Intel Plans Oregon Expansion for 7nm Chip Production - Report

[Guest]

Intel reportedly plans to build a 1.1-million-square-foot expansion to its Oregon research factory, D1X, to prepare for 7nm chips.

Intel Plans Oregon Expansion for 7nm Chip Production - Report : Read more

 

[redgarl]

Fabless in 10 years, no more advantage for Intel having fabs.

 

[michaelpublic2019]

I think you mean 'fabulous'.

 

[ingtar33]

so it looks like 10nm won't happen. How does intel get a pass for a completely paper launch? Furthermore how is it supposed to launch 7nm when it doesn't really have any fabs for it?

 

[hannibal]

Intel still is big player with enormous amounth of money. They can develop two three production technology at the same time. Paraller architecture development Also... all that just reduces the profit so They don`t do it unless They have to now They have to do it. Intel will come back that is sure, but it sure is nice to have amd in small lead at this moment. It forses Intel not to sit on their thumbs and do minor improvements. Now They really have to work for their money and that is good to customers!
Hopefully TSMC and Samsung can keep up with Intell this time so that it has to be on its toes all the time!

 

[lhughey]

Its gonna be 2014 before they touch 7nm

 

[pug_s]

Fabless in 10 years, no more advantage for Intel having fabs.

I don't know about that. If Intel is smart, they would split up its Fab and CPU manufacturer companies like what AMD did.

 

[shrapnel_indie]

so it looks like 10nm won't happen. How does intel get a pass for a completely paper launch? Furthermore how is it supposed to launch 7nm when it doesn't really have any fabs for it?

Well, they got a pass for what was nearly a paper launch of Broadwell. They scraped by with just a minimal launch in small quantities. and almost immediately launched its successor with new motherboard requirement.

They're managing some 10nm SKUs... but so far it's pretty limited to the mobile market (IIRC)

 

[Et2Brute]

So, by the time Intel has 7nm capability the semiconductor industry will be producing the next node. I think Intel is about to get an important lesson on how hard it is to play catch up.

 

[hannibal]

Intels 10nm is smaller than others 7nm.. so intels 7 is smaller than others 5nm... You can not compare the marketing numbers... They are just that marketing, without any real world relation. But yep. at least at this moment Intel is in situation to catch up, or at least it will be in the summer.

 

[jimmysmitty]

so it looks like 10nm won't happen. How does intel get a pass for a completely paper launch? Furthermore how is it supposed to launch 7nm when it doesn't really have any fabs for it?

10nm is launching later this year.

They started a bit ago retooling FAB42 in Chandler for 7nm, it was originally tooled for 10nm but never got opened. The FAB is built already just needs the equipment. That will probably be the primary site for it with DX1 now being a second site for 7nm production.

 

[none12345]

"Intels 10nm is smaller than others 7nm.. so intels 7 is smaller than others 5nm... You can not compare the marketing numbers... They are just that marketing, without any real world relation. But yep. at least at this moment Intel is in situation to catch up, or at least it will be in the summer."

On paper....or at best in the lab. But, if you cant actually manufacture product on it....is it really smaller?

Lets say they stick to their new timeline and have product for holiday 2019(and that likely means mobile not desktop....and well assume it means in volume this time). Did they meet that schedule with the same paper process they have been showing off for the last 3 years....or has it been gutted to make it manufacturable? Is it still smaller?

I suspect the 10nm they are pushing forward with now is not the same 10nm they have been showing off for the last few years. Its likely been altered, the question is how much has it been altered, is it still a better process on paper or not. This is the path i would take as CEO, id tell them to gut anything that is holding it back and get it out the door ASAP(well i would have told them that back at the end of 2017). There are a few things they can seemingly remove, mainly the cobalt stuff, that would lose some density, but should still result in a quite good process.

Even if their 10nm has not changed, and is still as dense as it was on paper, it will still be ~1.3 years behind. TSMC for instance has been in volume on 7nm for 6 months now, and is at about the same point or slightly behind on their second gen 7nm+ vs intels 10nm(if you believe intel's latest roadmap, which keeps getting pushed back). I cant recall if intels 10nm on paper was better then TSMC's 7nm+ on paper....i don't think it was. You could have intel come out with their 10nm finally at end of 2019, only for TSMC to come out with 7nm+ in early 2020, so intel might have 4 months of being a little faster, before they are much slower again, for another 1-2 years.

There is no sugar coating it at this point. Intel went from a 2(or even a bit more) year lead on process to being about 0.75 years behind(if they meet their current 10nm schedule). Intel's market dominance over the last decade was largely attributable to their process dominance.

Having to compete while behind on process will not be fun for them.

 

[Kaz_2_]

Everything for intel is in theory it might or not might happen. Tsmc has working 7nm and its has good yield they recoup their profit alreafy unlike intel

 

[SockPuppet]

"Intels 10nm is smaller than others 7nm.. so intels 7 is smaller than others 5nm... You can not compare the marketing numbers... They are just that marketing, without any real world relation. But yep. at least at this moment Intel is in situation to catch up, or at least it will be in the summer."

On paper....or at best in the lab. But, if you cant actually manufacture product on it....is it really smaller?

Lets say they stick to their new timeline and have product for holiday 2019(and that likely means mobile not desktop....and well assume it means in volume this time). Did they meet that schedule with the same paper process they have been showing off for the last 3 years....or has it been gutted to make it manufacturable? Is it still smaller?

I suspect the 10nm they are pushing forward with now is not the same 10nm they have been showing off for the last few years. Its likely been altered, the question is how much has it been altered, is it still a better process on paper or not. This is the path i would take as CEO, id tell them to gut anything that is holding it back and get it out the door ASAP(well i would have told them that back at the end of 2017). There are a few things they can seemingly remove, mainly the cobalt stuff, that would lose some density, but should still result in a quite good process.

Even if their 10nm has not changed, and is still as dense as it was on paper, it will still be ~1.3 years behind. TSMC for instance has been in volume on 7nm for 6 months now, and is at about the same point or slightly behind on their second gen 7nm+ vs intels 10nm(if you believe intel's latest roadmap, which keeps getting pushed back). I cant recall if intels 10nm on paper was better then TSMC's 7nm+ on paper....i don't think it was. You could have intel come out with their 10nm finally at end of 2019, only for TSMC to come out with 7nm+ in early 2020, so intel might have 4 months of being a little faster, before they are much slower again, for another 1-2 years.

There is no sugar coating it at this point. Intel went from a 2(or even a bit more) year lead on process to being about 0.75 years behind(if they meet their current 10nm schedule). Intel's market dominance over the last decade was largely attributable to their process dominance.

Having to compete while behind on process will not be fun for them.

Hence why you're not the CEO of anything, because you obviously have 0 clue as to how semiconductor manufacturing works.

 

[urbanman2004]

Sucks for Intel still being unable to produce any worthwhile 10nm yields. Their greed and laziness which contributed to their complacency made them forget all a/b AMD until now where it's too late. I'm laughing and crying tears of joy ????????

 

[mlee 2500]

Fabless in 10 years, no more advantage for Intel having fabs.

Their fabs historically gave their designs a competitive advantage. If that's no longer the case then you could be right.

 

[Afrospinach]

Sucks for Intel still being unable to produce any worthwhile 10nm yields. Their greed and laziness which contributed to their complacency made them forget all a/b AMD until now where it's too late. I'm laughing and crying tears of joy ????????

These comments continue to be absolutely remarkable.

Intel spends nearly as much per year on R&D as AMD has spent this entire century.

Intel makes many times more profit per quarter than AMD has made this entire century.

So they charge a lot for their chips? 87% market share says if anything, they are under priced. Really the evidence says they are neither incompetent, lazy and complacent OR particularly greedy.

I am not sure what great things you expect to happen because Intel is under a bit of pressure. Last time it happened we got the P4 from intel, commonly heralded as the greatest processor ever made, thanks to AMD. Sure. And AMD certainly stepped it up when the core architecture was released, too. A decade later. That competition mate, turning up aces everywhere it goes. It is almost like there is too much inertia in this industry to expect amazing things from competition.

 

[JamesSneed]

Sucks for Intel still being unable to produce any worthwhile 10nm yields. Their greed and laziness which contributed to their complacency made them forget all a/b AMD until now where it's too late. I'm laughing and crying tears of joy ????????

I personally hope AMD does get more market share as that will only be good for the consumer. However you couldn't be more wrong about Intel's 10nm. The issue is that Intel tried to go too dense using quad patterning instead of waiting to move to EUV which cost them in being able to hit high frequencies. Intel's 10nm is more dense than TSMC's 7nm. So if anything Intel was pushing the boundaries of there fab to fast and should have made 10nm less dense until they moved over to EUV.

 

[Co BIY]

I haven't heard a lot about the the importance of Global Foundries dropping out of the process race. For those companies working fabless I think that puts them at the mercy of TSMC.

Even if they have a market beating product, can they get it made? Will TSMC want a large cut of the profit ?

 

[mlee 2500]

Sucks for Intel still being unable to produce any worthwhile 10nm yields. Their greed and laziness which contributed to their complacency made them forget all a/b AMD until now where it's too late. I'm laughing and crying tears of joy ????????

I personally hope AMD does get more market share as that will only be good for the consumer. However you couldn't be more wrong about Intel's 10nm. The issue is that Intel tried to go too dense using quad patterning instead of waiting to move to EUV which cost them in being able to hit high frequencies. Intel's 10nm is more dense than TSMC's 7nm. So if anything Intel was pushing the boundaries of there fab to fast and should have made 10nm less dense until they moved over to EUV.

Yeah, that may be. Their leadership issues also didn't help. Either way both are likely transient issues.

 

[mlee 2500]

I haven't heard a lot about the the importance of Global Foundries dropping out of the process race. For those companies working fabless I think that puts them at the mercy of TSMC.

Even if they have a market beating product, can they get it made? Will TSMC want a large cut of the profit ?

Well, there' always Samsung. Otherwise TSMC would have Apple and numerous other fabless designers by the short and curly's.

I suspect that if it ever did get down to a one-horse race then one of these Fabless Companies with deep pockets might have to bite the bullet. But China (mainland communists) will probably wade into these particular waters with products at an extraordinary discount relative to a smallish performance deficiency (depending on how much IP they can beg, borrow, or steal).

Right now China's own domestic fab's are trailing-edge (~28nm at best), but that only needs to improve a little bit (if at all to) to compete for commodity computing platforms.

 

[Giroro]

Increased competition is good, but the total failure of Intel would be even worse for innovation and consumers than the duopoly that controls the market right now.

As an aside, it's kind of odd that so many of the people actively rooting against Intel speak more-or-less broken English comparable to how you get a lot of broken astroturf out of the Chinese state-sponsored troll farm any time there's a news story that points out one of the many egregious things happening in china right now.

 

[njwhite]

The Trump hating "news" didn't mention this.

 

[littleleo]

Its gonna be 2014 before they touch 7nm

Do you own a Calendar?