More FinFET delays

[Daniel Nenni]

Not a religious war, I'm not biased towards or against any foundry (and have been involved in process/foundry selection) -- but just a couple of points need making...

"10 customer 16nm tapeouts in 2014 so far" -- so maybe 5 customers have actually got 16FF silicon in their hands right now (see below)...
"TSMC is already in production with a 16nm FinFET network processor for HiSilicon" -- they've just delivered first samples, this is not "in production"

"Most of these people already have 16nm silicon" -- if 5=most (see above) it must be a pretty small event then, presumably only 7 or 8 companies attending?
"no fooling them with ambiguous slides and double speak" -- I know at least one of said companies is unhappy with 16FF not delivering claimed power savings -- so were they fooled?

From the TSMC OIP website: "With this forum, TSMC offers you an effective marketing venue..."

In other words, it's a marketing event for TSMC to promote their technology, and ecosystem (IP/software) partners to market their products to current and potential customers. Absolutely nothing wrong with that -- TSMC is a fantastic foundry with massive support -- but don't expect to see any shadow of doubt about how wonderful the technology and the ecosystem are and that all the problems are solved. I'm equally sure that if you went here:

GLOBALFOUNDRIES to Host Inaugural Global Technology Conference

you'd see exactly the same kind of presentations but with TSMC replaced by Globalfoundries, and everything being similarly wonderful. Nobody should be under any illusion that marketing events like these are in any way impartial, just put on the free rose-tinted glasses when you walk through the door.

As you say, your *opinion*... ;-)

Clearly your problem is with me personally, you made that evident by your first comments and this post is a byproduct of that.

For the record, test chips qualify as 16nm silicon in my definition and MANY people including EDA, IP, and early access fabless companies have 16nm Silicon. According to my notes more than 40 EDA tools and 700 IP have been qualified for 16nm. Do you have any idea how many test chips that required? You should also look at the agenda for OIP as it has 16nm silicon written all over it:

TSMC 2014 - Attendee Registration

I attend dozens of these conferences every year (including ISSCC and IEDM), I even help organize conference sessions. Last year I did a full day on FinFETs at EDPS and next year I'm doing a session on FinFET versus FD-SOI design challenges. I work inside the fabless semiconductor ecosystem by day and write about it extensively at night. I'm one of the last people that TSMC or anyone else for that matter should hand a pair of rose colored glasses to.

 

[infras]

Was IanD banned for his comments?

 

[Daniel Nenni]

"Samsung on the other hand licensed their 14nm FinFET process from IBM. Samsung also got the 28nm process from IBM which did not yield properly"

No, Samsung did not license FinFET from IBM, and BTW their 28nm is what powers iphone 5s.

All companies have yiled problems once in a while and that's the nature of this industry. Suppose TSMC has problem ramping 16FF in time for next Apple prodcut. So what? They might as well ship the next product on 20nm (no one waits for an uncertain yield ramp to tape-out a product anyway).

Marketing always tend to show things rosy. But looks like we are entring a new era with the introduction of FinFET. Honestly, I'd feel much better if foundries would call their first FinFET technology a 20nm FF. Same thing that they did at 28nm. Each company has several versions and yet they are all called 28nm, because they share the same base GR. That was the first place things went wrong. Then came the performance claims. Virtually everyone in the industry made a claim of 30-50% power saving. If that's the case why there is a need to 16FF+? I was looking at some of the charts a top tier fabless shared with me two years ago when they were evaluating 16FF and to my surprise their post layout was faster than pre-layout! The third thing that went wrong, IMO, was several fabless companies insisting in skipping 20nm and moving directly from 28nm to FinFET. I know they spent time and resources just to realize it's better to ship 20nm products. And now is the 4th thing; tying the industries success to the success of FinFET. If TSMC says they share 90% of the 20nm equipment with FinFET (which should be the case for most processes and for all foundries) why Ultratech is claiming their shipment is delayed beacuse of FinFET ramp delay?

TSMC's strength is the ability to serve a wide range of customers across many nodes and I admire them for that. I don't care if they are first to ramp 16nm, 10nm, or whatever. That's a game a certain company likes to play; As I commented somewhere else, to see where the industry stands you should not go to IEDM, ISSCC, or companies sponsored conferences. Go to Apple store, Best buy, or whatever your favorit store is.

Okay, four comments:

(1) I was told by GF during our briefing on Monday that Samsung Licensed 14nm from IBM. I believe it was an architecture license so maybe process was the wrong word. I will double check. Here is my write-up of the briefing in case you are interested:

https://www.semiwiki.com/forum/content/3972-gf-ibm-deal-explained.html

This is also an interesting take on the IBM acquisition from a former TI guy:

https://www.semiwiki.com/forum/content/3974-ibm-leaves-semiconductors-end-era.html


(2) For what it is worth, here is my take on Samsung 28nm:

SemiWiki - Can Samsung Deliver As Promised?

https://www.semiwiki.com/forum/content/2180-samsung-28nm-still-does-not-yield.html


(3) I agree with you on the process naming. Intel started the obfusication @ 22nm "TriGate" and it went down hill from there. I'm not sure if there is a solution to this since marketing people now seem to control the naming. It really is confusing to outsiders but that drives more traffic to sites like SemiWiki so there you go.


(4) The Ultratech CEO said "Once again a major logic manufacturer delayed their FinFET ramp" and everyone started pointing fingers at TSMC. It could be Samung, GF, UMC, or maybe even SMIC, right? And the media is having a field day. Check out this interpretation of Art’s comments from the Motley Fool “Senior” Technology Specialist:

However, after listening to the earnings call of chip equipment vendor Ultratech (NASDAQ: UTEK ) , it's clear to me that neither TSMC nor Samsung quite has the FinFET transistor structure (which promises higher performing transistors at lower power) figured out. This, as far as I can tell, strongly suggests that Intel's manufacturing lead remains intact.

Absolutely ridiculous. He owns Intel stock of course.


Bottom line: The "major logic manufacturer that pushed out an equipment order" is not TSMC, I'm 99.99% sure of that. If I had to pick one from the other possibilities it would be UMC. They licensed IBM 14nm and I have not heard of any production equipment moving in yet.

 

[astilo]

Bottom line: The "major logic manufacturer that pushed out an equipment order" is not TSMC, I'm 99.99% sure of that. If I had to pick one from the other possibilities it would be UMC. They licensed IBM 14nm and I have not heard of any production equipment moving in yet.

Well, it would be really hard to believe that either TSMC or Samsung/GF are currently manufacturing 14nm wafers (test chips, prototypes, qual and pre prod) without having the equipments in house. Easy bet I guess.

 

[IanD]

Was IanD banned for his comments?

It would seem so, and I've had some very illuminating comments on the fact...

 

[IanD]

Assuming it's not Intel might also be wrong: the report said:

"As we have discussed on past conference calls, the difficult implementation of 3D FinFET microprocessors to high production manufacturing. Once again a major logic manufacturer delayed their FinFET ramp. We had then requested to prepare LSA tools for shipment for the end of the third quarter which was delayed. These LSA shipments for the most part caused our third quarter revenue to be less than projected. These LSA systems have been rescheduled for shipment in the fourth quarter."

It's public that Intel have delayed their 14nm rollout/shipments, given the mention of "3D FinFET microprocessors" it could equally well be them.

 

[hist78]

I have a question and it's something not totally related to this discussion. WTH, who is that Ultratech's CEO? How can he reveal some sensitive information without his customer's approval? Even he didn't say which company exactly but this foundry industry has a small and limited number of players. People WILL find out who is that "unnamed" comany sooner or later.

How can he use his customer's private information as an execuse for his own revenue problem?

"Trust" is the critical element and requirement in this industry!

 

[IanD]

I have a question and it's something not totally related to this discussion. WTH, who is that Ultratech's CEO? How can he reveal some sensitive information without his customer's approval? Even he didn't say which company exactly but this foundry industry has a small and limited number of players. People WILL find out who is that "unnamed" comany sooner or later.

How can he use his customer's private information as an execuse for his own revenue problem?

Because he thinks that if he doesn't say who the customer is he hasn't given anything away. I wonder if his customer agrees?

 

[hist78]

I doubt if any foundry will agree to let a supplier to reveal foundry's production delay or problem.

That Ultratech's CEO is basically telling the world that he or his company can't be trusted.

 

[Daniel Nenni]

It would seem so, and I've had some very illuminating comments on the fact...

Clearly you are not banned but we both were warned that cooler heads should prevail and rightly so. My appologies to all.

 

[Brianhayes]

I doubt if any foundry will agree to let a supplier to reveal foundry's production delay or problem.

That Ultratech's CEO is basically telling the world that he or his company can't be trusted.

This is my first post so I will start with my credentials. I have zero knowledge of semiconductors but I am an investor in Arm, TSMC and Apple.
i would suggest everyone reads all that Zafiropoulos said before deciding that his statements have credibility and, in particular, his response to the question posed by Ted Moreau. In my opinion if C C Wei were to read this he would not buy Z's products unless he absolutely had no alternative. Z flatly contradicts what C C Wei said in the TSMC transcript and one should remember that such statements , if made negligently or falsely, are I believe subject to criminal penalties. In other respects Z's remarks with respect to 16nm in general and TSMC in particular are riddled with inconsistencies. In particular he admits at one point that he does not have a strong position in 16nm! It could be that C C Wei has no opinion of his product.
Just my opinion of course!

 

[SheDevil]

From the TSMC OIP website: "With this forum, TSMC offers you an effective marketing venue..."

In other words, it's a marketing event for TSMC to promote their technology, and ecosystem (IP/software) partners to market their products to current and potential customers. Absolutely nothing wrong with that -- TSMC is a fantastic foundry with massive support -- but don't expect to see any shadow of doubt about how wonderful the technology and the ecosystem are and that all the problems are solved.

You are quoting from the page that they use to get partners to rent booth space? Every conference with exhibits has that page. Exhibits help cover expenses, otherwise attendance would be a fraction of what it is. I'm guessing you did not attend OIP but please look at the abstracts before commenting on the value of the event. By the way it is not easy getting a paper into a TSMC conference. The review process is daunting!

 

[hist78]

Clearly you are not banned but we both were warned that cooler heads should prevail and rightly so. My appologies to all.

 

[Li Yisuo]

 

[IanD]

You are quoting from the page that they use to get partners to rent booth space? Every conference with exhibits has that page. Exhibits help cover expenses, otherwise attendance would be a fraction of what it is. I'm guessing you did not attend OIP but please look at the abstracts before commenting on the value of the event. By the way it is not easy getting a paper into a TSMC conference. The review process is daunting!

What charming faith you have ;-)

Nobody would deny that a lot of useful information is presented at industry-sponsored events like the TSMC and GF ones. But equally nobody should be under the illusion that they're impartial, as shown by the huge public disagreements -- for example, on wafer cost or gate density or cost -- between the CEO presentations of Intel, TSMC and GF at their respective events, they can't all be right. The purpose of such presentations is to talk up your own technology and ecosystem (and sometimes to talk down the competition), tell everybody that you're on track and doing a great job, and convince new customers to choose you instead of the competition.

All perfectly justifiable and good business practice, but don't expect anyone to ever admit that the competition might have any advantages, and owning up to any problems or delays would just give the press and your competitors a big stick to beat you with, so this doesn't happen unless the train crash is out in public and undeniable -- like GF canning 14XM in favour of licensing Samsung's 14FF process less than a year after they'd been telling the world how wonderful it was and how it would be first to the market...

So to get (hopefully) relatively unbiased comparisons of technologies people read sites like this, where some people occasionally publish really helpful comparisons with most of the PR fluff stripped away. The difficulty is often still spotting whether what they're saying has a strong slant in favour of one company or another -- not as strong as the CEO presentation mentioned above, but still very definitely not using a level playing field.

And very often there isn't a simple answer, which is "best" depends on the application -- for example, people flat-out saying "FinFET is the best solution for everything" or "FDSOI is better and cheaper" are ignoring reality, which says that they are different technologies, both have advantages and disadvantages and are best suited to different applications -- and making biased statements like this just makes them less likely to be taken seriously.

 

[otistd]

Do you really feel that SemiAccurate is a good source for this type of information?

Personally, I have found that SemiWiki and SemiAccurate are both, well, semi-accurate as far as the information they provide. They are all opinion pieces- and written with definite points of view- and they gloss over a lot of unattractive facts that subtract from or even sometimes contradict their viewpoint. Like the price tag for GF taking IBM semi- is well south of the $2+Billion asserted in your previous piece- but that is glossed over (not by PM though). On the SA side, predicting a GF-IBM Semi merger has been going on for years now- everyone saw it coming, so predicting it months ago is really not so impressive. I'm not paying SA for the full article, so I really don't have much comment about the accuracy of this particular piece- although you and a lot of other people seem to think the yield issue is real- and it's either at TSMC or Samsung/GF or both. My bet is both- at least once they are both ramping seriously- FinFETs with HKMG are just not easy to do and bound to cause problems. Intel acknowledged their 14nm yield problems (mostly because they missed their timing probably), but nobody else is acknowledging anything- let alone explaining it- so where are the calls for openness like what you wanted from Intel previously- I'll assume that's coming along with the outing of who is having yield problems. The fact is, you are very clearly biased towards pure play foundries- and towards TSMC in particular- anyone who has followed your posts knows that- IanD notes it below in the comments too. I've come to accept it, and still value everything you put up here because it gives very good insight into the fabless ecosystem, but it is a bit grating at times, so I'll admit to chuckling over the SA "paid some Ninny" line- that was a good one- and goes to what many people suspect, which is you have a strong financial conflict of interest when it comes to TSMC.

By the way, I think Chenming Hu was at Berkeley, not TSMC, when he pioneered FinFETs- otherwise TSMC would have the patent, he elected not to patent. However, TSMC has made some important contributions to FinFETs, like showing how to strain the channel, and Intel's K. Kuhn has acknowledged that publicly.

 

[CharlieD]

I have a question and it's something not totally related to this discussion. WTH, who is that Ultratech's CEO? How can he reveal some sensitive information without his customer's approval? Even he didn't say which company exactly but this foundry industry has a small and limited number of players. People WILL find out who is that "unnamed" comany sooner or later.

How can he use his customer's private information as an execuse for his own revenue problem?

"Trust" is the critical element and requirement in this industry!

The short answer is he cannot. All of the foundry NDA's that I have seen expressly prohibit the mention of yield. Yield differs between designs so this is completely understandable for a foundry. For IDMs yield numbers are more relevant but I doubt Intel would allow a vendor to disclose numbers.

 

[CharlieD]

You are quoting from the page that they use to get partners to rent booth space? Every conference with exhibits has that page. Exhibits help cover expenses, otherwise attendance would be a fraction of what it is. I'm guessing you did not attend OIP but please look at the abstracts before commenting on the value of the event. By the way it is not easy getting a paper into a TSMC conference. The review process is daunting!

I reside in Silicon Valley and attend many of these events with coworkers. We are also early access foundry customers so we know the wheat versus chaff. The TSMC events this year have been excellent and correlate with what we already know. Sure there is showmanship but that is to be expected. It is also an opportunity to see old friends and compare notes. I highly doubt IanD attends them so I would not give credence to what he wrote. I did not attend the Globalfoundry event this month but a coworker did and felt it was very marketing heavy. My $.002.

 

[SheDevil]

I reside in Silicon Valley and attend many of these events with coworkers. We are also early access foundry customers so we know the wheat versus chaff. The TSMC events this year have been excellent and correlate with what we already know. Sure there is showmanship but that is to be expected. It is also an opportunity to see old friends and compare notes. I highly doubt IanD attends them so I would not give credence to what he wrote. I did not attend the Globalfoundry event this month but a coworker did and felt it was very marketing heavy. My $.002.

 

[Daniel Nenni]

This is my first post so I will start with my credentials. I have zero knowledge of semiconductors but I am an investor in Arm, TSMC and Apple.
i would suggest everyone reads all that Zafiropoulos said before deciding that his statements have credibility and, in particular, his response to the question posed by Ted Moreau. In my opinion if C C Wei were to read this he would not buy Z's products unless he absolutely had no alternative. Z flatly contradicts what C C Wei said in the TSMC transcript and one should remember that such statements , if made negligently or falsely, are I believe subject to criminal penalties. In other respects Z's remarks with respect to 16nm in general and TSMC in particular are riddled with inconsistencies. In particular he admits at one point that he does not have a strong position in 16nm! It could be that C C Wei has no opinion of his product.
Just my opinion of course!

Honestly it was confusing for me too. I do not remember hearing about legal issues as a result of one of these calls, have you? I will check for case law.