Apple 'gets serious' about moving chip production away from Samsung - report - Page 3

 

First off, Q1 was 16T and Q2 was 11T which makes 27T, not 31T. Where you you pull that 31T figure out of? My figures are straight from Samsung's quarterly reports.

 

Nobody knows what Apple will spend in 2012, but their "contract" was apparently for $9.7 billion and expected to surpass that number. Unless Samsung turns things around Q3 and Q4 they won't be hitting that $60 billion they did last year. More like $50 billion, and unless my math is off, $9.7 billion is pretty close to 20% of $50 billion. If they hit $60 billion then Apple's share drops to 16%.

 

What you fail to mention is that Samsung's smartphone sales have been soaring the last 2 years, yet their semiconductor & display deliveries have actually been going down. So much for Samsung being able to make up the loss of semi sales to outside customers with increased sales to Samsung Mobile for their own Smartphones. And this has been going on before this news that Apple is starting to ween themselves from Samsung. Even HTC announced last week they were going to reduce orders from Samsung.

 

And all those billions are going into the pockets of Samsung competitors.

 

The A15 isn't the savior of SoC's. Even according to ARM the A15 is only 40% faster than the A9 that's currently powering the GS3 and other phones. It's an improvement, but people expecting the A15 to blow away everything else will be disappointed.

 

Based on the information so far there's no reason to think the A15 will be any better than Apple's or Qualcomm's own custom ARM implementations. And both Apple and Qualcomm have managed to ship actual devices in consumer products while Samsung's, Nvidia's and TI's A15 processors aren't due until the end of 2012 (not likely) or early 2013 (very likely).

/dupe
Edited by DocNo42 - 10/16/12 at 8:34am

 

 

 

Ok, I'm getting a bit tired of your BS. You are now essentially making sh*t up and cherry picking data to support your nonsensical fanboyism. So why cite/compare unconfirmed component contract figure for unfinished fiscal 2012 sales when Apple doesn't give out quarterly breakdown of their component purchase? And why even bother with 2012 when we have complete data for the fiscal year 2011?  Let me explain why - your bullsh*t doesn't add up.  

 

Nope, you are making some more sh*t up. I already mentioned that "Samsung IM - IT & Mobile division - is still growing at 40%-50%". I also noted in my earlier comment that everyone in semi and display biz is losing money now (but that it's the "nature of their business" - ie, cyclical).  What's HTC got to do with this thread?  HTC made the same announcement back in 2010 that they were switching to Sony's TFT LCD (due to the AMOLE shortages), but it rather quickly ended back in the arms of Samsung again (Samsung bought out Sony's entire stake in Samsung & Sony SLCD venture in 2011). Samsung DP also makes 97% of all AMOLED manufactured worldwide, HTC has nowhere else place to go - we call that a PR stunt. 

 

 

		FY '12
		1Q		2Q
			OP%/Weight		OP%/Weight
Sales		45.27		47.60
	Semiconductor	7.98	17.6%	8.60	18.1%
	DP	8.54	18.9%	8.25	17.3%
	IM	23.22	51.3%	24.04	50.5%

 

 

FY '11
1Q		2Q		3Q		4Q
	OP%/ Weight		OP%/ Weight		OP%/ Weight		OP%/ Weight
36.99		39.44		41.27		47.30
9.18	24.8%	9.16	23.2%	9.48	23.0%	9.17	19.4%
6.51	17.6%	7.09	18.0%	7.08	17.2%	8.55	18.1%

Edited by tooltalk - 10/16/12 at 8:32am

Sigh... it has more cores or a newer reference design so therefore it must be better, right?

Good gawd - are you not paying attention? Android doesn't even take advantage of quad cores - the dual core iPhone 5 still bests quad core Android phones. What good is having a gazillion cores if your software can't take advantage of them?

And the last I checked, we are talking about mobile devices. More cores equals larger chips and also increased power use.

If you paid attention closely to the A6, Apple stuck with two cores, on an older instruction set, but due to hand optimization those cores are more efficient on a per clock cycle. It's better on power use than the reference designs. And it's physically smaller too. All significant advantages for mobile devices.

Mobile is a COMPLETELY different game than the "PC Wars" of the 90's. Size, power consumption and processor efficiency are all paramount. On the desktop physical size and power don't necessarily matter as much, but per core efficiency is still relevant since, especially on Windows, multithreading utilization is poor. Why do you think Intel spent so much time optimizing turbo boost to enhance single core performance? That kind of stuff may be OK for the desktop, but in mobile there simply isn't the physical space and essentially unlimited power budget as on the desktop.

It's pretty simple to do what Samsung did and take a reference design, rearrange it a little and kick out a new chip. What Apple is doing is not so easy to reproduce. When Apple does kick out an A15 or whatever design, it will be substantially different in power and performance because they aren't simply taking a reference design that has the "kitchen sink" approach to appeal to as many customers as possible, but a design that has a laser focus on just the aspects that are really important. As Apple continues to develop this expertise and refine their design prowess, this gap will increase even more dramatically. Competitors will no longer be able to simply pick up commodity chips to match Apple's hardware performance.

I know stories are focusing on Amazon and foundries, but I highly suspect they are after chip design and not production as much. Bezos is no fool and it's clear he's the only one that really seems to understand the real reasons and vision behind Apple's success. Android and Windows phone are a flash in the pan over the long term* unless someone steps up to match the end to end integration of Apple, and apparently where Amazon is going. Google's clumsy fumbling of Motorola and failing to do anything significant with it further underscores just how clueless they truly are.

* And that long term focus is where many people also completely underestimate Apple. I don't know why. They have demonstrated it twice now, first with the iPod, then with the iPhone/iPad. Ignore at your own peril!

Thanks,

I understand you much better now that you have shared this tangle [table, I mean.]

 

Are you serious? The current Apple A6 is basically equivalent to a quad-core Exynos 4. A 40% improvement on top of that absolutely would blow away everything else.

Intel wishes it could get 40% additional performance every time they changed the microarchitecture of their chip.

All the A15 dual-core designs (and Qualcomm S4) outperform A9-based quad cores.  So it's no surprise Apple's does too.  

 

A more fair comparison would be the A15-based Exynos processor in the new Chromebook.  Qualcomm's S4 quad-core will be appearing in phones soon too, which will likely outperform the A6 in multi-threaded applications...