Take the 2-minute tour ×
Electrical Engineering Stack Exchange is a question and answer site for electronics and electrical engineering professionals, students, and enthusiasts. It's 100% free, no registration required.

Latest chips from Intel were 22nm, and they are now targeting 14nm (coming in 2014 or 2015 it seems).

On the other hand, Global Foundries or TSMC struggle since several years to go beyond the 28nm for their clients (mostly Nvidia and AMD, as I can tell from reading news about GPUs). They have delayed 20nm for quite a long time.

Why is Intel (or looks like) so advanced compared to GloFo or TSMC ?

share|improve this question

closed as primarily opinion-based by The Photon, PeterJ, Chetan Bhargava, markrages Jul 8 at 0:40

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.If this question can be reworded to fit the rules in the help center, please edit the question.

    
My initial reaction is the P word. –  Majenko Jul 7 at 23:15
2  
If you want a list of all the technical challenges to implementing 28 nm or smaller, this question is too broad. If you want an economic answer, it's just that Intel spends more money on new equipment. If you want to know why the foundries don't spend more money, you're asking us to speculate on things that either (a) we don't know or (b) we can't talk about. –  The Photon Jul 7 at 23:24
    
@ThePhoton I had no expectations about possible answers, but the economic one from tcrosley seems good. –  teh internets is made of catz Jul 8 at 0:07
    
@Majenko Could you make your point clearer, everyone's not a british native here. –  teh internets is made of catz Jul 8 at 0:09
1  
I am not so sure that Intel's advanced design and processing capabilities are so much of opinion as suggested by the folks that voted to shut down this question. –  Michael Karas Jul 8 at 5:02

2 Answers 2

up vote 7 down vote accepted

It's all about size. Intel has recent revenue of $48 billion and 107,000 employees.

GlobalFoundires, which is a divestiture of the manufacturing arm of AMD, and manufactures chips for AMD, Qualcomm, Broadcom, and STMicroelectronic among others, has recent revenues of $4.6 billion and 13,000 employees. So it is about a tenth the size of Intel.

Meanwhile, Intel is the leader in R&D spending in the semiconductor industry -- they spend over twice as much in R&D ($10 billion) as the total revenues of GlobalFoundries.

TSMC, based in Taiwan, is in a little better shape (revenues of $20 billion and 37,000 employees) but it spends only 8% of its revenues on R&D ($1.6 billion) compared to Intel's 21%. So its R&D spending is only 1/7 of Intel's even though its revenues are almost half of Intel's.

share|improve this answer
    
It should probably be noted that Intel's R&D and other spending is not entirely for manufacturing. They also design processors. Furthermore, Intel has mostly concentrated on high performance logic (CPU-oriented) process technology whereas most foundries have more diverse requirements, which would tend to dilute R&D investments. (Somewhat recently they have opened their manufacturing to outside designs; combined with the greater SoC focus brought by increasing integration Intel is likely to become more foundry-like.) Also, the Common Platform Tech. Alliance (GF, Samsung, IBM) shares R&D. –  Paul A. Clayton Jul 8 at 4:20
    
@PaulA.Clayton Good points. It sounds like, despite the R&D dollars siphoned off for processor development and other miscellaneous activities, the dollars spent on improving manufacturer processes of high-end microprocessors keeps it ahead in the nm-size war which the OP was referring to. –  tcrosley Jul 8 at 7:30

To get a little more technical and expanding upon tcrosley answer (which is good btw), there are two ways a processor can really go. Either a company will:

  • Invest into a higher speed core for the processor or
  • Invest into more cores of the same speed

Intel has taken the higher speed core course. They have invested in having much faster processors over just adding more cores (like AMD).
Because of the x86 architecture and its history, many many many programs do not take advantage of the multiple core (eight-core) that AMD offers, which Intel has realized. Their investments have increased in better and smaller transistors which allow for better cores.

While the multi-core is great, there are not many applications that utilize all of them but on the off chance you need them, Intel has included multiple cores in the processors for PCs as well.

All in all, this is a huge result of the higher revenue of Intel

With all that said, there are so so so many factors that go into designing the die, some that are not apparent when looking at the specs of the processor as a whole.

share|improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.