TSMC (TSM) Q4 2014 Results Dicscussion

[user_2013101]

Not all the CapEx is 16nm.

16 nm buildup

I didn’t say the capex is all for 16nm. But, it is MOSTLY for 16nm.

Lora Ho - SVP and CFO

Randy, in my remark I was talking about 80% of the CapEx goes to leading edge technology. That actually cover very big part of 16 FinFET capacity and also the 10 nanometer for the engineering line and R&D expenditure that altogether leading edge technology will be 80%.

80% is $9.6B. 10nm is merely in R&D and risk production. $8B, out of 9.6, for 16nm is a reasonable guess. Throwing $8B at 16nm buildup, when 80-100% of A9 production had gone to Samsung-GF or 16nm delayed to 2016?

1Q15 revenue 43% above 1Q14

If the low end of 1Q guidance comes true, the revenue will grow 43% yoy.

If 2Q through 4Q grow only half of 1Q yoy, 2015 will grow over 20%. The 16% forecast is likely to be a lowball number.

Summary

Nevertheless, nothing can stop negative rumors. Anyone can create rumor pages, sometimes citing unnamed “industrial sources.” I include one such amusing rumor at end.

Most analysts are not thinking straight, either. But, that’s nothing new. In the past 10 years, majority of time, TSM ratings have been neutral or sell. Some analysts had concluded that TSMC is out. Only Intel and Samsung are left standing to compete at cutting edge technologies.

For entertainment

This news came out after the TSMC conference call:

TSMC 16nm FinFET Faces Major Delay - Qualcomm Jumps Ship To Samsung

 

[lefty]

16 nm buildup

I didn’t say the capex is all for 16nm. But, it is MOSTLY for 16nm.

Lora Ho - SVP and CFO

Randy, in my remark I was talking about 80% of the CapEx goes to leading edge technology. That actually cover very big part of 16 FinFET capacity and also the 10 nanometer for the engineering line and R&D expenditure that altogether leading edge technology will be 80%.

80% is $9.6B. 10nm is merely in R&D and risk production. $8B, out of 9.6, for 16nm is a reasonable guess. Throwing $8B at 16nm buildup, when 80-100% of A9 production had gone to Samsung-GF or 16nm delayed to 2016?

1Q15 revenue 43% above 1Q14

If the low end of 1Q guidance comes true, the revenue will grow 43% yoy.

If 2Q through 4Q grow only half of 1Q yoy, 2015 will grow over 20%. The 16% forecast is likely to be a lowball number.

Summary

Nevertheless, nothing can stop negative rumors. Anyone can create rumor pages, sometimes citing unnamed “industrial sources.” I include one such amusing rumor at end.

Most analysts are not thinking straight, either. But, that’s nothing new. In the past 10 years, majority of time, TSM ratings have been neutral or sell. Some analysts had concluded that TSMC is out. Only Intel and Samsung are left standing to compete at cutting edge technologies.

For entertainment

This news came out after the TSMC conference call:

TSMC 16nm FinFET Faces Major Delay - Qualcomm Jumps Ship To Samsung

The guidence is for several percentage points above the average foundry revenue growth of 12%, in other words, about 17% -20%. Definately, not 43%. Most of that growth will be 20nm, because it's going from 9% in 2014 to 20% in 2015.

Edit: By the way, the article on WccfTech is just a repost of the rumour that was posted on DigiTimes 1 day *before* earnings call. Am I the only one that finds it suspicious of the timing of these negative rumours?

 

[user_2013101]

The guidence is for several percentage points above the average foundry revenue growth of 12%, in other words, about 17% -20%. Definately, not 43%.

I didn’t suggest 43% growth for the entire year; that’s for 1Q only.

I went back to check the transcript, TSMC first said several percentage points over the 12% foundry rate. One analyst, Andrew Lu from Barclays, asked if 15% revenue growth for 2015 is reasonable estimate, and the answer was positive.

My personal opinion is that the growth is likely to exceed 15%. 20% or higher is more realistic.

 

[Fred P]

In the interest of full disclosure I have a substantial long term holding in WX and TSM of over 10K shares in each. I have read over 20k pages on each.

Id be very careful with WX.

 

[lefty]

There was a interesting question in the call that I just noticed now:

Andrew Lu - Barclays

I remember last Investor Conference C.C Wei mentioned 16 FinFET revenue have a high single digit by Q4 this year and may be few percentages by Q3. Is that number unchanged?
C. C. Wei - Co-CEO

Unchanged.
Andrew Lu - Barclays

You sound so less [ph] confident.
C. C. Wei - Co-CEO

The more I say the more that information from the customer will be released soon.

In other words he was afraid to elaborate on this in case any customer sensitive information would be released. That seems to indicate that maybe the previous figures have indeed changed.

 

[user_2013101]

There was a interesting question in the call that I just noticed now:

In other words he was afraid to elaborate on this in case any customer sensitive information would be released. That seems to indicate that maybe the previous figures have indeed changed.

The contradiction

The estimate of 7-9% 16nm sales in 4Q15, as well as the smaller 14/16nm market shares this year, contradict the guidances of revenue and capex, as explained in my earlier posts.

IMO, for unknown reasons, TSMC does not want to give out any evidence that it had won most of A9 production. Had it estimated the high teens percentage of 16nm 4Q revenue (which I believe will be the case), the world would know that it had most, if not all, of A9 orders.

Uniformly bearish on TSMC

The following are examples of prevailing media sentiments on TSMC, AFTER the Jan 15 conference call.

“Analysts Turn Negative on Outlook for TSMC”

"TSMC is expected to lose out to Samsung in high-stakes battle to make Apple’s next smartphone."

"TSMC officials conceded that the company would lose market share in 2015, while Samsung executives have sounded increasingly confident about winning back Apple’s business"

In fact, all brokerage houses have either neutral or sell ratings on TSM. If anyone knows of buy or strong-buy rating, please post.

Investment basics

It’s a risky investment in a stock with universal bullish sentiment, because most likely all the good news in anticipation have been priced in already and any slight disappointment may cause outsized corrections. Little upside, but a lot of downside risks.

Today’s TSM is in the exact opposite situation: universal bearish sentiments, with high probability of huge upward surprises in revenue, profits, A9 productions, etc. In the longer term, 2-3 years from now, not only that TSMC is likely to deliver 10nm chips ahead of Intel and Samsung, but that its 10nm chips will be superior.

Ever heard “climb a wall of worries”?

However, with such bearish sentiment, in addition to a somewhat unstable overall stock market, some sort of corrections on TSM may be unavoidable in the coming months. IMO, it’s a good idea to accumulate TSM on weakness. TSM is a must-own in the semi sector.

Disclaimer

The posters are not licensed financial advisers. The forum posts are an expression of personal opinions, and does not constitute recommendation of investments whatsoever.

 

[hist78]

Because user_2013101's comment, I decided to read the official TSMC conference call transcript. I found out there are errors in the Seeking Alpha’s version of transcript.

Official transcript:
Taiwan Semiconductor Manufacturing Company Limited#

One of interesting Q&A topic is TSMC 2015 revenue and growth. Here is a segment from the transcript:

####

Andrew Lu - Barclays Capital - Analyst
Dr. Chang, Dr. Liu and Dr. C.C. Wei and CFO. (Spoken in foreign language). The first one is regarding the revenue outlook for the next few quarters. Are you expecting any single quarter, for the next few quarters, Q2, Q3, Q4, revenue below first quarter?

Elizabeth Sun - Taiwan Semiconductor Manufacturing Company Ltd - Director of Corporate Communications
Andrew is asking us to give him a guidance whether or not our Q2, Q3, Q4 revenue will be lower than Q1 level.

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
Will Q2, Q3, Q4 be lower than Q1?

Andrew Lu - Barclays Capital - Analyst
Any single quarter in your internal forecast saying will be lower than the first quarter?

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
What is Q1 times four?

Andrew Lu - Barclays Capital - Analyst
Not what the --.

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
I will work out the answer here, right here.

Andrew Lu - Barclays Capital - Analyst
Because that's the following what I am going to calculate, because, based on the estimates, Q1 is quite similar to Q4 from the revenue, from the EPS point of view, from OP margin guidance, gross margin guidance. If we times four, revenue is up 16% year over year and the EPS up 20% year over year.

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
That would be consistent with what Mark said. He said that we will outperform the foundry growth, which is, what, 12%? He said we would outperform it by several points.

Andrew Lu - Barclays Capital - Analyst
Yes. But this is based on flattish environment. We've got no growth for the next few quarters. That's why I ask for any downside result.

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
All right. Let me just tell you what I think. I think we have upside. Okay?

Andrew Lu - Barclays Capital - Analyst
Okay. That's now?

Morris Chang - Taiwan Semiconductor Manufacturing Company Ltd - Chairman
In this year. Yes.

Elizabeth Sun - Taiwan Semiconductor Manufacturing Company Ltd - Director of Corporate Communications
And Andrew, I also have to remind you, the foundry numbers are based in US dollars, but the fourth-quarter or the first-quarter revenues are based in NT dollars.

####

Dr. Morris Chang gave his 2015 revenue hint (by my own feeling) saying "What is Q1 times four?" I did a calculation and that "Q1 times four" represents TSMC 2015 revenue can be almost US$29 billion, close to 15% increase from 2014 revenue. With all the negative rumors floating on the market, this number seems not real at all. But if you trust Dr. Chang and TSMC's credibility, this is a very positive sign.

BTW, Seeking Alpha transcribed Dr. Chang's "What is Q1 times four?" comment to "What is the Q1 times for?". One missing character 'u' can mean a huge difference!

 

[lefty]

The contradiction

The estimate of 7-9% 16nm sales in 4Q15, as well as the smaller 14/16nm market shares this year, contradict the guidances of revenue and capex, as explained in my earlier posts.

IMO, for unknown reasons, TSMC does not want to give out any evidence that it had won most of A9 production. Had it estimated the high teens percentage of 16nm 4Q revenue (which I believe will be the case), the world would know that it had most, if not all, of A9 orders.

Uniformly bearish on TSMC

The following are examples of prevailing media sentiments on TSMC, AFTER the Jan 15 conference call.

“Analysts Turn Negative on Outlook for TSMC”

"TSMC is expected to lose out to Samsung in high-stakes battle to make Apple’s next smartphone."

"TSMC officials conceded that the company would lose market share in 2015, while Samsung executives have sounded increasingly confident about winning back Apple’s business"

In fact, all brokerage houses have either neutral or sell ratings on TSM. If anyone knows of buy or strong-buy rating, please post.

Investment basics

It’s a risky investment in a stock with universal bullish sentiment, because most likely all the good news in anticipation have been priced in already and any slight disappointment may cause outsized corrections. Little upside, but a lot of downside risks.

Today’s TSM is in the exact opposite situation: universal bearish sentiments, with high probability of huge upward surprises in revenue, profits, A9 productions, etc. In the longer term, 2-3 years from now, not only that TSMC is likely to deliver 10nm chips ahead of Intel and Samsung, but that its 10nm chips will be superior.

Ever heard “climb a wall of worries”?

However, with such bearish sentiment, in addition to a somewhat unstable overall stock market, some sort of corrections on TSM may be unavoidable in the coming months. IMO, it’s a good idea to accumulate TSM on weakness. TSM is a must-own in the semi sector.

Disclaimer

The posters are not licensed financial advisers. The forum posts are an expression of personal opinions, and does not constitute recommendation of investments whatsoever.

TSMC have said that Samsung have bigger market share in FinFET for 2015. They repeated that statement many times, so it's pretty unrefutable. Now, if TSMC get the bulk of the A9 contract, as you claim, that means that some other TSMC customers have jumped ship to Samsung's 14nm.

 

[user_2013101]

TSMC have said that Samsung have bigger market share in FinFET for 2015. They repeated that statement many times, so it's pretty unrefutable.

Lefty,

Suggestion: Please just quote the first sentence or paragraph of a long post you reply to, so that reader don’t need to scroll all the way down to read your comment. Thanks. (You can manually delete the rest of quoted post)

Despite my repeated statements that the talk of smaller 14/16nm share is in conflict to the forecast of revenue and capex, you focus solely on this point and maintain a bearish view, similar to the analysts under the aggressive Samsung propaganda.

I don’t have problem with your view; that is exactly why TSM is a buy to me - universal bearish on a great company that will explode upward this year.

Even Nenni, who usually has favorable view on TSMC, also believe most of A9 orders had gone to Samsung. This is telling. Now TSMC producing only 50% of A9s will beat the expectations. In the next 6-9 months, when it becomes clear that TSMC produces 70-100% of A9s, on top of other gains like IoT, TSM will skyrocket. We have seen TSM gap-up 8% in a day.

Do you feel strange that the leading foundry repeatedly emphasizes it is losing market shares to the competitor in the race of cutting edge 14/16nm? No other companies would do so. Usually, companies will spin such bad news; or at least express it in a euphoric way. Why does TSMC express it this way?

P.S.
Most fab-less companies do use second production sources. But, very little new shifts to Samsung. It’s just Samsung’s PR.

 

[lefty]

Well, I would say my view is realist, rather than bearish. But I never said TSMC will do bad. 15-20% revenue improvement means TSMC is a buy, whatever Samsung is doing, or not.
If Mr Chang repeatedly says competitor gets bigger market share, then yes, I believe him.

 

[Daniel Nenni]

Well, I would say my view is realist, rather than bearish. But I never said TSMC will do bad. 15-20% revenue improvement means TSMC is a buy, whatever Samsung is doing, or not. If Mr Chang repeatedly says competitor gets bigger market share, then yes, I believe him.

In my experience Dr. Morris Chang uses the conference calls to motivate TSMC employees rather than pander to Wall Street. I have heard him many a time sound the competition alarm bell only to see TSM stock drop as a result. The comment about losing share at 16nm is the latest example. The result being 16FF+ is a much better node than 16nm and will arrive before originally expected. I also believe TSMC North America is being much more aggressive on the business side and will in fact recover much of the lost market share in 2016.

You would do better to ask your dog to pick a stock than me but I did tell my Wall Street followers that TSMC would own 28nm and 20nm and they continue to thank me for that. For 16/14nm I see a market split between Samsung and TSMC but feel TSMC will continue to post strong numbers based on continued 28nm and 20nm business. My prediction is that TSMC will transition quickly to 10nm and again hold a dominant position. The last five years have been exciting for the fabless semiconductor ecosystem and I feel the next 5 years will be more so. SoCs will continue to flourish, new applications are coming, and with IoT we are seeing just the tip of the iceberg. Just my opinion of course but it is a great time to be fabless.

 

[Arthur Hanson]

But why did TSMC sell their ASML shares? Anyone know?

Cash to work with, their expansion takes cash.

 

[ippisl]

Arthur , that's very interesting.Where can i find info about wuxi and mems ?

And apple's breakthrough device is the watch i presume, and nvidia's are their new tegra chips (which could become very popular) i presume ?

 

[user_2013101]

[h=4]The reality that is buried under hostile PR[/h]

The fast ramp of TSMC’s 20nm node probably breaks the semi industry record. But, other than perhaps Nenni mentioned it once, no media coverage on this fact. Instead, media and bloggers focus on the “irreparable low-yield” of the 20nm that only Apple can afford to eat such poor yield.

They probably will invent new 20nm problems when more 20nm chips, from QCOM, NVDA, MediaTek, etc., hit the market later this year.

The superior density 20nm has achieved is completely obscured by the Intel’s leads in “logical density” that cannot be found in physicals.

Now media is spreading major 16nm delays, lost of A9 orders, etc. Pretty much, It’s over for TSMC in the race of cutting edge processes.

[h=4]The irony[/h]
Comparing to producing 100% of A8s, the case of TSMC producing only 50% of A9s should be considered a major setback.

Not any more!

Thanks to the negative PR from Intel and Samsung, now 50% of A9s beats expectation. 70-100% will be explosive.

It’s funny how rivals’ negative campaigns have helped TSMC, who seems to be passively playing along with the negatives.

 

[lefty]

I noticed quite a shocking statement from an EE Times article:

Qualcomm is said to have exclusive rights to the new 16 nm TSMC process for perhaps six months or so.

7 Takeaways From ISS 2015 | EE Times

Could that be true?

 

[user_2013101]

I noticed quite a shocking statement from an EE Times article:

7 Takeaways From ISS 2015 | EE Times

It is emphasized in boldface under the title of the article: “Samsung, GF winning 14/16 nm race.” No, I didn’t make it up: indeed, the media had created a virtual reality that TSMC had lost the race.

So, why QCOM needs exclusivity on the inferior 16nm that had lost the race? Didn’t QCOM already dump TSMC and move on to Samsung’s 14nm?

Sometimes, reading these stories is more amusing than watching Saturday Night Live.

 

[user_2013101]

The latest from the unnamed “industry sources” in Korea:

"According to industry sources on Jan. 19, Apple has recently decided to request Samsung to manufacture 70 percent of A9 APs, core components for the iPhone 6S and iPhone 7."

Jan 20, 2015
Romance of Necessity: Apple Decides to Purchase 70% of Mobile AP A9 from Samsung | BusinessKorea

But, Samsung had started A9 production already back in early December, according to “industry sources” in Korea.

Dec 12, 2014
Samsung Electronics began production of Apple A9 in 14nm FinFET
???? IT??? ??! ????

 

[hist78]

It is emphasized in boldface under the title of the article: “Samsung, GF winning 14/16 nm race.” No, I didn’t make it up: indeed, the media had created a virtual reality that TSMC had lost the race.

So, why QCOM needs exclusivity on the inferior 16nm that had lost the race? Didn’t QCOM already dump TSMC and move on to Samsung’s 14nm?

Sometimes, reading these stories is more amusing than watching Saturday Night Live.

If this "exclusive rights" is true, it will be interesting to see what exactly the Qualcomm/TSMC 16nm product will be. TSMC won't accept and Qualcomm won't ask the 16nm "exclusive rights" just because they happen to see each other at Taipei airport on a sunny day.

 

[Brianhayes]

For me the most significant news is the hike in Capex. Given that TSMC has 53% of the market as against less than 20% for Intel with Capex that has never been greater than Intel what will happen now that, as one analyst question put it, they are now top of the tree. As Morris Chang has said again and again all Capex is spent to customer commitment.

Another interesting psssage sage was when Mark Liu was directed by the Chairman to talk about going into servers. He said they were in "partnership" with Arm an "innovative" company that produces new products every year. The last product launched by Arm was the A12, hardly relevant to servers. Was he talking about Maya and Artemis?

These guys don't say much but what they do say is worth a great deal more than Intel's spin machine and the acolytes it inspires. I am not a semiconductor guy but a "finance guy" as DAN puts it - at least I was - but 53% of the market and outgrowing the supposed market leader by a large margin ... I call that dominance.

 

[user_2013101]

Posted under another thread:

Latest of TSMC 10nm progress