• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Anyone else planning on entering the PC CPU manufacturing market any time soon?


Golden Member
May 1, 2001
I'd guess "no" since the PC market seems to be declining, and AMD has such struggles to make money, but thought I'd ask just out of curiosity. I miss the days when Cyrus (I think that was its name) was in the market and giving Intel fits performance wise (at least as far as integer performance, think Intel always rules at FP performance).


dark zero

Platinum Member
Jun 2, 2015
The PC and Laptop market are dying.... let them die, then Phone market will die for a while..... the innovations is long time gone....


Mar 27, 2009
Intel is engaged in some form of atom core licensing to Rock chip and Spreadtrum.

I know the Rockchip SoC is fabbed at TSMC.

That is about as close as we are are getting (that I know of) to new players in the PC market.

EDIT: Apparently Intel Quark is also open.


Golden Member
Dec 17, 2008
Not on the high end such as desktop and traditional mobile form factors.

Many people enter and leave the cpu design market for things on the low tdp and low end price point with stuff such as arm cpus in android, apple, chromebooks, linux etc. These cpus while not x86 are extremely high on the price per performance front that you actually see Intel offer cut rate prices on their cpus in those markets and this also spills over into the windows front which for the most part requires x86.

Right now what is separating a atom tablet with a 22nm baytrail 1.8 ghz for $99, a 22nm baytrail atom 1.8 ghz for $199, a 14nm atom cherrytrail x5 at 2.24 ghz for $299, and an 14nm atom cherrytrail x7 2.4 ghz at $499 is not the cost of the cpu but artifical market segmentations by the OEMs (and to some extent intel but mostly the oems) where the $199 product has 2 gbs of ram vs 1 gb for the $99, and the $299 x5 atom has 4gbs of ram vs 2gbs for $199. That and differentation on things like storage space of the emmc of 16, 32, and 64gbs and whether the screen resolution is 1280x800, 1920x1080, 1920x1200, or some other higher number

The tray price for the best atom x7-Z8700 is $35 and that is prior to discounts due to bulk ordering, while the atom 22nm 3735 tray price is $21 before discounts with bulk ordering (the real price for this cpu is <$10 for oems). The atom x5 8300 tray price is actually less than the 3735 at $20 but due to lesser rebates you have not seen the oems complete the transition to the 14nm part just yet.

All of this is to say in the mobile space the big performance cost is not the cpu but instead the oems trying to actually make money and force you to pay more for the features you want, for the oems have now agreed unofficially on a rough price points for various performances. This is not collusion but merely market maturity where if you undercut your competitor you may gain marginal market share but the market share you gain does not make up for the less profit you obtain by lowering your profit per unit sold.


And yes atom in its current form now sucks and is no longer a market leader for things such as new arm skus are much faster in the mobile space when compared to atom. CPU designs such as the cortex A57 such as in the samsung 14nm exynos 7420 as well more custom implementations of arm's instruction set such as the latest ipad air and ipad pro chips are just flat out superior to the current atom. This arm dominance on the high end (and not the mainstream price/performance from) is only going to continue to occur until intel releases their architecture change for atom which will be q2 2016 so we are talking optimistically the very early half of 2016 for these cpus to be mainstream, with a few devices being out in q2 2016 due to the turnaround where it takes time for oems to release hardware.

And even then arm may still keep the high end due to things like the very talented apple design team on the cpu front, as well as things such as cortex a72


On the desktop and traditional laptop form factor intel is just king of the hill with intel either dominating or amd being almost compeitive in some market segements but still not superior to the intel option at specific price points (being equal is being competitive, being superior means you have to actually being superior).

And while this is disappointing for traditional pc users you have to remember how much cpu single thread performance you get with a $50 celeron, or a $75 pentium, or a $130 i3. A 2.8 ghz haswell dual core for $42 is not slow, it is no workstation processor but often if your budget is under $700 (for pc, monitor, os) your money would be better spent on things like ssd, graphics, etc, or you just do a minor bump up to one of the other cost competitive intel cpus. You go for a well rounded machine.

And due to the how mobile is focusing more on graphics and doing fix function hardware this same logic applies for tablet and cell phone cpus. Can we run this software on fix function hardware, can we use the gpu instead of the cpu, can we make the code more efficient so the experience is more snappy?

All of this is to say cpu performance in the markets you want is a well defined market that if we were talking stalks we would call these stocks "blue chip stocks." New business do not enter the same marketplace as blue chip stocks unless they have a competitive advantage that this competitive advantage would allow them to compete with the entrenched market and economies of scales that the existing market leaders in that industry have already entrenched.

It is in places such as cell phones and tablets that you see people willing to compete and push the margin but even these industries are becoming stabilized and not revolutionary and merely evolutionary improvements year over year. Yet you see new people enter these cpu markets for the actual cpu design can be handled by arm and you do not need a R&D lab for that, and the foundry business does not mean you need to spend resources actually fabbing the chip for you pay TMSC or another foundry to build the cpu whose blueprints you hand over combine with money, and they hand back working silicon.

But even in the cell phone and tablet marketplace you better be doing somethign that gives you a competitive advantage over the other cpu makers, some form of fix function hardware or necessary features that you need to have, or else you would be buying just off the market silicon from someone who has already come up with a design such as qualcomm, mediatek, intel, samsung, or even just the stock arm designs with mali graphics, etc.

There is most likely not going to be anybody buying into the desktop marketplace over the next 5 years unless they buy the almost zombie yet still treading water AMD.

The story of the almost zombie yet still not dead yet for I am treading water AMD is a very a sad story. But tears can't keep you alive nor can they bring back the dead or improve your balance sheet. Business in many ways is just like the games of thrones, you play the game and either you win or you die. AMD's problem is you can't make good enough products when your opponents for the most part are always making superior products and price them at similar prices to your inferior / good enough cpu at specific price points for what roughly 30 out of 35 years AMD has been making x86 chips for ibm computers and ibm clones vs intel. I am sorry but pentium 4 was not AMD's success but instead intel's blunder, and even during the times AMD had very competitive cpus if not slightly superior (such as the early days prior to core 2 with the AMD X2) Intel had economies of scale that allowed them to price their products so competitively AMD gain no effective marketshare or could capitalize on the success of some very talented engineers who yield a great cpu.


Apr 22, 2012
I even doubt you will see any new ARM manufactors as such. The entry barrier is getting too high.

For x86 there is simply no room at all.

Same in the GPU section for that matter. No room either.


Senior member
Nov 6, 2009
Don't know about you, but yes I have plans to enter the CPU market


Apr 22, 2012
Didn't some entity in China say they were investing $150 Billion in building chips?
Its actually 161 billion $. But its over many years.

They want to buy already established companies. The Micron deal for example failed. They may look at Qualcomm now or MediaTek.