There are several technology limiting factors playing in the x86-processor field these days:
1. Windows is the most used platform on consumer x86-processors, and Windows has reached it's height on what it requires in terms of system performance with Windows Vista. Vista was so heavy and full featured, that the computers in the market were not ready for it. They would be now though, run Vista on a brand new computer and you'll like it for sure, but thta's another story. Basically, since W7, MS has been kicking out features to make Windoze lighter and thus faster, W8 has even less features and is therefore again a bit faster, but it's regression, it's not improvement. The result is that the hardware requirements for Windows are actually dropping since Vista, so people don't upgrade their platform as rapidly. Also - except in North-America - Windows is losing terrain fast to GNU/Linux, Microsoft says that a lot of Chinese are running WIndows, but how many of those are legal copies, and fact is that almost all branded systems sold in the Far East come with Linpus, a GNU/Linux distro, not with Windoze, and the Chinese government for instance has a deal with Red Hat (the American open source software giant) for a customised chinese official version of Red Hat that is called Red Flag Linux, and obviously that's a lot of quality to get for free, so why would anyone spend money on Windoze right. In Europe, it's the same story, the former Eastern Block countries have never experienced the golden era of Microsoft in the eighties, so most run GNU/Linux, also because they don't all have huge budgets to spend, and they'd rather spend it on tangible things that "do something" like hardware, than on some legacy operating system, and torrents are not really a problem in Europe or Asia, so everyone just pulls a cracked windows off a torrent to play equally cracked or even bought games with (most games in europe are cheaper on DVD than they are on steam or online, because online or on steam, the prices are US prices, but frankly, the game distributors would not sell a great deal of games if they would only sell at those prices outside of the US, in Germany for instance, games are expensive, because people have more disposable income, but in Belgium or France, games are much much cheaper, because otherwise they wouldn't sell any).
2.The same principle goes for hardware: because windows doesn't require a hardware upgrade in windows countries, and GNU/Linux certainly doesn't require a hardware upgrade, people just don't buy hardware upgrades. Thsi results in a situation whereby hardware prices have never been lower, because the market is so tough. Noone really needs the upgrade, except video renderers and gamers if they want to play the few titles that cannot run at max setting on 5-6 year old hardware (there are not a lot of games like that, most games run just fine on an Intel Core Duo E8400 with an nVidia 9800GT 512MB). At the same time, people want portability, and the popualrity of RISC-platforms running some kind of linux based operating system, has never been bigger. There has never been any CISC-platform capable of selling 1.5 million units per day like linux-based RISC-computing platform are doing now. That distorts the market big time, or better, it open a new perspective: at the moment, you can buy a full featured x86 CISC laptop for less than 300 USD, whereas a Galaxy S3 will cost you twice as much. However, the laptop is much more expensive to make, it has 9 Li-ion cells, the S3 only has a single small one, it has a large screen, the S3 has a small one, it has a lot of internal memory, the S3 not, it has an expensive large silicon processor, the S3 CPU and GPU assembly costs less than 11 USD to produce, it has a lot of material, keyboard, HDD, power supply, etc, the S3 needs none of that, and the S3 has built-in GPS and a bunch of cameras and a touch screen, which the laptop doesn't have. And linux makes it possible to have a faster and more fluid user experience on the S3 than you would have on the laptop running windoze. So obviously, Intel and AMD are questioning their CISC-CPU production. Supercomputing saw a revival of CISC-CPU's with the AMD Opterons because they were cheap, but that fashion seems over also, because of obvious energy constraints and the price pressure on the CPU's, the energy needed to run those CPU's is actually costlier than the CPU's themselves, so large computing facilities switch to RISC arrays to save energy, and guess what, you need to buy more of them to have the same performance, but those RISC processors, as production reaches enormous quantities, are becoming cheaper too. So again, Intel and AMD are questioning the huge costs involved in developing new CISC-processors, because we're getting in a situation where the only reason why they are bothering is because of Windoze-PC's, which only seems to be a necessary for the conservative consumer market in North-America.
I'm involved in high performance computing, and frankly, when I see the price of a RISC-based high performance rig, it's comparable to the price of a full blown CISC-based consumer rig, and for that price, you're getting a lot more hardware quality and performance (like at least 10 times the performance of an i7-3970 rig). Of course, this is still a long way from what the average consumer wants to pay for a system, but for hardcore gamers, it would fit their budget. So once games are not locked down on the Windoze platform anymore, which is happening right now, because frankly, making products only for the conservative brand-infested North-American market, is just not viable any longer, the preferred technology will switch to RISC very quickly. Now there is no doubt in my mind that the consumer preferred technology under the influence of Windoze will stagnated further in North-America, it is a very conservative market and brands are more important than specifications there, so Intel and AMD will probably keep making x86-CISC-CPU's for quite some time, and will keep concentrating on producing them cheaper, with smaller dies, less silicon, and for a more expensive price. And that is happening already. What game would not be perfectly capable of gaming - even Crysis 3 - on a 5 year old Intel Quad Core or even Intel Core Duo extreme, paired with a more modern GPU? Since Vista, we've been at a point where the CPU performance doesn't matter quite that much anymore on a windoze system.
Now Android is a linux pioneer project, it will eventually phase out like iOS does, but then Google hasn't invested quite the same amount, because they've pulled a lot of assets from the community, and got those for a really low price in comparison to closed source systems. In the end, the linux kernel will be used for other proprietary systems, and soon, RISC-devices will be running full blown GNU/Linux distros (which is already possible, I know, because I run my tablet with Debian ARM -in dual boot with Cyanogenmod Android for the time being-, and it works just fine), and a lot of people in Europe and Asia have working GNU/Linux distros for phones that work, but that cannot be implemented due to ISPs and commercial arrangements of communications companies, which are often owned or co-owned by North-American communications kartels that lock down the platform, but they will realise where the money is in the near future also. In fact, it's started already, because the EU for instance has put a legal cap on how much communications kartels can charge for mobile data and mobile data in roaming, and that will only become more strict in the future, which really helps normal competition, because prices for data plans in the EU are dropping constantly, especially since ever more users switch to prepaid SIMs, which gives them tariff freedom, because in the EU, a communications company cannot hold your telephone number for ransom, if you switch provider, you can take your number with you. That leads to a situation whereby people buy their devices instead of leasing them with a subscription, which puts pressure on the prices of the devices, and leads to a different offering of devices being available. For instance, the <150 EUR Android phone category is well represented everywhere but in North-America. Phones that were originally thought for china and india, are now best-sellers in continental europe, like the Huawei Y300. Nokia sells more System 40 Asha-series phones in continental Europe (which were originally thought for the far east and south-east asia and other developing markets) than they sell Windoze phones (the Nokia 900 series phones are dumped for less than 200 EUR through Aldi stores, which is a german chain of bottom price food stores that is very popular in continental europe, that is the phone without any contract and without any branding or odexing or other lockdown, yes there is no trace of such dumping practices in european online stores, because those could be seen by north-american or UK consumers that still pay over 500 USD for the same phone, locked down and with a cutthroat contract). I know this may seem unbelievable to North-American consumers, but there is a reason why Nokia has problems and Huawei has not. Japan is also very commercial, has an economy that is to a large extent a clone of the North-American economy, and look at what's happening there: Sharp has taken in Samsung as shareholder to survive, because they make the screens for Apple products, and those have been doing very bad lately, so Sharp is losing heaps of money, and the only thing they can still sell, are small screens for low budget devices, not retina displays, not super high pixel density, but small screens that can be made by just about anyone. Of course such a small display costs Sharp much more to produce than it costs some mainland china factory to produce them, so what will sharp do: they will incorporate the higher display technology in the small displays and will try to sell them as premium products, for a low price, because they have the technology to make these high pixel density screens, and the chinese not yet, so in the near future, samsung will be bringing out 150 USD android phones with retina displays, not in north-america, but everywhere else in the world. That's just how that goes.
IBM saw this coming in the nineties already. They have learned the most valuable lesson in the eighties, where they have paid a lot of money to microsoft just to be undercut and lose the entire PC market, so they have learned in the eighties that it doesn't pay to try to lock a market down, and when linux came in 1991, IBM jumped on it, they sold their legacy OS/2 Warp to Microsoft, that was looking for a system to replace the 20 year old MS-DOS at that time, and Microsoft made NT from that, and is still using it as a base for Windows, and that's now a 20 year old platform, just like MS-DOS made it through 20 years and then definitely was used up. IBM then continued with RISC and linux/unix, which was picked up by Apple, because the RISC-based PowerPC chip was cheap and efficient, and the unix-based platform was ready, cheap and performing well, and they started a new interface revolution with it, but Apple was not strong enough to fight the market lockdown Microsoft had firmly in place at that time, and IBM didn't want to stand on the barricades anymore. So IBM evolved further with ever better technology, one look at Watson and you know that they know how to make a computer, but preferred to stay in the shadow, they are one of the largest contributors to linux, together with Intel, RedHat and Novell, and they are doing well, they didn't go with Windows like HP/Compaq/Nokia did, and they sold Lenovo that was largely Windows-dependant, so they didn't have to sack thousand of people in the last couple of years and didn't have to change their CEO a thousand times, etc... Intel and AMD are now under extreme stress, because they see Broadcom and Atmel and Marvell and nVidia making these cheap RISC chips that are doing incredibly well, and they don't have anything like that yet. Intel is jumping on the peripherals market like a lion and is throwing out all support products for the CISC-platform, like their motherboards, etc... Intel and AMD are now pulling the plug on development of their existing CISC-based chips to jump on development of RISC-based processors, and they will succeed in the end, but they will not be on top of the heap at first, but with all the experience they have, especially Intel, they will find a technological edge sooner or later, that will differentiate them and push technological advancement of RISC CPU's into a maelstrom, leading to incredible performance increases of RISC CPU's in a short period of time.
There are a lot more factors than this, but in any case, everything is pointing in the same direction, and that direction is away from the CISC platform, away from Microsoft and other consumer-leeching kartels, towards a more natural situation of priority to technological advancement. I think that 3 years from now, we'll be rocking a RISC platform with the latest steam games, and will have our legacy CISC-desktop in the closet, to pull out when we feel like playing an old game. So in any case, I would not invest in expensive hardware anymore until about the end of 2014, when the shift will happen, except in North-America, where the technology in the consumer market will stagnate until at least 2017 if Microsoft doesn't shift to the subscription model for Windoze, and to 2015 is Microsoft does switch to a subscription model for Windoze.
Anyway, that's what I think, I could be wrong, but I hope I'm not, because I've invested in a modular scalable array system already for my high performance computing needs and I would love to be able to run games on it natively lol.