as I said fierce dog fighting for ARM business - second tier foundries like UMC and SMIC are encroaching on TSMC bread and butter business....
this is a long slow grinding process but Intel will prevail cherrypicking and leave the scraps to the rest -
Intel has no desire to put ARMH or TSMC out of business -
that is not there intention but if ARM or TSMC are in Intel's way they will do whatever it takes
No - I don't work for Apple....but let me explain (at least try to)
when Apple procures memory chips from Micron rather than Samsung - that is transparent - it does NOT matter - Samsung or Micron makes NO difference - you follow?
however that does not apply to the APU (SoC) aka NON transparent because Samsung uses a different (FinFet) process than TSMC -
you think for example the shape of the fins of Samsung and TSMC are identical ?
Apple has to tape out for Samsung as well for TSMC -
there is an Apple design that plugs into a Samsung process and there is an Apple design that plugs into a TSMC process....and both are not compatible aka NON transparent....
and that leads to a great deal of redundance for Apple....
TI has a 12 inch state of the art analog fab -
QCOM will probably follow TI's path and exit the (mobile) chip business.
Wow, you are smarter than Robert and everybody at ARM, Apple, Samsung, TSMC, QCOM, BRCM, TI, and the rest of the fab-lite or fabless companies? Maybe you should run for president. Or even better, you should start your own investment firm so we can all bask in your magnificence?
you obviously don't follow - I don't think Apple wants to depend just on one supplier for very good reasons - unfortunately Samsung / TSMC process is not transparent to Apple custom design -
very similar situation like gate first versus gate last ....NOT transparent
Robert is an idiot - see his comment TSMC cutting capex
How many suppliers does Apple use?
Imagine 10nm ...
Apple custom design would to have to tape out two designs - one for Samsung process and one for TSMC process
that would be "fun" - talking about redundance
do you follow - it's the interaction between design engineering and manufacturing
Is the foundry business dead ?
It is limping
Imagine Apple having to tape out two designs - one for Samsung process and one for TSMC process -
Apple uses custom design - now wouldn't that be "fun" @ 10nm -
talking about redundance
you think Intel is stupid?
Apple's nightmare -
just like I said .....multiple suppliers leads to redundance....Nenni won't tell you....
GF bought the process from Samung ...correct?
they try to "transfer" the process but struggle because of difference in toolsets - it just does not copy exactly ....
I agree that there always seems to be rumors like this. However, where there is smoke.... I think it only makes sense for Apple to keep trying to get other sources for their chips. At a minimum it gives them some negotiating power with samsung. More importantly having a sole source can be a weak link in the supply chain. It is always better if they can have multiple vendors providing the same component. If one vendor has a strike, or is otherwise unable to provide the component, it is good to have another vendor able to pick up some or all of the slack. So Apple should be working on this. Regardless of what we think about samsung as a competitor, it just not make sense to have them as the sole source for critical components of Apple products, even in the best case scenario.
Single source is just how the SoC business works. It would take significant manpower to tape out a chip on two processes in the same time as one process.
TSMC has a BIG mouth - I follow them for more than 15 years -
in Q1 2015 the capex run rate (annualized) was $ 6 billion!!!!!
Of course that's NOT sustainable but TSMC better show some money at this point it's more like "talk is cheap" -
ASML wins - EUV or not
actually EUV would be nice - otherwise the foundries would be completely out of business
Now, an analyst is reporting that Apple, in a last-second move, has yanked a significant percentage of its orders back from Samsung due to low yields at Samsung’s partner, GlobalFoundries.
I believe it - Samsung sold its "process" to GF - but toolsets among both vary - it's not transparent ....
that's why I mentioned Intel "copy exactly" a few days ago.....
chipmakers often run split lots for the same application - etch it on Lam or AMAT and the results would differ -
it's NOT transparent,,,,,and it gets only worse as geometries shrink
how many fabs does Apple need to supply its APU chips ?
this is becoming a soap opera
The dual-supplier strategy is common.
that applies for DRAM and NAND or displays - not for APU - it's not transparent for APUs - having two or more suppliers for APU causes redundance for Apple
You can find it on SemiWiki
the foundry business is dead -Apple is not the typical foundry customer - the "typical" foundry customers can not afford 20nm - not even talking about FinFet.
Apple probably sourcing from TSMC and Samsung - both use completely different tool sets - IOW
Apple has to deal with redundance - even GF's toolset is not identical with Samsung
You know why Intel applies "copy exactly" ?
how many will be able to migrate to 10nm ?
How many fabs does Apple need to support "14nm FinFet"?
3 or 4?
Of course capital intensity is going up - I am riding this wave for years....i
On the capital intensity I will give you the long-term answer and then come back to what we're seeing in 2015. The longer term answer and I'll refer you to some of the stuff we showed at the investor meeting. We do agree capital intensity is going to go up as measured by capital cost per square inch of silicon. We believe that through 14, 10 and with some insight all the way down the 7 nanometer we can offset that increase in capital cost per square inch of silicon by improving our density. And so we can keep the cost per transistor coming down at the historical close
The next question comes from Stacy Rasgon from Bernstein.
Stacy Rasgon - Bernstein Research
I wanted to try to verify something, just so I understand it. So, is that gross margin is down in the second half were your 14-nanometer start ups are coming down, your revenues up a bunch, your tablet profit-loss is again lot worse, obviously you’ve 10-nanometer ramping. And you talked about reducing your 22-nanometer utilization in building inventory. So, that suggests that utilization has to be coming down a lot in the second half. You got to be building inventory a lot in Q2, but it's mostly 22-nanometers that you’ll then be selling in the second half when Windows 10 and Skylake are launching. Do I have those dynamics correct? And then what is the risk that you’re going to be building inventory I guess in the first half for Q2 that will be tougher to sale in the second-half given the new products that ought to be launching at that that point?
Stacy Smith - CFO
I think that the thing in that isn't accurate and is confusing…
Stacy Rasgon - Bernstein Research
some irrational bashing - he just can't follow .....corned like rat
Lowering CAPEX means less production .
So let me start and then I’ll let Stacy jump in. What we said was that this year’s capital is a little bit less than what we would have normally projected for the year. We were able to get as we said better yields better utilization and we’re able to -- and this happens from time-to-time. You can look back at our past it happened in the '08, '09 and from there we get that there is a slight increase in our projections for total volume at the same time we're going to a transition we’re able to move capital very-very quickly.
Lowering CAPEX means less production .
you're an idiot - you're a clueless day trader -
my advise read the transcript
there was no surprise about capex cut backs - that's for the dummies.
what was expected ...dummy ?
you have NO clue - just the usual garbage posted on yahoo
certainly lower demand but coupled with excellent yields led to that decision - Intel feels very comfortable with its technology lead
busy tonight - TSMC trucking in water ?
only a moron would pack that many chip plants on one tiny island -
intel is not delaying 10nm
don't bother to respond -
take your Alzheimer meds