How long can intel maintain such a large process lead over rivals?

[Blades]

Assuming that is the case, it would be interesting to know why the equipment is so costly? Does it require high precision mechanics or similar? Does the equipment contain many more parts than the previous generation of manufacturing equipment? Or something else?

You can check out this video from inside the Intel 22 nm manufacturing fab:

http://www.youtube.com/watch?v=lrsPzbUJwl8&feature=related

Looks pretty high tech... But still most of it seems to consist of moving wafers and chips between different manufacturing steps (and verifying their functionality). But this was done in previous manufacturing fabs too. So I still don't see how the equipment can get THAT much more expensive with each generation...?

What is Auburndale?

 

[cytg111]

Intel is expanding its lead. And the competition is dropping of. New processes got a ROI to forfill. For example for 14nm only Intel and TSMC can actually make a profit out of it. Samsung is in a 50/50 area. And beyond 14nm for logic. Samsung is out for sure.

The others are already 3½-4½ years behind if you look beyond the node scale itself.

Intel could bruteforce a bad design and still win. But there are limits to how bad the design could be. Essentially Intel could produce a quadcore IB today with 3-4x the IGP power of the current HD4000. And still make it cheaper than what AMD can make its dualcore Trinity with CMT. Not to mention AMD would have to share profits with the foundry they use.

Another point is that Intel, unlike any other can design its manufactoring and chip design together for maximum performance, yield and so on.

Its simple economics tho. Intel spends several times more on chipdesign than AMD as well. AMD would need a miracle to beat Intel. While Intel could still beat AMD on a bad day. Its the same that happend to the mass of previous x86 makers. The scale of economics killed them off one by one. Thats what competition is.

So to your question. They cant compete.

- That is so friggin sad.

 

[happysmiles]

AMD is on the path to recovery and ARM is growing insanely fast.

The whole game is changing but I definitely agree that MS and Intel are overcharging.

 

[Arachnotronic]

AMD is on the path to recovery and ARM is growing insanely fast.

The whole game is changing but I definitely agree that MS and Intel are overcharging.

Huh? AMD's given up trying to take the high end space -- they're going to chug competently along at mainstream price-points. Not going to die, but they're not exactly going to "dethrone" Intel.

ARM? I'm sick to death of the hype around ARM and how they're going to "take the lead" and "beat intel". Anybody actually look at their finances? Revenues for the trailing twelve months were $800M with net profit of $200M. This company is not even close to a real competitor to Intel.

Oh, and Intel charges too much? $200 for a 2500K is an incredible deal.

 

[AtenRa]

Oh, and Intel charges too much? $200 for a 2500K is an incredible deal.

Im sorry to say it but after 16-17 months that 2500K has been released, $200 is not an incredible deal.

 

[frozentundra123456]

Huh? AMD's given up trying to take the high end space -- they're going to chug competently along at mainstream price-points. Not going to die, but they're not exactly going to "dethrone" Intel.

ARM? I'm sick to death of the hype around ARM and how they're going to "take the lead" and "beat intel". Anybody actually look at their finances? Revenues for the trailing twelve months were $800M with net profit of $200M. This company is not even close to a real competitor to Intel.

Oh, and Intel charges too much? $200 for a 2500K is an incredible deal.

I agree that cpu prices are incredibly cheap compared to historical levels. The only place i think Intel is overcharging is for six core chips, which they keep out of the mainstream to keep the price up.

However, I also think tablets are overpriced as well. I bought one for 200.00 and it was worth that. However, 500 or more dollars for some of the high end tablets is completely too expensive for something that is basically a content consumption device, and I am sorry, but Android completely sucks compared to windows.

 

[Fox5]

I was reading this recently and it really put the costs of such endeavors into perspective. Needless to say, it's a metric butt-ton of money required

Makes going fabless sound like a pretty good deal until you realize your arch rival has the best fabs and the most money

APUs may make more sense in this model, have one basic design that can be mass produced to meet market needs.

It probably sucks for AMD though, at least for CPUs, they're locked into Global Foundries if they want anything high performance, GF probably gets to charge whatever they want. Of course, AMD could just get out of the high performance game altogether and produce low power chips at TSMC.

As far as foundries, I think we'll see even more consolidation. Intel will stand on its own but open its fabs up to a few partners. TSMC will stand on its own, at least for a while. GloFo will continue to integrate with other small foundries, and possibly eventually bring Samsung into the fold, or be brought into Samsung's fold. In the really long run, TSMC might gobble up all the independent foundries though.

 

[ShintaiDK]

Huh? AMD's given up trying to take the high end space -- they're going to chug competently along at mainstream price-points. Not going to die, but they're not exactly going to "dethrone" Intel.

ARM? I'm sick to death of the hype around ARM and how they're going to "take the lead" and "beat intel". Anybody actually look at their finances? Revenues for the trailing twelve months were $800M with net profit of $200M. This company is not even close to a real competitor to Intel.

Oh, and Intel charges too much? $200 for a 2500K is an incredible deal.

^^ This. Also 64bit ARM CPU is _maybe_ coming in 2014. And even the fastest (Demo only) ARM CPU that TSMC made at 3.2Gz that uses alot of power cant even compete with regular Atoms/Bobcats.

x86 will kill ARM.

But hey, it doesnt sell as well as hypes and rumours

 

[ShintaiDK]

APUs may make more sense in this model, have one basic design that can be mass produced to meet market needs.

It probably sucks for AMD though, at least for CPUs, they're locked into Global Foundries if they want anything high performance, GF probably gets to charge whatever they want. Of course, AMD could just get out of the high performance game altogether and produce low power chips at TSMC.

As far as foundries, I think we'll see even more consolidation. Intel will stand on its own but open its fabs up to a few partners. TSMC will stand on its own, at least for a while. GloFo will continue to integrate with other small foundries, and possibly eventually bring Samsung into the fold, or be brought into Samsung's fold. In the really long run, TSMC might gobble up all the independent foundries though.

TSMC is better and bigger than than Glofo. Same reason your GPU comes from that. The size difference between TSMC and Glofo is essentially like Intel vs AMD.

Manufactoring wise it doesnt matter if you have 1 single design or 10 designs if we look away from the design cost. Its just another mask to run. Thats what TSMC and others do very cheaply. The only thing that matters is how many wafers you send through. (And get payed for.)

TSMC produces what equals to over 520000 12" wafers a month. Or 1.2mio 8" wafers. Fab15 should also get online around now. A 9.4billion$ fab.

 

[blastingcap]

^^ This. Also 64bit ARM CPU is _maybe_ coming in 2014. And even the fastest (Demo only) ARM CPU that TSMC made at 3.2Gz that uses alot of power cant even compete with regular Atoms/Bobcats.

x86 will kill ARM.

But hey, it doesnt sell as well as hypes and rumours

I am not an EE but maybe someone who is can comment on this: I heard that as process sizes shrink, leakage matters more and more as a percentage of energy expenditure, compared to switching, and therefore one may expect that ARM's energy efficiency advantage would shrink as well, so long as Intel maintains its process advantage (since ARM-based designs currently and for the foreseeable future will be built on larger processes than Intel chips, and assuming Intel can continue to press its process advantage). True or false?

 

[KompuKare]

Oh, and Intel charges too much? $200 for a 2500K is an incredible deal.

Oh no, that one again. Yes, for people who are used to paying $200, $300+ for a CPU the 2500K is a good deal. But to repeat myself again, there are a lot of people who have never paid more than $100 for a CPU. I know hardware forums are full of people who buy every new things, but...

For instance since the days of the Celeron 300A my budget has and thinking has been constant: I spend around roughly £50 for a CPU and the same for a mobo and always overclock 30-50%, but I only upgrade if I can get twice my previous performance.

Thing is atm, I've had my 45nm Pentium E5300 @ 3.5GHz for a good while because there is nothing new which offers me that kind of value. About the closest would have been the Thuban's but even they were above that budget and I'd really like better idle power.

 

[ShintaiDK]

Some obviously already forget what old prices was:
http://techreport.com/articles.x/8295/1

200$ for 2500K/3570K is a steal for basicly the fastest gaming CPU in the world.

 

[Arzachel]

There is a ton of misconception in this thread.

Intel isn't in a direct competition with ARM (the company), comparing financial stats is stupid. ARM sells licenses for their CPU arch, they don't have the overhead of actually manufacturing the chips, nor do they have anything to do with the platforms the chips end up with. Quallcom would be a closer comparison. x86 isn't going to "kill" ARM (the arch) either, because you simply can't license x86 which is a dealbreaker for quite a few companies.

 

[2is]

AMD is on the path to recovery and ARM is growing insanely fast.

The whole game is changing but I definitely agree that MS and Intel are overcharging.

I like people who make me laugh and you just made me laugh.

 

[Idontcare]

I am not an EE but maybe someone who is can comment on this: I heard that as process sizes shrink, leakage matters more and more as a percentage of energy expenditure, compared to switching, and therefore one may expect that ARM's energy efficiency advantage would shrink as well, so long as Intel maintains its process advantage (since ARM-based designs currently and for the foreseeable future will be built on larger processes than Intel chips, and assuming Intel can continue to press its process advantage). True or false?

Process-wise that is where you end up when projecting forward a few current rules-of-thumb.

What it doesn't capture is the gross-margin story, and whether or not Intel cares to compete against ARM products if ARM developers are willing to operate at sub-30% GM's while Intel's shareholders and BOD are expecting Intel's executive team to seek out business in only >50% GM areas.

We (us laypeople) talk about process technology from a technical perspective as enabling things - enabling lower power or higher clocks - but the reason the nodes are being developed is solely for gross-margin enrichment and ASP enhancement.

14nm would not be developed if Intel did not expect it to make them more money than 22nm. Their shareholders would not be happy to be told 14nm is being built so Intel can dominate the ARM space with 30% GM SKU's.

In the end it is the ARM developers that hold the trump card, they can operate at break-even if need be, they aren't expected to operate in the >50% GM regime.

 

[BiG K]

The problem is, when AMD had a better product than Intel, Intel was actively stifling AMD through giving computer manufacturers incentives to only exclusively use Intel chips. So AMD couldn't line it's pockets to fund further development in the later years. When Intel finally had a better product than AMD, it didn't matter that Intel got fined £1 billion, it didn't matter because Intel had very deep pockets by then. However it's the ripple effect from all of this, Intel has been pushed for so long by all the major retailers, that the average joe will simply go for an Intel now, hence the manufacturers offer more Intel products....which just gets worse for AMD, as AMD is only on the "cheap" computers and to the average joe, cheap = shit.

The thing is, if AMD were to die, there would be no competition to Intel whatsoever, then we, as the consumer are actually negatively affected. Without any competition, there is no incentive whatsoever to innovate. A minor improvement can be charged for a higher price. So we, as the consumer get hit with higher prices.

Look what happened with Microsoft. They dominated the market and then we had to pay £90+ for each version of Windows that was pretty much the same as before. They left Windows XP for so long, simply because what was the point in making something new? If someone wanted a new OS, they had to pick Windows. (I know linux is available but it wasn't exactly user friendly now was it?). When Apple caught up, Microsoft has dramatically shifted it's policy!
However int he processor field, the goalposts are so far ahead, we will NEVER see a new company coming out with X86 processors (as only a handful of companies have x86 licenses to start with) and then the cost to invest in fab, design etc will likely run to well over £20 billion and that is ALOT of money to invest in a company that might not succeed.

I have always rocked AMD processors, I can trade the 10% of performance gain in the knowledge that I am making a difference. Plus I just prefer AMD. :wub:

AMD K6-2 3D Now! > AMD Athlon XP 2100+ > AMD Athlon 64 3500+ > AMD Phenom 9550

And now I have an AMD E450 and an AMD A8-3820. I feel my life is complete having owned a K6, K7, K8, K10, Bobcat and Bulldozer!

 

[ShintaiDK]

The thing is, if AMD were to die, there would be no competition to Intel whatsoever, then we, as the consumer are actually negatively affected. Without any competition, there is no incentive whatsoever to innovate. A minor improvement can be charged for a higher price. So we, as the consumer get hit with higher prices.

I strongly disagree.

First of all Intel havent had _any_ competition since july 2006.

Secondly, Intel is forced to innovation not by AMD, but by consumers. If a new CPU is too expensive and doesnt yield a satisfying improvement over the previous. People wont buy it. Its that simple. And Intel needs to ship 100mio every quarter to reach the profit. Intels current position actually gave us the cheapest, fastest CPUs we ever had. In 2005 you would pay around 800$ for a 2500K/3570K.

Lastly, when we sit in a duopoly or lower in a fixed market we actually have a stalled monopoly. Neither AMD or Intel dares to do something that might seem too risky. Intel already tried with IA64, but AMD just pissed in the wind with x64 and won that game hands down. x86 was secured for another 30 years+.

Regarding to MS. WinME was a disaster, Vista was a disaster. And if Win8 fails as it most likely will. Heads will roll. Steve Ballmer will be gone. If MS gets lazy someone else just takes over and it goes fast. Stockowners know that. What keeps MS alive isnt its monopoly. Its the inability for the competition to do anything. Just look at the combined joke the *nix community accounts for. Someone gets angry over what editor is used or something and bum. New branch. And look how big Apple is now compared to MS in just a few years. If MS delivers another flop and Apple gets it act together. MS might be the little one in another 10 years.

While I am not in favour of very prolonged monopolies in innovative sectors. Having a chip monopoly for ~10 years would be a good thing to get the risky changes needed. Just like it was on the software side for 20 years.

 

[2is]

Intel has competition, just not in the high end segment. If you think not having AMD around would keep intel CPU prices where they're at you're diluting yourself.

 

[Ferzerp]

you're diluting yourself

Like a homeopath?

 

[chimaxi83]

Intel has competition, just not in the high end segment. If you think not having AMD around would keep intel CPU prices where they're at you're diluting yourself.

*deluding

 

[2is]

Like a homeopath?

Yes! At least 16oz of water.

 

[rgallant]

try beer

 

[Charles Kozierok]

Everything has its pros and cons, even HKMG, SOI (Both PD and FD) and Tri-Gate. You cant just say that GloFo or TSMC is 4 years back in lithographic process in general because Intel started HKMG 4 years ago.

I agree that everything has pros and cons, but some things have more pros than cons and thus become industry standard, and others are more balanced and become trade-offs or design choices.

SOI falls into the latter category. You can't really use that as an example of where Intel lags in process technology, because they had the ability to implement it and chose not to do so.

 

[Arachnotronic]

In the end it is the ARM developers that hold the trump card, they can operate at break-even if need be, they aren't expected to operate in the >50% GM regime.

Right. Qualcomm, NVIDIA, TI, and Samsung shareholders would be happy with break-even operation

It'd also be stupid for Intel to hit a wall with their process development ; saving gross margins now at the expense of keeping a technological lead will only result in Intel becoming an irrelevant player in all of their segments.

 

[blastingcap]

Process-wise that is where you end up when projecting forward a few current rules-of-thumb.

What it doesn't capture is the gross-margin story, and whether or not Intel cares to compete against ARM products if ARM developers are willing to operate at sub-30% GM's while Intel's shareholders and BOD are expecting Intel's executive team to seek out business in only >50% GM areas.

We (us laypeople) talk about process technology from a technical perspective as enabling things - enabling lower power or higher clocks - but the reason the nodes are being developed is solely for gross-margin enrichment and ASP enhancement.

14nm would not be developed if Intel did not expect it to make them more money than 22nm. Their shareholders would not be happy to be told 14nm is being built so Intel can dominate the ARM space with 30% GM SKU's.

In the end it is the ARM developers that hold the trump card, they can operate at break-even if need be, they aren't expected to operate in the >50% GM regime.

Thanks, though I'm not sure Intel wouldn't be willing to suffer lower margins if it meant cementing the domination of x86 into the mobile computing era. Especially when even Microsoft is developing an OS running on ARM. Short term pain for long-term gain. Or do you think that Intel will allow ARM to continue to dominate mobile and maybe move on up into servers and even desktops, for the sake of preserving gross margins in the near term?