Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/buying-opportunity-of-tsm.5690/page-3
        )

    [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] => 2021770
            [XFI] => 1050270
        )

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

Buying opportunity of TSM?

This is a very interesting thread, for it makes it very obvious that SemiWiki is a power in its own right. It is also because of the way it's structured, power with built in checks and balances set up in such a way as to bring out the best in people, without reservation. Coming from totally outside the semi industry, I have seen and experienced this far more than most. My first interest in the semi industry was an investment in TSM many years ago, with just a few to follow. It is a nexus of information, debate, educational resources and conference calendar in real time since it's online and open. I have posted several questions over the time I have been a member and except for one question on using light, always received several expanded explanations and answers. Many times these answers have included excellent background and perspective. I'm honored to be a member of the community and to contribute to it's growth in size and scope as most members strive to do. Thank you to the Founders and Members, the best is yet to come as semiconductors and spin off technologies expand the already key part they play in the world.
 
My outlook that TSMC to build most, or all, of A9s has not changed. Apple probably had made the decision around November last year. We outsiders certainly don’t have insider info. But, circumstantial evidences are abundant for such outlook.

Last November, TSMC board approved special US$5.5 billions of capex.

2014/11/11
TSMC Board of Directors Meeting Resolutions

On Christmas day last year, someone broke out the news that GlobalFoundries delayed and warehoused the equipment deliveries.

On January 15, TSMC announced that the capex for this year to increase 20% over last year, reaching $11.5-12 billions.

Starting in January, Exynos 7420 was reported as a 14nm chip, versus reported as a 20nm chip before January. Early this month, at MWC, Samsung officially confirmed that Galaxy S6 will be entirely powered by the 14nm Exynos. My interpretation is that Samsung had lost the A9 orders and its 14nm capacity has been moved to produce Exynos, instead.

Samsung has no increased hiring at Austin, TX, where the A9s are supposed to be built, while TSMC has been increasing hiring by the thousands.

[h=4] Strange sentiment [/h]
No matter how many evidences I cited, most people choose to believe Samsung’s media misinformation – all based on unnamed "industry source" in Korea or "people with knowledge of the situation."

The following is an example of such fabrication.

February 4, 2015
Apple Taps Samsung to Manufacture A9 iPhone Chip | Re/code
Click to expand...

[h=4] The irony [/h]
The irony is that the relentless Samsung propaganda had actually helped TSMC.

Comparing to 100% of A8 production, it should be considered a setback if TSMC produces 70% of A9s this year. However, since Samsung has created a parallel reality that TSMC had lost most or all of A9 orders AND fallen miserably behind Samsung in the cutting-edge technologies, now 50% of A9s is considered a great win for TSMC.

Most likely TSMC will produce 70-100% of A9s. Samsung still leads in memory chips, but realistically it probably had lost the race in the logic chips.

Conclusion: TSM price has great upward potential.
 
Last edited:
It is a surprise to the reporter; but no surprise to me at all.

March 16, 2015

16nm FinFET game on; TSMC has started to aggressively ramp 16nm as we think it has won the majority of AAPL after all– Field work indicates that in the past week, TSMC has started to aggressively pull-in 16nm order and shipment timing with SPE suppliers across the board.

From Barron's:
Surprise! Taiwan Semi Set to Make Apple’s ‘A9′ Chip, Says Cowen
Surprise! Taiwan Semi Set to Make Apple
Click to expand...
 
Bernstein Research, who previously proclaimed that Samsung is the primary supplier of A9s, now said TSMC will produce 70% of A9s.

However, these ANAL-ists still maintain that QCOM will switch to Samsung, who not only abandoned SD 810, but also plan to sell the 14nm Exynos to other phone vendors to directly compete with QCOM.

But, hey, QCOM loves to shift to Samsung as the primary production partner?

March 19, 2015

Bernstein Research:

Previously we modeled Samsung to be the sole supplier of the next-gen iPhone processor (A9) at 14/16nm, with TSMC retaining the processors in the next iPad (A9X, 16nm) and low-end iPhone.

However, our recent checks suggest that Samsung experienced new yield difficulty recently.

We now estimate TSMC share in AAPL’s overall business will be ~70%, instead of ~50%, starting 3Q15.
...
This will be the first time that AAPL has split the foundry service of the same processor at two suppliers. We think AAPL now has a very strong design team to make the variance of the chips at two foundries small enough so that regular end users can’t tell.

TSMC, Samsung: The Latest Twist At Apple, Qualcomm - Asia Stocks to Watch - Barrons.com
Click to expand...
 
Last edited:
I'd take anything like this from analysts with a huge pinch of salt; not only do we not know what their motives for such a statement are (talking up investments?), anyone who says that the same chip can be made in Samsung and TSMC "so that regular end users can't tell" is smoking something pretty strong -- the processes are different in both density and performance, it would take two separate designs and layouts and the chips would end up with different speed (TSMC faster?) and power (Samsung lower?). I don't think even Apple would do two separate developments for the same chip in two different processes -- A9 in one and A9x in the other, very possibly -- A9 in both, very unlikely.

user_2013101 said:
Bernstein Research, who previously proclaimed that Samsung is the primary supplier of A9s, now said TSMC will produce 70% of A9s.

However, these ANAL-ists still maintain that QCOM will switch to Samsung, who not only abandoned SD 810, but also plan to sell the 14nm Exynos to other phone vendors to directly compete with QCOM.

But, hey, QCOM loves to shift to Samsung as the primary production partner?
Click to expand...
 
IanD said:
I'd take anything like this from analysts with a huge pinch of salt; not only do we not know what their motives for such a statement are (talking up investments?), anyone who says that the same chip can be made in Samsung and TSMC "so that regular end users can't tell" is smoking something pretty strong -- the processes are different in both density and performance, it would take two separate designs and layouts and the chips would end up with different speed (TSMC faster?) and power (Samsung lower?). I don't think even Apple would do two separate developments for the same chip in two different processes -- A9 in one and A9x in the other, very possibly -- A9 in both, very unlikely.
Click to expand...

I just thought about what most people (including me) perceive that Apple will have only two new iPhone processors coming out in 2015. But wait, Apple can name the new iPhone processors whatever they like and whatever number of the sub-model names for this A9 category. Apple can also name the new iPhones with all kinds of model names depending on the screen type, screen size, case material, case color, waterproofing, battery life/size, camera types, memory amount, countries/regions of the phone sold, and most important, the processor's sub-model, etc.

For example, a new processor made by TSMC 16nm+ can be called A9t and used only on those new large screen iPhone models and sold only in mainland China and US while Samsung's 14nm produced processors can be called A9m and used only on phones with red or silver color cases named Extreme Edition.

As long as Apple calls those new iPhones (with different processors) with different model names and customers can reasonably tell they are buying different models, I think the potential problems are much less than we thought. And remember that today various iPhone models are assembled at multiple factories/companies and multiple production lines using various Apple A* processors already.

Intel can have all kinds sub-model names for an i7 or Core M series. Samsung can use different processors for the same phone series in different world regions. Then considering Apple's huge volume and market reach, they do have some flexibility to do whatever they feel right, just like Intel and Samsung already did.
 
Last edited:
There is new news as regards TSMC 20nm sales from the Credit Suisse Annual Asian Investment Conference:
Business has slowed, raising customer inventory levels. TSMC has noted a slowdown in the past 4-5 weeks due to US$ strength impacting European and emerging market purchasing power. As a result, the company’s near-term demand is being reduced across multiple nodes and end markets (PC/mobile/tablet/communications). As a result of lower end demand, it now believes inventory will be a few days above seasonal exiting 1Q15 so may take some time to work down. It now expects foundry may not grow the 12% it previously expected, but it believes TSMC will still grow a few points above the industry due to market share gains. 2. Some impact from the Qualcomm design loss and slower high-end Android – few hundred million impact on capex. TSMC acknowledges some impact to its 20nm revenue due to Samsung using its own processor over Qualcomm for the Galaxy S6 and also weaker high-end Android demand for other global brands. It believes 20nm will still double YoY (from US$2bn to US$4bn – implying about 13-15% of sales) vs. original target for 20% of sales. It will now do tool reuse to convert more to 16nm, citing that it has 95% equipment reuse. The company can save a few hundred million on capex through the reuse and conversion of 20nm capacity to 16nm.
Click to expand...

TSMC Says Business Has Slowed In The Last Month At Credit Suisse Conference - Asia Stocks to Watch - Barrons.com
 
[h=4] TSMC: The Writing Is On The Wall – Sell [/h]

That’s the title of an analyst report. Linked below.

March 25, 2015
TSMC: The Writing Is On The Wall
Click to expand...

And, not surprisingly, old rumors against TSMC returned with a vengeance, such as QCOM to use Samsung as the primary foundry, etc. Now even MediaTek is abandoning TSMC, if you believe the media.

Periodically, such bearish reports cause corrections, or mark the correction bottoms. This time, the decline, however, is more widespread. Over 2 days 3/25 and 3/26, semi index SOX down 5.92%; TSM down 5.3%. INTC closed below $30 on 3/25, the first time since last June, but bounced back up the next day.

SOX and INTC rose on Friday 3/27, but TSM declined further for a total of 6.3% in 3 days. Now at the low of March 10.

TSM - SharpCharts Workbench - StockCharts.com

On January 15, TSMC said the foundry industry will grow 12% this year and TSMC will be several percentage points above the industry. Therefore, analysts set the estimate at 15% or higher.

On March 24, TSMC’s CFO revised the outlook that, due to slowdown in the past 4-5 weeks, 12% industry-wide growth may not be achieved, but TSMC will remain higher than the industry. See report linked below.

In the same report, TSMC’s comment on the 16nm node appears to confirm that the company has won most of A9 orders.

March 25, 2015
16nm confidence increasing and scaling up quickly. TSMC believes 16nm will register revenue starting in 3Q15 and still be a steeper ramp than 20nm (which ramped from <1% in 2Q15 to 9% in 3Q15 to 21% in 4Q15), with much broader products and applications ramping in 2016.

TSMC Says Business Has Slowed In The Last Month At Credit Suisse Conference - Asia Stocks to Watch - Barrons.com
Click to expand...

[h=4] Strong bias against TSM [/h]
It appears that, perhaps due to the consistent PR bombardment from Intel and Samsung, TSMC is generally considered as intrinsically inferior. The rumors of Samsung winning most of A9 orders or major customers abandoning TSMC, are promptly taken as facts without critical reviews, because such rumors fit right into the popular perception.

Didn’t TSMC just deliver record profits in 4Q14? The media tells us that it is purely luck. The company had peaked, all downhill from here.

1|16|2015
after the world’s largest contract chipmaker announced a record quarterly profit for the end of 2014 that may mark the start of a downturn for the company during 2015.

Analysts Turn Negative on Outlook for TSMC | EE Times
Click to expand...

Any little negatives cause strong sell rating. Maybe, the company cannot achieve 15% growth this year, but will still be double digits, not to mention 50% raise of dividends.

For comparison, in 2H14, Samsung saw revenue down 20%, profits cut over one third. I am not aware that Samsung was downgraded to strong sell. Intel, with lower growth rate, was not downgraded to strong sell, either, after its recent warning.

“TSMC: The Writing Is On The Wall – Sell.” This kind of pounding-table sell is particularly reserved for TSM. I don’t remember any other company ever received such extremely bearish call, not even the bankrupting and fraudulent Enron.

TSM investors need to understand this deep-rooted unfavorable bias and the periodic corrections. TSM also endures higher volatility.

Huge bearish bets on TSM

More than merely bias. Huge bearish bets had been placed on TSM. See reports below:

March 26, 2015
What's behind Taiwan Semi puts
What's behind Taiwan Semi puts - Yahoo Finance
Click to expand...

Friday, 20 Mar 2015
Trader bets $1 million against this Apple supplier
http://www.cnbc.com/id/102522677?__source=yahoo|finance|headline|headline|story&par=yahoo&doc=102522677
Click to expand...

[h=4] Panic sell-off and buying opportunity? [/h]
On the single day 5.3% plunge of 3/25, the TSM trading volume was over twice of average. It appears that sell-stops were hit and then cascaded downward into a panic sell-off.

The next day, similar high volume. The third day, volume decreased but still higher than the 3-month average.

TSM - prices and volumes
http://finance.yahoo.com/q/hp?s=TSM+Historical+Prices

Had the market already shaken off weak holders of TSM and near a bottom? Is it a buying opportunity here?

I think so. TSM can still go lower short-term. But, not a bad idea to start accumulating.

Perhaps 2Q will be flat sequentially, but sill double-digit growth yoy. From 3Q to the end of 2016, TSMC revenue should see persistent up ticks. Personally, I think 15% annual revenue growth is no problem, with the possibility of hitting 20%.

Readers need to do their own due diligence and decide for themselves. See disclaimer below.

[h=4]Disclaimer [/h]
The posters are not licensed financial advisers. The forum posts are an expression of personal opinions, and does not constitute recommendation of investments whatsoever.
 
Last edited:
Another piece of evidence that points to last week’s panic selling of TSMC shares. See news below.

March 30, 2015

From March 23 to 27, foreign institutional investors sold a net NT$33.25 billion (US$1.06 billion) worth of local shares on the main board, with the selling largely focused on TSMC, the most heavily weighted stock on the market.

TSMC records largest foreign institutional net sell last week - Taiwan News Online
Click to expand...

Maybe weak holders were flush out during the sudden plunge, now suffer from shell shock, and are afraid to buy TSM again?
 
I've being noticing a pattern over the last year. Around the time of TSMC's earnings call a barrage of negative rumours start doing the rounds - there's already one out today - and as April 16th looms closer we will see more. Then, afterwards it'll go quite again. The timing of these rumours is no coincidence.
 
lefty said:
I've being noticing a pattern over the last year. Around the time of TSMC's earnings call a barrage of negative rumours start doing the rounds - there's already one out today - and as April 16th looms closer we will see more. Then, afterwards it'll go quite again. The timing of these rumours is no coincidence.
Click to expand...

I think, the attacks on TSM from the combined forces of media and Wall Street seem more frequent, not limiting to the timeframe leading to earning calls.

For example, the assault, which led to the low of March 10, happened right in the middle of Jan and April earning calls.

In a way, TSMC should pride itself on these attacks, because its competitors felt threatened enough to resort to dirty tricks. I don’t remember Intel PR ever bothered with AMD, nor did Samsung ever targeted HTC or GlobalFoundries.

The best they can do is to knock 10% off TSM from all time high’s, only to create buying opportunities, without altering anything fundamentally.
 
lefty said:
That's exactly what they want! 10% is enough to make some one a nice profit.
Click to expand...

Indeed, those large put buyers made huge and quick profits, if they closed out the options near the low of last Friday.

The losers are those who sold TSM shares in panic, especially near the low’s.
 
[h=4]Another example of media bias against TSM [/h]

The potential lower revenue of 20nm is painted as a big negative.

First of all, the lower 20nm revenue is offset by the higher and earlier 16nm revenue from the same TSMC revelation at Credit Suisse Conference.

Secondly, lower 20nm income may not happen at all. It’s true that the A8 production is to decline and output of SD810 will be lower (due to losing Galaxy S6 sockets). However, MediaTek’s 20nm SoCs and iPhone 6C powered by A8 are expected to launch in the second half. It’s unclear if 20nm revenue will actually drop below the earlier estimates.

The reporters decided to ignore the new 20nm chips and better 16nm incomes, in order to present a bearish picture.

Many instances of such bearish bias. I don’t have time to write them all. Readers need to do their own researches.
 
You must log in or register to reply here.
Back
Top