Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/tsmc-will-have-the-lion%E2%80%99s-share-of-apple%E2%80%99s-a9-pie-citi-analysis.5619/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2020771
            [XFI] => 1050170
        )

    [wordpress] => /var/www/html
)

TSMC Will Have The Lion’s Share Of Apple’s A9 Pie: Citi Analysis

hist78

Active member
It's like watching TV series that every week has some new twist.

TSMC Will Have The Lion


"We believe that 14nm strategic collaboration between SEC and Globalfoundry (GF) is way too early to play out in 2015. The technology transfer across the companies and the product qualification period would take much longer than is generally expected. In our view, it is too late for GF to ramp up 14nm capacity and capability to meet the production schedule for Apple and any customers’ new product launch cycles in 2015. Stripping out GF’s capacity, the conservative plan of SEC to expand 14nm FINFET capacity seems to be tight in terms of fulfilling demand for its captive Exynos AP and Apple’s A9, as well as for Qualcomm (QCOM) or other key customers."
 

Daniel Nenni

Admin
Staff member
Thank you for posting this. We can look back on it and remember who was right and who was wrong. Not that it matters to them. By the way, the Samsung Exynos demand and QCOM demand overlap since Samsung is QCOM's #1 customer.
 
Last edited:

hist78

Active member
Thank you for posting this. We can look back on it and remember who was right and who was wrong. Not that it matters to them. By the way, the Samsung Exynos demand and QCOM demand overlap since Samsung is QCOM's #1 customer.

You pointed out some interesting situation that Samsung will face:

1. Samsung can use more Exynos processors on Samsung's smartphones to reduce the dependency on Qualcomm. But the more it makes Exynos the less capacity Samsung can have to make Apple's A9 processors. Save money to pay Qualcomm or lose(reduce) contract revenue in making A9 for Apple?

2. Samsung can also scale back the production of Exynos processors and use more Qualcomm's Snapdragon 810. But wait, did we hear that Samsung dropped the Snapdragon 810 for using in the upcoming new phones is because S810 doesn't meet Samsung's quality and performance expectation?

3. If Samsung wants grab the opportunity to get as much orders as possible from Apple A9x then it needs more capacity. They can use more Qualcomm's S810 made by TSMC. Not only Qualcomm will be very happy to accommodate their needs but also TSMC will benefit greatly from the increased S810 production.

So for Samsung to get more A9x orders from Apple and at the same time to sell more high end and high margins Samsung's smart phones, Samsung really needs TSMC's help! Strange and interesting!
 

Daniel Nenni

Admin
Staff member
Do you know what the combined Samsung and GF 14nm capacity is? That would be the key to this debate it seems.
 

hist78

Active member
Do you know what the combined Samsung and GF 14nm capacity is? That would be the key to this debate it seems.


Good question. Both 14nm and 16nm are not out yet and I'm not sure if there is any reliable capacity number for Samsung and TSMC. According to the Citi's research, Samsung alone can't supply enough A9 processors to Apple.

According to the following report: "(Apple’s foundry partner has to prepare 60-70k wafers per month (WPM) to meet demand in the peak season. This translates to >US$10bn capex)."

Samsung And TSMC Will Share The Apple Pie: CLSA - Asia Stocks to Watch - Barrons.com

Samsung has spent huge CAPEX in the past several years but I believe that majority of them were allocated to their memory making business. Samsung memory division has a much lager revenue than the foundry's.
 
And then all (but one) Ax processors indeed turned out to be made by Samsung. What's your point?

All of Korean reports that Samsung had won and started A9 production, have been proven false. See post below:

https://www.semiwiki.com/forum/f293/samsung’s-14nm-mass-production-5591-post19040.html#post19040

Please also read the entire thread:

https://www.semiwiki.com/forum/f293/samsung’s-14nm-mass-production-5591.html

It may be helpful, if you can please do a bit of reading before posting.
 
Last edited:

hist78

Active member
Ladies and Gentlemen,

Before we all jump into some heated debate, please read that articles Daniel mentioned:

iPhone 6S on Release Date Confirmed with 14nm A9, DRAM Chips Made by Samsung

and another article the above article was based on:

Samsung to supply chips for new LG, Apple smartphones
http://www.koreatimes.co.kr/www/news/tech/2015/02/133_173939.html

I don't have problem with any forecast or predication, as long as they clearly identified sources or stated it's just a speculation by the writer. But I do have problems with these two articles because:

1. The ibtimes.com article Daniel mentioned uses "
iPhone 6S on Release Date Confirmed with 14nm A9, DRAM Chips Made by Samsung – Report" as the title. But after reading the second article (Korea Times) this "confirmed" article was based, I think the word "confirmed" is really a stretch from the original Korea Time's article and the "release date" is nowhere to be seen.

2. In the Korea Times article, it said: "In a related note, Apple has chosen Samsung to provide mobile processors that will be used to power the upcoming iPhone 6S which is to be released later this year."

I read this Korea Times article and I can't figure out who is the source of this "related note". I think our SemiWiki members have done a much better job than the Korea Times to identify the sources in most SemiWiki's posts.

3. The Korea Times article said: "For applications, Samsung will be responsible for 80 percent of the production of the A9 chipset, while Taiwan Semiconductor Manufacturing Company (TSMC) will take care of the remaining 30 percent."

Sorry, I thought 80% + 30% > 100% ?
icon9.png


4. Personally I'm not sure Samsung can have 14nm capacity ramp up so fast to cover all those orders from iPhone 6s, iPhone 7x, new Qualcomm chips, and Samsung's Exynos processors. And don't forget the rumored Nvidia and AMD's orders. Samsung must have either achieved an amazing efficiency or acquired some amazing capacity in order to support those several big customers. Or a third possibility: Samsung may indeed have won these major customers except the actual order quantity is much smaller.
 
Last edited:
Top