Arthur Hanson
Well-known member
hist78, about the holidays
Buying opportunity of TSM?
- Thread starter user_2013101
- Start date
user_2013101
Banned
I have reiterated repeatedly that TSMC will produce 70% or more of A9s. If confirmed true in the coming months, TSM will go up quite nicely. It’s a buying opportunity now.
user_2013101
Banned
Nothing new to me. I have been saying so since last November, despite that the one-sided media reports to the contrary.
So sorry but you are wrong. Please do not embarrass yourself further.
Brianhayes
New member
Thanks. The debate about the A9 is fascinating but has little influence on my decision. The Chairman said in the last earnings call that his top 30 customers were buying aggressively. He is spending $12 billion capex against customer commitments and he is delivering a CAGR over 10% (sales and profits) likely to continue for some time. Pacific Crest do not even attend his conference calls. So I bought more shares.
Not an investment tip I hasten to say as I am not qualified but I have been watching TSM as it overhauls Intel for about 4 years.
hist78
Well-known member
I sense that a "Word War" will begin soon
But before you guys start killing each other, can someone educate me a little bit about how can, if possible, an A9 (not A9X) be manufactured by two different manufacturing processes (14nm of Samsung and 16nm/16nm+ of TSMC)? Does that mean Apple is so rich that they actually have two teams doing the same A9 chip development?
And how about the final product assembly? Does that mean Foxcom will use A9 from foundry A and another one or two companies will use the other A9 made by the foundry B? For me the complexity is enormous for one assembler, say Foxcom, to deal with two A9 from Samsung and TSMC at the same time.
hist78
Well-known member
Can TSMC come to Intel's rescue? As long as Intel has a good design and can sell lots of that particular product, nobody really cares who actually manufactured the "product". IMHO, Intel doesn't need to manufacture everything they sell. The time-to-market demand and the ever growing foundry start up cost must be in Intel's senior management's mind.
Brianhayes
New member
I agree with you but they have been spending $billions on advertising their manufacturing superiority, which IMHO does not exist. A new strategy and CEO and a realisation that sooner or later they will lose the x86 monopoly and I would buy the shares. I am not holding my breath.
It is hard, but not as hard as you are thinking. I start from the bottom. Foxcom doesn´t even need to know from where a chip is coming, they only see the final package, so they simply do not care.I sense that a "Word War" will begin soon
But before you guys start killing each other, can someone educate me a little bit about how can, if possible, an A9 (not A9X) be manufactured by two different manufacturing processes (14nm of Samsung and 16nm/16nm+ of TSMC)? Does that mean Apple is so rich that they actually have two teams doing the same A9 chip development?
And how about the final product assembly? Does that mean Foxcom will use A9 from foundry A and another one or two companies will use the other A9 made by the foundry B? For me the complexity is enormous for one assembler, say Foxcom, to deal with two A9 from Samsung and TSMC at the same time.
Of course Apple can afford a double design. That actually was my guess. You know that companies like Qualcomm and Mediatek are already doing a massive use of second and sometimes also third sourcing, don´t you?
Not at 16/14nm they won't. In fact "massive use of second and third source manufacturing" has not happened since 40nm.
The same designs or new design starts?
If you compare TSMC 28nm wafer shipments versus UMC+SMIC+GF 28nm my guesstimate would be a 10:1 ratio. Even if ALL of those were the same designs second sourced, which they are not, it would be closer to "miniscule" amounts of second sourcing than "massive". Based on press releases however it would be closer to a 1:1 ratio.
I guess this is a glass half full versus half empty reporting?
TSMC's sales up 33.8% Y/Y in February (TSM) Seeking Alpha
TSMC Feb. sales down 28.1% month-on-month China Post
TSMC's sales up 33.8% Y/Y in February (TSM) Seeking Alpha
TSMC Feb. sales down 28.1% month-on-month China Post
hist78
Well-known member
Regarding the -28.1% M/M drop, the Seeking Alpha report said it's due to seasonality and three fewer days. Then the China Post's article mentioned "Analysts said with the number of working days returning to normal in March, " but didn't explain in detail.
I explained it in this thread:
https://www.semiwiki.com/forum/f302/buying-opportunity-tsm-5690-2.html#post19561
"In addition to February has 3 less calendar days than January, TSMC and all factories in Taiwan observed 6 days of Chinese New Year holidays. Also Feb 27 is Taiwan's memorial day. That means 25% of February were holidays in Taiwan. This is a significant impact to any factories (including TSMC) operate in Taiwan. "
February is not just three days less than January for factories operate in Taiwan. But I think many people outside of Taiwan may not understand this situation at all.
Last edited:
This is yet another rumor...but this one comes from Tom's Hardware....one of the more reputable sites on the Internet in my opinion.
[h=1]Most Of Apple's A9/A9X Chips To Be Manufactured On TSMC's 16nm FinFET Process[/h]Please remember that this is just a rumor (unfortunately).
[h=1]Most Of Apple's A9/A9X Chips To Be Manufactured On TSMC's 16nm FinFET Process[/h]Please remember that this is just a rumor (unfortunately).
I agree with you. I was just pointing out the different headlines covering the same story. One seemed negative and the other positive.
No, but TSMC will get the A9x which represents 30% of the Apple business. Next year TSMC will get the A10 and A10x that will be a big party!
hist78
Well-known member
I see. Thanks. I think a more serious problem is those online media/blogging outlets like to quote from each other. Sometimes it may end up altering the original statement or inflating the original meanings. My another observation in this TSMC vs Samsung A9 orders reporting is that often we can only see reports mentioned their source is "according to source close to XYZ" or "source close to the supply chain". The recent analyst reports regrading TSMC A9 at least have identified themselves as "Deutsche Bank - Hong Kong" or "Daiwa Securities analysts Rick Hsu and Olivia Hsu".
Last edited:
Arthur Hanson
Well-known member
From my own research, TSM is going to do well in a rising semi market due to the variety of semi technologies they deal with. They locate their fabs close together so they can share and shift resources easily (Not rumor, TSM has clearly stated this). This gives them several advantages over other fabs. Also Taiwan has a very large, concentrated ecosystem as far as suppliers, support companies and the educational system, all topped up by a very supportive government with many end users. All this and many other factors give TSM an edge others haven't mentioned. TSM has the most complete ecosystem on all sides of any foundry in the world and this is an advantage they use and actively cultivate. Dan and many other members of SemiWiki are a key part of this ecosystem.