power to a 10900k? I know the chips fit and it can run newer chips but I'm not sure it can handle such a massive jump in power.