Intel 10nm delayed by 9 months? (Semiengineering)

[witeken]

I don't see anything new here. Someone from Intel told some time ago that first 10nm products are expected to launch early 2017. That means 10nm production start not before H2 2016.

http://gulfnews.com/business/sectors/technology/intel-to-launch-10nm-chips-in-early-2017-1.1443856

That statement has been retracted by Intel (do you see the problem with "someone" from Intel?). Intel's "official" disclosure about 10nm:

“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 signaling 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 signaling exactly when, what and where.” --Brian Krzanich, CEO Intel, IM’14

“We do think that we've been giving too much insight too far in advance and so we'll talk about 10nm some time in the next 12 or 18 months [from Nov '14] and when it's appropriate.” Stacy Smith, CFO Intel, IM’14

“But you are seeing in the fourth quarter, you’re seeing the front edge of the startup cost associated with the 10-nanometer and that’s kind of right in line with the historical timing of what you’d expect.” Stacy Smith, CFO Intel, IM’14

Unless the delay is too significant (which I actually doubt given the optimism from Intel at ISSCC), there will be 10nm in 2016.

 

[mikk]

Difference is that Brian Krzanich is more likely in a marketing speech mode (as usual) when he says such things in public. Remember what he said in the past, for example Cherry Trail coming 2014. The statement from Taha Khalifa is much more plausible. He wouldn't say this if Intel really planned a 2016 start at this time.

 

[Arachnotronic]

Impressive how R&D budgets of billions of $ and thousands of engineers can be reduced to 3 simple numbers

Just because the end result can be crudely approximated in three simple numbers doesn't mean that the effort to build these things isn't extraordinarily complex.

The argument to prove that Fermat's Last Theorem was true was extraordinarily difficult to come up with, but the "result" can be understood by average high school kids

 

[witeken]

marketing speech mode (as usual)

I'm not convinced. It's called "investor meeting". It has absolutely nothing to do with marketing.

Or are you going to say that Bill Holt was also in marketing speech mode when he said the 14nm problems would be fixed within 1 quarter. It turned out to be 2 or 3. Things can change, just like Cherry Trail.

 

[Arachnotronic]

I'm not convinced. It's called "investor meeting". It has absolutely nothing to do with marketing.

Or are you going to say that Bill Holt was also in marketing speech mode when he said the 14nm problems would be fixed within 1 quarter. It turned out to be 2 or 3. Things can change, just like Cherry Trail.

Investor meetings are absolutely marketing. Executives are literally trying to convince investors that they should invest money in their company rather than the thousands of other publicly traded stocks in the market.

It's a different kind of marketing, where company shares and not commercial goods are the items in question, but marketing nonetheless.

 

[witeken]

Investor meetings are absolutely marketing. Executives are literally trying to convince investors that they should invest money in their company rather than the thousands of other publicly traded stocks in the market.

It's a different kind of marketing, where company shares and not commercial goods are the items in question, but marketing nonetheless.

The presentations, sure, but the Q&A?

And anyway, that's irrelevant. The relevant points is if there's something wrong with the specific quotes.

 

[BD231]

5 years later and the only thing worth noting about Intel CPU's are delays, bravo !

 

[krumme]

5 years later and the only thing worth noting about Intel CPU's are delays, bravo !

It sucks but its deliberate and planned. They have ajusted capex earlier we might start to see the results of that. People is buying iphone and s6 not intel stuff and for servers they own it anyway.

Bs aside, try harder whatever, its all lesser reasons to "upgrade" fast. I think they will go for slower ramp to getter depreciation profile for the new market. No need for space rockets for all some will get a recycled one. Eco model.

 

[Hulk]

5 years later and the only thing worth noting about Intel CPU's are delays, bravo !

And yet most of us are still using them.

 

[mikk]

I'm not convinced. It's called "investor meeting". It has absolutely nothing to do with marketing.

Investor meeting? This what Krzanich told at the Investor Meeeting mid-November:

Cherry Trail out by the end of the year.

At that time CHT was planned for mid-2015. Krzanich doesn't tell us the truth in public.

 

[frozentundra123456]

5 years later and the only thing worth noting about Intel CPU's are delays, bravo !

Well, if you want to start your own chip design company and think you can do better go for it.

I dont buy all this intel is lazy, is holding back because of lack of competition, blah, blah.
CPU design is in a mature stage where all the easy improvements have been made. Each incremental improvement past a certain point becomes more and more difficult. It is extraordinally difficult to design a cpu with a billion or more transistors of nanometer dimensions. I really dont know what people are expecting.

And if you want to compare to the progress of ARM, yes they are progressing much faster. But that is because they are in the early growth phase where large improvements can be made easily. I predict that eventually the rapid improvement in ARM performance will level off as well.

 

[mrmt]

That statement has been retracted by Intel (do you see the problem with "someone" from Intel?). Intel's "official" disclosure about 10nm:

BK and SS didn't commit themselves with any timeline with those blanket statements. 10nm may or may not be delayed and these statements by no means will haunt the execs.

 

[coolpurplefan]

I don't see why they'd want to rush anyway. According to this, they may have to change materials to go to 7nm: http://www.kitguru.net/components/g...development-7nm-will-require-new-transistors/

more here:
http://www.extremetech.com/computin...that-will-take-us-to-the-limits-of-moores-law

http://www.extremetech.com/computing/190845-intel-forges-ahead-to-7nm-without-the-use-of-euv-lasers

 

[Idontcare]

@IDC:

Would you say given what we're potentially seeing - that Intel has slept too much on it dominant node position and may now start loosing some of it's advantage?

It seems despite the turncoats in taiwan - both TSMC & Samsung are (while not 100% comparable in parametric\physical sizes) maybe catching up to confident chipzilla?

I haven't seen that in the trenches. Intel isn't doing things that one would expect if they were getting arrogant or complacent. And we know what that looks like from the days of AMD's intentional 65nm development slowdown.

What I am seeing is that the defectivity issue is a nearly insurmountable challenge at 10nm node physical dimensions. The metrology is barely capable of generating usable data. Everything is operating at the hairy edge of its technical limitations.

And if you can't get data with which you can have confidence in (statistical confidence) then it becomes a lot more challenging to identify the problem, make decisions, and measure the effect of those changes on the same timescale necessary to stick with a 2 year node cadence.

There is a price-premium to be paid by whichever company is out in front, as everyone else can draft behind with confidence knowing they don't need to be spending R&D money scouting out path-finding options or alternatives. Just invest money into replicating that which the leading company has already proven to work.

 

[Phynaz]

It sucks but its deliberate and planned. They have ajusted capex earlier we might start to see the results of that. People is buying iphone and s6 not intel stuff and for servers they own it anyway.

Bs aside, try harder whatever, its all lesser reasons to "upgrade" fast. I think they will go for slower ramp to getter depreciation profile for the new market. No need for space rockets for all some will get a recycled one. Eco model.

Intel just cut $1.3B off CAPEX. No word yet where that is coming from, but the equipment makers stock prices are taking a pounding.

http://finance.yahoo.com/news/intels-revenue-flat-ago-201614545.html

 

[liahos1]

Intel just cut $1.3B off CAPEX. No word yet where that is coming from, but the equipment makers stock prices are taking a pounding.

http://finance.yahoo.com/news/intels-revenue-flat-ago-201614545.html

they said 14nm yields were ramping better than expected. less desktop demand from SMB's was driving less usage of 22nm and they were repurposing equipment to 14nm.

 

[witeken]

they said 14nm yields were ramping better than expected. less desktop demand from SMB's was driving less usage of 22nm and they were repurposing equipment to 14nm.

+ 1.0 point: Lower factory start-up costs

The high 10nm spending that was forecast for Q1 didn't happen.

Edit: there will be more start-up in Q3-4 for 10nm. HVM in Q1?

 

[BD231]

Well, if you want to start your own chip design company and think you can do better go for it.

I dont buy all this intel is lazy, is holding back because of lack of competition, blah, blah.
CPU design is in a mature stage where all the easy improvements have been made. Each incremental improvement past a certain point becomes more and more difficult. It is extraordinally difficult to design a cpu with a billion or more transistors of nanometer dimensions. I really dont know what people are expecting.

And if you want to compare to the progress of ARM, yes they are progressing much faster. But that is because they are in the early growth phase where large improvements can be made easily. I predict that eventually the rapid improvement in ARM performance will level off as well.

Lesser companies do more with less every year, I'm not buyin.

 

[witeken]

From the Q&A, BK said that the capex has nothing to do with 10nm timing, but they haven't talked about 10nm, which they're going to do at a later date.

 

[Fjodor2001]

10nm is essentially TSMC's 14nm.

Does TSMC agree with this? In that case how do they justify calling it 10 nm?

 

[Intel FX Master]

Do most of us need more performance from the CPU right now ? My i5 4690K is more than enough. I want more software optimization that what I think we need.

 

[tommo123]

Lesser companies do more with less every year, I'm not buyin.

who puts out better CPUs than intel? or develops more advances process tech?

 

[Fjodor2001]

Do most of us need more performance from the CPU right now ? My i5 4690K is more than enough. I want more software optimization that what I think we need.

Don't worry, 10 nm is not expected to bring any significant performance improvements anyway.

 

[mrmt]

From the Q&A, BK said that the capex has nothing to do with 10nm timing, but they haven't talked about 10nm, which they're going to do at a later date.

I think this has more to do with smaller 14nm needs than projected at the node ramp up. Instead of Broadwell at a mature state, Skylake ramp up and Airmont in Q115, all we got is... Broadwell in relatively small quantities, which will be replaced by skylake some time later in the year and the mobile sales projections will also be reduced, hence the need for extra contrarevenue in 2015.

 

[witeken]

I think this has more to do with smaller 14nm needs than projected at the node ramp up. Instead of Broadwell at a mature state, Skylake ramp up and Airmont in Q115, all we got is... Broadwell in relatively small quantities, which will be replaced by skylake some time later in the year and the mobile sales projections will also be reduced, hence the need for extra contrarevenue in 2015.

According to BK, 14nm ramp was going a bit faster than expected.