Yield challenges in FinFET foundry land

[Arachnotronic]

Is this for high volume production or testing 10nm production ??

They claimed to have 10nm test chips running in late 2013; the 4Q 2015 production date was supposed to be for high volume, commercial production.

 

[witeken]

Bohr was interviewed at IEDM 2014 (Dec. 2014). Intel's last public statement on 10nm was that the plan was to go into production in Q4 2015. Here in Q1 2015, Intel should be close to the finish line if they are indeed going to keep to their original schedule.

That said, Intel expected 14nm to be production viable by Q1 2014, but the process did not qualify for production until Q2 2014. I think this will cause at least a 1 quarter push-out of 10nm with production start in Q1 or Q2 of 2016.

Given that TSMC will go into production on its 10nm in Q4 2016, I would expect Intel to have an ace up their sleeve for this process (crazy tight gate/metal pitches, III-V materials, something...) in order to maintain the "process lead" that it seems so confident that it has.

TSMC's last statement was 2017 HVM.

Now I’ll give you a few words on 10 nanometer development update. Our 10 nanometer technology development is progressing and our qualification schedule at the end of 2015 -- end of this year remains the same. We are now working with customers for their product tape outs. We expect this volume production in 2017.

This is still the same schedule as here: http://seekingalpha.com/article/196...semiconductor-have-very-different-definitions

Also, if you say Intel's 10nm will be delayed due to 14nm delay, then you should apply the same logic to TSMC's 10nm, which would then suffer from 16nm's 2-3Q delay.

Also, TSMC 10nm isn't a competitor for Intel's 10nm, but for Intel's 14nm. Both are second gen FinFETs.

--
I found the comment, it was from Holt: http://intelstudios.edgesuite.net/im/2013/archive/qa1/archive.html (23:30)

 

[Arachnotronic]

TSMC's last statement was 2017 HVM.



This is still the same schedule as here: http://seekingalpha.com/article/196...semiconductor-have-very-different-definitions

Also, if you say Intel's 10nm will be delayed due to 14nm delay, then you should apply the same logic to TSMC's 10nm, which would then suffer from 16nm's 2-3Q delay.

Also, TSMC 10nm isn't a competitor for Intel's 10nm, but for Intel's 14nm. Both are second gen FinFETs.

--
I found the comment, it was from Holt: http://intelstudios.edgesuite.net/im/2013/archive/qa1/archive.html (23:30)

TSMC clarified on its "10nm" timeline.

Roland Shu - Citibank
Just 10 nanometer question to C.C. Since C.C., you said we are expecting to valid [indiscernible] 10 nanometer in 2017. But I remember in the past two quarters actually our goal was up to pull in 10 nanometer [indiscernible] by end of 2015. So [indiscernible] a little bit or not?

C. C. Wei - Co-CEO
Let me explain that because 10 nanometer is the [indiscernible] about 70 to 80. So you got to start in 2016 to have output in 2017. So what I am talking about is 2017 should start to have revenue.

Also, while I would have totally believed that TSMC "10nm" was roughly equivalent with Intel's 14nm before I saw TSMC's claims on scaling at 10nm, I think that no longer will hold true.

If TSMC scales its HD SRAM cell size of 0.07um^2 by 0.45x as claimed, then this leads to an HD SRAM cell size of 0.0315um^2. Intel's 14nm HD SRAM cell weighs in at 0.05um^2. This is appreciably better than what TSMC is doing at 16FF obviously, but quite a bit larger than what TSMC claims is in the pipe for 10nm.

I still think Intel will be "ahead" at 10nm in the way Intel 14nm is ahead of TSMC 16FF (about 1 year in terms of production start timing, and then obviously the density/transistor performance advantages), but I don't think the lead is as big as TSMC's leading edge = Intel's N-1 process.

Now, against Samsung 10nm, Intel's 14nm seems to have very comparable gate/metal pitch metrics and HD SRAM cell size.

Regarding the delay...no, I'm not confident that logic applies here. TSMC is on record stating Q4 2016 production for 2017 revenue for 10nm. Intel, which was previously on record for Q4 2015 10nm production, has decided that it doesn't want to tell us much!

We'll see how it all plays out.

 

[AtenRa]

They claimed to have 10nm test chips running in late 2013; the 4Q 2015 production date was supposed to be for high volume, commercial production.

Intel was saying Q4 2015 for 10nm when they were believing 14nm would be Q4 2013. I dont believe they will start HVP before Q4 2016 the earliest.

 

[Arachnotronic]

Intel was saying Q4 2015 for 10nm when they were believing 14nm would be Q4 2013. I dont believe they will start HVP before Q4 2016 the earliest.

No. The 10nm timeline was given at the Nov. 2013 investor meeting when 14nm delay was already well-known. Now, at the time I believe Intel was expecting a 1 quarter delay, but they got a 2 quarter delay. That's why I think that HVM will happen in Q1 2016.

I further expect Skylake to be the "Broadwell" of 2016, with the Skylake-U/H parts taking care of the first 3/4 of 2016 PC demand, and Cannonlake coming in around IDF 2016 timeframe for holiday PC sales.

If Intel is smart, they will put Broxton into high volume production by EOY 2015 for MWC 2016 unveil, and then bring out a 10nm Atom right alongside Cannonlake in late 2016. This would give Intel a time to market edge on low power 10nm parts, and help bolster its competitive positioning in mobile.

 

[elemein]

The excuses... Now the "merits of a process" are totally unrelated to the products produced on it? You guys are bending over backwards mentally to explain the last five years of intel's foundry progress.

You know when I was thinking about the massive leaps in performance ARM is giving (I.e. 100% or often more improvement without even a change in process) I thought about intel's history and for the life of me I can't think of a single generation of Intel chips where the next one (I.e. Haswell to Broadwell i7) gives 100% improvement even once!

Intel has always been a terrible processor company with a decent foundry. The problem is that so much of their business is built on bullying competitors, lying, and outright removing any competition that they never learned to make a decent product.

What they did learn, is how to convince a generation of AT members, supposedly educated individuals, that 2+2=5. Do the math! Find me one generation of processor where Intel can give even just %100 improvement!

I won't ask you to give the 200-500% increases apple often gives.

Cue the rabble of Intel apologists to come explain why all the benchmarks are unfair, Intel is still the fastest and finally they really do have a new faster architecture, they are just unloading this crappy one on those silly customers for margin and will soon release skylake which will blow us all away.

You must be joking. We already went over this TreVader and you said you understood why it was hard to make large increases.

Edit: Didnt even notice this had a pg2, sorry, my bad ignore.

 

[TreVader]

1. My point about the technical merits of a process was only to point out that you can in fact have inferior products on a superior node, because the node is only an enabler for your chip design. It is not the sole and probably not even the most important factor.

2. I think you don't have a good idea about how similar CPUs really are. It's not like AMD or Intel or ARM or Apple has some secret sauce or significantly better engineers that makes their CPU significantly faster than everybody else's. What sets them apart is the tradeoffs they choose to make during the design process and how well those tradeoffs fit their product into market segments. Do you think Apple/ARM or anybody else will keep giving you 100% increases per generation? The answer is clearly no if you have any understanding of how the industry as a whole is progressing. So when you say "Intel is a terrible processor company", I have a hard time understanding your position. They have released the best desktop/server class CPUs for years now. However I would not disagree that they don't have their mobile products positioned low enough on the power scale to compete with ARM and Apple designs. That doesn't mean they're a "terrible processor company", it means they're a processor company who has traditionally focused on frequency targets and has had trouble shifting their focus onto power targets. Although it boggles my mind why Intel is even trying, you don't make 60% margins in mobile.

Thank you for clarifying your statement. The way I read it, it sounded like an absurdity.


Interesting points, I can see why Intel may not have made such giant performance jumps vs Apple and Qualcomm.

 

[witeken]

Regarding the delay...no, I'm not confident that logic applies here. TSMC is on record stating Q4 2016 production for 2017 revenue for 10nm. Intel, which was previously on record for Q4 2015 10nm production, has decided that it doesn't want to tell us much!

We'll see how it all plays out.

I think I or someone else has done it before, but it's possible to take a look at the delta between risk and volume production of earlier nodes. This allows for an educated guess. From "customers can start early work" EOY'15 to HVM in '16 seems too optimistic, certainly if you keep in mind the delta becomes longer. ShintaiDK or Homeles showed a slide that illustrates that, a long time ago.

I somehow managed to find it quite quickly:

 

[witeken]

Intel was saying Q4 2015 for 10nm when they were believing 14nm would be Q4 2013. I dont believe they will start HVP before Q4 2016 the earliest.

Not true.

The slide where Brian Krzanich states 10nm will go into production in 2015...

was shown at the Investor Meeting 2013 in November, a long time after they knew 14nm had yield issues (which they announced to their investors at the Q3 earnings call). They thought the delay would be 3 months back then, (but it turned out to be 6 months and 14nm probably still isn't at the same level as 22 probably) but at the Q2 earnings call, BK reiterated that they had not done any changes or shift to their 10nm schedule, and now, as can be seen in their Q4 margins, Intel has indeed started with the start-up costs of 10nm:

http://files.shareholder.com/downlo...2757B4612/CFO_Commentary_Q4_2014__-_FINAL.pdf

- 0.5 point: Higher factory start-up costs

If I go 2 years back in time, then the Q1'13 CFO commentary shows the Q1 costs (there was none in Q4'12):

- 2.5 points: Higher factory start-up costs

Q2:

- 2.0 points: Other Cost of Sales (primarily higher 14nm factory start-up costs)

In Q3 it comes down:

+ 0.5 point: Lower factory start-up costs

Back to '15, Intel is expecting the bulk of the start-up costs in H1. In Q4 it was already 0.5 points (originally forecast to be 1.0 point). For Q1:

• - 1.5 points: Higher factory start-up costs
Also of interest:
• - 0.5 point: Higher platform* write-offs (primarily higher pre-qualification costs on Skylake)

The Skylake write-offs are in line with Haswell's; about 1 quarter behind (so 9 quarters delta). IIRC, Haswell was originally planned to launch early in Q2, but was delayed to mid-'13 for inventory reasons or so.
The start-up costs are as you'd expect for a Q4 production start.

So you are 4 quarters off in your bet.

 

[witeken]

I further expect Skylake to be the "Broadwell" of 2016, with the Skylake-U/H parts taking care of the first 3/4 of 2016 PC demand, and Cannonlake coming in around IDF 2016 timeframe for holiday PC sales.

No, 10nm should come earlier. I expect 12-13 months after Skylake.

 

[Arachnotronic]

No, 10nm should come earlier. I expect 12-13 months after Skylake.

12-13 months after Skylake is exactly the timeline I just said

Skylake is holiday '15, Cannonlake will likely be holiday '16.

 

[witeken]

12-13 months after Skylake is exactly the timeline I just said

Skylake is holiday '15, Cannonlake will likely be holiday '16.

If you look at my post above, you'll see that Skylake is on track for Computex launch.

 

[aigomorla]

when i was sampling processors...

roughly, a cpu sits as "house internal" 2 years before its released.
As the name implies.. it sits inside intel's factories, tested by there engineers.
It NEVER leaves the house... and rarely does the public ever know of its existence.


Then the cpu gets upgraded to "Alpha vendor Engineering Sample" 1.5 yrs to 1 year before launch.
(at this stage the cpu has no name... its given a generic Tag...
Note name on CPU-Z

Then the cpu then gets upgraded to "Beta Engineering Sample with Product Tag" 1 year to 6 months before release...
This is the common ES's you see where it has a proper tag / name.

Usually by the time it gets a product name, the yields are typically within acceptable standards, unless again as i stated, intel wants to tweak an already efficient yield even more.

Then after that you get your first batch of retail cpu's. And production tweaking doesnt stop there.
You then get step revisions, to further increase yield capacity, and overall efficiency of the cpu.

So when intel is late, what does that tell me?
They have no competition, no need to rush, and wants to optimize the yield, and stepping revisions so we dont get variants in stepping like we used to with all the other processors which came before.... ie.. Q6600 -> B1 - B2 -> B3 -> G0

So my guess... intel already has its working 10nm sample.

 

[Arachnotronic]

If you look at my post above, you'll see that Skylake is on track for Computex launch.

Maybe Skylake-S (since the desktop is well overdue for a CPU refresh), but I don't think we'll see Skylake mobile parts until later.

 

[TreVader]

No rush? Intel has never been challenged the way they are today. I think they have every reason to rush, and I think they are rushing as fast as they possibly can.

 

[III-V]

No rush? Intel has never been challenged the way they are today. I think they have every reason to rush, and I think they are rushing as fast as they possibly can.

There's not so much need for them to rush their flagship line. Their Atom line, though, certainly.

 

[Idontcare]

IDC,

I would be very curious as to any insights on why Intel might be quiet about 10nm in light of the quotes witeken posted. Their R&D head is claiming that 10nm won't have the same 14nm yield challenges, their CFO is saying that the startup costs will be on a "normal" cadence, and at the 2013 investor meeting they already said they're aiming for Q4'15 production.

And yet, they don't want to directly answer questions on 10nm right now.

What's your take on this?

I'm personally perplexed because they don't say the same thing to the suppliers. Suppliers are being told they have 10nm yield and timeline issue and that everyone is scrambling in recovery mode.

The public statements provide perfect cover for not having to explain that 10nm is shaping up to be an accentuated repeat of 14nm yield-induced delays, but isn't it illegal or something for official statements to be misleading?

So my conclusion at this point is that nothing right now makes sense. There is too much contradictory information. If you are an analyst then you are being directed to believe Intel is withholding information as matter of strategic upper-hand maneuvering, but if you are a supplier then you are scrambling to get customer demos cycled fast enough with high enough priority because yield issues are dominating the development engineers minds right now.

My conclusion is that I don't even want to wade into the debate, time will sort this all out and then we'll know soon enough which part was malarkey.

I will say this though, when Intel sent 600 Ireland expats packing (sent back to Ireland and cancelled their 14nm training) they made that decision because they already knew (had all the data they needed) that 14nm yields were unacceptably low and that a big delay was in the offing, and yet they played it down publicly and acted as if it was all part of a "smart capex alignment" move owing to unforeseen shifts in market demand.

the ramp of 14-nm production in Ireland has been pushed out six months to late 2013, attributing the delay to a slowdown in demand rather than technical problems.

You know Intel knew of the 14nm yield issues at the time because you can easily look at their own published 22nm vs 14nm yield graphs and note that the decision was based on the craptastic 14nm yields at that time, not unforeseen market dynamics. They knew the problem was technical and internal but they spun it as being external and unforeseen. These guys are not above withholding materially relevant information when it comes justifying are rationalizing major internal decisions.

Is the public just being fed more smokescreen on 10nm? History suggests yes.

 

[Arachnotronic]

I'm personally perplexed because they don't say the same thing to the suppliers. Suppliers are being told they have 10nm yield and timeline issue and that everyone is scrambling in recovery mode.

Wait, what? You mean Intel is telling its customers that it is having yield issues with 10nm even though Mark Bohr is on record stating that he doesn't expect such issues at 10nm?

If what you are saying is true, then no wonder they are saying nothing about 10nm status/timelines. This allows them to not make any knowingly false or misleading statements because they aren't making any statements.

At the time of the 2013 investor meeting, things may have been going according to schedule for them to say that they planned to go into production in 4Q 2015.

 

[witeken]

I'm personally perplexed because they don't say the same thing to the suppliers. Suppliers are being told they have 10nm yield and timeline issue and that everyone is scrambling in recovery mode.

Citation needed. I haven't heard any credible statements about Intel's 10nm except from Intel.

The public statements provide perfect cover for not having to explain that 10nm is shaping up to be an accentuated repeat of 14nm yield-induced delays, but isn't it illegal or something for official statements to be misleading?

We'll see... (but from what I know I consider your post a conspiracy theory)

I will say this though, when Intel sent 600 Ireland expats packing (sent back to Ireland and cancelled their 14nm training) they made that decision because they already knew (had all the data they needed) that 14nm yields were unacceptably low and that a big delay was in the offing, and yet they played it down publicly and acted as if it was all part of a "smart capex alignment" move owing to unforeseen shifts in market demand.

Very interesting article, but I think you misinterpreted it: Intel didn't say anything, the whole article is a rumor, so how could they have played it down? And by the way, I'm not sure if Intel was really worrying:

Sure, it starts lower (but it's still a long way from 0), but until July, there weren't any major problems and yields were improving considerably, and at the Investor Meeting, they predicted a 3 month delay based on the yield learning pace.

We also don't know how it went before that time, since the article is about half a year before where the graph starts. Also, history suggests that when Intel faces problems, they just try harder. Intel knows they have to deliver 14nm within at most 2.5 years, so why would this article have anything to do with yields when the process will only start HVM more than a year later? I think there are too many unknowns to base anything on this article.

You know Intel knew of the 14nm yield issues at the time because you can easily look at their own published 22nm vs 14nm yield graphs and note that the decision was based on the craptastic 14nm yields at that time, not unforeseen market dynamics. They knew the problem was technical and internal but they spun it as being external and unforeseen. These guys are not above withholding materially relevant information when it comes justifying are rationalizing major internal decisions.

See above, the article was just a rumor which Intel didn't comment on, and it's unsure whether Intel really thought the lower yield learning was indeed an issue (that would cause more than 1-2 months delay) until June. You can't predict yield learning, as far as I know.

Is the public just being fed more smokescreen on 10nm? History suggests yes.

I strongly disagree, but we'll see .

 

[witeken]

Wait, what? You mean Intel is telling its customers that it is having yield issues with 10nm even though Mark Bohr is on record stating that he doesn't expect such issues at 10nm?

If what you are saying is true, then no wonder they are saying nothing about 10nm status/timelines. This allows them to not make any knowingly false or misleading statements because they aren't making any statements.

At the time of the 2013 investor meeting, things may have been going according to schedule for them to say that they planned to go into production in 4Q 2015.

I like this take from Russ Fisher:

Now we are ending 2012 and the four 22nm fabs are online. According to Mark Bohr's comments during multiple interviews at the recent Intel Developers Forum, "22nm yields are excellent," "22nm yields are better than we expected," and "22nm yields are the same as 32nm." Bohr is an Intel Technical Fellow and Chief Process Technologist. Bohr, like many technical professionals, is genetically incapable of lying or even fibbing. Now, if Bohr were a marketing guy, his comments might be suspect.

1. Mark Bohr says 22nm yields are excellent
2. Russ says he doesn't lie
3. Intel confirmed 22nm yield recently
4. => Mark Bohr doesn't lie

5. Bohr says "We don’t expect we’ll have similar problems at 10 nm, because we’ve learned and we’re trying harder."
6. => History suggest he's telling the truth.

 

[aigomorla]

No rush? Intel has never been challenged the way they are today. I think they have every reason to rush, and I think they are rushing as fast as they possibly can.

challenged by who?
who is in the x86 market against intel?

who do you think makes all the chipsets for the boards we use on our x86 processor?

What would be the benifit of intel to rush a processor on low end, when the market saturation of SoC is already at its near peak?
It would not be worth its cost unless intel decided to goto war against samsung / apple which they are specifically not intending on doing so.

There's not so much need for them to rush their flagship line. Their Atom line, though, certainly.

why?

The Atom line is meant as low power x86 processor.
Its not intended on use for cell phones.
The Tablet possibly, but Atom's are used on servers like the C2750 series, which is fairly polished atom server.

The only competitor they have for something like is AMD, which is currently cutting themselves lean with all the people who got laid off from upper management.

Again... is intel going to war with samsung or apple like how crucial / micron went to war against samsung and gave us all those delicious SSD prices?

Going to war against another competitor would oversaturate an already saturated market, with very little returns for the company itself.
I dont think intel see's that as any cost benefit or gain, and they most definitely doing want to kill out all competition in every processor market field.

 

[witeken]

challenged by who?

Paranoia.

Again... is intel going to war with samsung or apple like how crucial / micron went to war against samsung and gave us all those delicious SSD prices?

Going to war against another competitor would oversaturate an already saturated market, with very little returns for the company itself.
I dont think intel see's that as any cost benefit or gain, and they most definitely doing want to kill out all competition in every processor market field.

Brian Krzanich might have been excited that he succeeded in his 46M/4X goal for 2014, but really, Intel's plan has always been world domination. Intel, as a computing company, simply cannot afford not to have tablet, IoT and smartphone foothold. You can't do that without having a compelling offer: both in terms of price and features.

 

[Idontcare]

IDC,

I would be very curious as to any insights on why Intel might be quiet about 10nm in light of the quotes witeken posted. Their R&D head is claiming that 10nm won't have the same 14nm yield challenges, their CFO is saying that the startup costs will be on a "normal" cadence, and at the 2013 investor meeting they already said they're aiming for Q4'15 production.

And yet, they don't want to directly answer questions on 10nm right now.

What's your take on this?

Citation needed. I haven't heard any credible statements about Intel's 10nm except from Intel.

Citation? I think you misunderstood my post, I was responding to Arachnotronic's post which specifically asked me for my "take".

If my "take" was based on a series of citable public domain sources then it wouldn't really be anything "insightful" now would it? Instead it would just be me parroting what a bunch of other people have already said and written elsewhere.

In the same token, my "take" is going to be formed by things I happen to experience first-hand, which may be things that the average forum member might not have the requisite level of education or work experience to correctly decipher, understand, or be privy to in the first place.

If I am to ignore my position and experiences, restricting my conversation to the realm of that which can solely be cited in the public domain then there isn't much point to me being an active forum member, I may as well become a lurker and just read what other people assemble from press releases on the web.

I can't speak to the motivations of high-level individuals at Intel or the reason for them delivering the statements they gave during interviews, but I can tell you there is a bit of a contradiction between what I read in those interviews versus the angst and trepidation that is unavoidable when 10nm is the topic of discussion with certain individuals who make a livelihood by delivering the very thing certain high-level people have gone on record as saying is hunky-dory and peachy-keen.

Look at the recent Nvidia 970 snafu where you have a PR team that made assumptions about the technical aspects of the product and didn't bring in the technical team to verify. The technical team always knew the reality of the 970 SKU (they created it, after all), and there were a lot of marketing and PR Nvidia people who didn't know, but that didn't keep those Nvidia people from misstating reality and it going uncorrected for months after the fact.

 

[witeken]

Citation? I think you misunderstood my post, I was responding to Arachnotronic's post which specifically asked me for my "take".

If my "take" was based on a series of citable public domain sources then it wouldn't really be anything "insightful" now would it? Instead it would just be me parroting what a bunch of other people have already said and written elsewhere.

In the same token, my "take" is going to be formed by things I happen to experience first-hand, which may be things that the average forum member might not have the requisite level of education or work experience to correctly decipher, understand, or be privy to in the first place.

If I am to ignore my position and experiences, restricting my conversation to the realm of that which can solely be cited in the public domain then there isn't much point to me being an active forum member, I may as well become a lurker and just read what other people assemble from press releases on the web.

I can't speak to the motivations of high-level individuals at Intel or the reason for them delivering the statements they gave during interviews, but I can tell you there is a bit of a contradiction between what I read in those interviews versus the angst and trepidation that is unavoidable when 10nm is the topic of discussion with certain individuals who make a livelihood by delivering the very thing certain high-level people have gone on record as saying is hunky-dory and peachy-keen.

I hadn't realized your claims were based on secret information, fair enough.

So then I guess the question becomes: how much insight do those people have in Intel's yield learning? 'Cause else it wouldn't be very insightful after all, since yields are tightly secured, right?

Look at the recent Nvidia 970 snafu where you have a PR team that made assumptions about the technical aspects of the product and didn't bring in the technical team to verify. The technical team always knew the reality of the 970 SKU (they created it, after all), and there were a lot of marketing and PR Nvidia people who didn't know, but that didn't keep those Nvidia people from misstating reality and it going uncorrected for months after the fact.

But Bohr isn't exactly someone from the marketing team, if that's what you meant. He's in fact leading the 7nm effort, and he's also seriously investigating immersion lithography, so 7nm probably isn't planned for 2019 either, and 10nm must fall somewhere in between.

Intel also isn't going to go full throttle on 10nm start-up multi-billion spending spree costs in Q1 when in fact there will be nothing to produce in high volumes at the end of the year. Moore's Law is still alive at Intel, not sure about other companies though.

“We felt like we went on a little early with 14nm as far as timing and performance and features and we saw actually competitors adjust to that. So we're gonna be a little bit more prudent, a little smarter about signalling to the industry exactly when, what and where. And you'll have to trust a little bit the 50 year history we have with Moore's Law and that we should be able to keep it going for 51 or 52 years. So we're gonna be a little careful there about that signalling exactly when, what and where.” --Brian Krzanich, CEO Intel, IM’14

What's Homeles' take on this -- if he's still around?

 

[mrmt]

Citation needed. I haven't heard any credible statements about Intel's 10nm except from Intel.

I've been reading and listening to the last Q&A with investors, the main problem IMO is Intel lack of commitments to firm timelines with regards to their 10nm process. When Intel was a closed foundry business a new node wouldn't change much in Intel financial modelling except for the cost drivers, today with Intel trying to become an open foundry it changes everything, because the node should also become a revenue stream.

And think about it, it wouldn't make sense for Intel to open up their bleeding edge foundries to other MPU companies while at the same time disclosing less information about its future nodes. It doesn't help Intel to gather news customers, it doesn't generate brand awareness, it doesn't inspire confidence in Intel R&D pipeline and in the 10nm product (the node itself) overall. We can see Intel committing to product launch dates, a node should be no different, but yet Intel refuses to do so. Were Intel really sure of what they could deliver they would say a simply "you know about our tick-tock cadence, the new node is well on its way".

Am I saying that Intel will be getting late with 10nm? No. What I'm saying is that 10nm is giving them enough headaches to the point they can't commit themselves to a given timeline, and if so the risks of a delay are palpable.

“We felt like we went on a little early with 14nm as far as timing and performance and features and we saw actually competitors adjust to that. So we're gonna be a little bit more prudent, a little smarter about signalling to the industry exactly when, what and where.

Had Intel ramp up 14nm with the same speed of 32m and 22nm, BK wouldn't be saying that because he would have had 14 products on the market at the time their rivals would be scrambling to launch 20nm. Because they were late, they can say on the record that they disclosed things a little too yearly (because they disclosed based on an unreal but optimistic projection, not based on the real 14nm timeline).

Which brings us to 10nm. If they are taking too long to disclose things and history is a reference, then they aren't really sure about 10nm.