Discussion What if Intel skipped their internal shrink node factories and went with TSMC?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

misuspita

Member
Jul 15, 2006
139
99
101
How long would it take for TSMC to increase their capacity enough to handle what Intel would require?

12 months? 18 months? 24 months? 36 months?
Actually not phisically building the fabs would stop them from expanding at the rate needed to accommodate Intel, as that would proceed quickly enough.

But getting the equipment ready, meaning what ASML makes, the tools that actually make the silicon chips, it would be impossible


They shipped 30 EUV tools this year. How many would be needed to fully equip 3 fabs of 7 or 5 nm needed for Intel only?

That would be the main brake in doing such a thing, assuming that would be needed.
 
  • Like
Reactions: scannall

Kocicak

Senior member
Jan 17, 2019
279
149
76
Intel could just pay TSMC upfront to build one gigafab for them with some modern (and expected to be usable for longer time) production technology (5 nm?) and have the entire capacity reserved for them. They are in such a pickle that TSMC could make a lot of money off them.
 

Kocicak

Senior member
Jan 17, 2019
279
149
76
Semiaccurate writes about something horrible happening at Intel causing another server CPUs delay. It is so horrible that the author is worried that Intel may not survive. Has anybody some clue about what is happening?
 

A///

Senior member
Feb 24, 2017
201
75
71
There are certainly a Taiwanese company first and most but that doesn't mean they are immune to not leaking trade secrets specially if they operate in China and have staff in the country. I have a suspicion multinational companies are easier to exploit.
You mean like how Intel has a fab in China for their NAND, and has two assembly sites in China for their processors? All the semiconductor companies that run their own fabs are multinational and have fabs in various countries including but not limited to just China.

Semiaccurate writes about something horrible happening at Intel causing another server CPUs delay. It is so horrible that the author is worried that Intel may not survive. Has anybody some clue about what is happening?
If they do, they probably have sources and aren't going to burn them.
 
  • Like
Reactions: ksec

Topweasel

Diamond Member
Oct 19, 2000
5,326
1,524
136
t isn’t necessarily the case that TSMC is that many nodes ahead of Intel since the numbers (like 14 nm, 7 nm, etc) are completely meaningless from both companies; they are really just names. You would need to compare specific specs from both processes to see how they compare. intel had the process lead for a long time when competing with AMD/Global Foundries. They aren’t going to have that going forward since TSMC has a massive portion of the market. They have the money and resources to push the process tech, so I don’t really expect intel to take the lead again. They may have some advantages in doing being both the designer and the fab, but TSMC has just a massive portion of the market, so that may not be a big enough advantage going forward.
My point to it was that TSMC is pretty much established their capacity for 5nm and 3nm. 5nm is spooling up and their 3nm factory has already broken ground (and I am sure some area's like their development parts are already up and running). So besides moving to TSMC 7nm in their designs (different then their 14nm), TSMC is already prepped and far a long on another 2 nodes. Not really trying to get into the Intel vs. everyone elses node size debate.
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,031
612
136
Semiaccurate writes about something horrible happening at Intel causing another server CPUs delay.
Wait to see the Xe MCM chips, before saying something bad and not good happened to XCC Xeon.

 

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
Intel cannot keep up with demand for their 14nm chips. How many Intel fabs are producing 14nm chips - 7, 8? Yet somehow TSMC is going to come to their rescue. Are people here high? This is just the most absurd speculation I’ve seen in a while. The capital costs and time to build would be preposterous.
 

Kocicak

Senior member
Jan 17, 2019
279
149
76
Ajay - I do not know the current output and yields of various production processes from various Intel fabs, do you? One large fab making wafers on a modern process which works could be a huge help in my opinion, your opinion may differ.

About the Semiaccurate secret article - I found out that you need to pay 1000 dollars per year to read those articles. Is anybody really paying this sort of money???
 

DrMrLordX

Lifer
Apr 27, 2000
16,020
4,970
136
Semiaccurate writes about something horrible happening at Intel causing another server CPUs delay. It is so horrible that the author is worried that Intel may not survive. Has anybody some clue about what is happening?
Intel is sitting on enough cash for years of operations. Not sure where their cash stockpiles are now, but in 2017 they had over $24 billion stashed overseas alone (not sure what they're carrying domestically). The main concerns would be: talent retention and loss of market cap. Secondary problems would include atrophy of product distribution channels and loss of name recognition. But they should be able to coast for awhile.

edit: okay, Intel is down to $12billion in cash-on-hand for Q3 2019. Must be due to some of those stock buybacks.

 
  • Like
Reactions: krumme

moinmoin

Golden Member
Jun 1, 2017
1,657
1,594
106
Intel cannot keep up with demand for their 14nm chips. How many Intel fabs are producing 14nm chips - 7, 8?
According to wikipedia 8 "fabs" at 3 places. I put fabs in quotes since only one of them appears to be dedicated to 14nm? Intel's problem is that their fabs are a moving target. They don't build a 22nm fab and that is it, but they are constantly convert existing fabs. In the last couple years they botched that process several times, converting to the non-working 10nm too early, then reverting that etc.

Yet somehow TSMC is going to come to their rescue. Are people here high? This is just the most absurd speculation I’ve seen in a while. The capital costs and time to build would be preposterous.
You are completely correct. Nobody can save Intel from themselves, they have to do that on their own either way.
 

Thunder 57

Golden Member
Aug 19, 2007
1,431
1,241
136
Semiaccurate writes about something horrible happening at Intel causing another server CPUs delay. It is so horrible that the author is worried that Intel may not survive. Has anybody some clue about what is happening?
Yes. Charlie is about as anti-Intel as they come. To be fair, he has gotten some things right. But he puts a negative spin on every little thing.

Ajay - I do not know the current output and yields of various production processes from various Intel fabs, do you? One large fab making wafers on a modern process which works could be a huge help in my opinion, your opinion may differ.

About the Semiaccurate secret article - I found out that you need to pay 1000 dollars per year to read those articles. Is anybody really paying this sort of money???
I hope not. If so, well, you know what they say about a fool and his money being parted.
 
  • Like
Reactions: CHADBOGA

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
According to wikipedia 8 "fabs" at 3 places. I put fabs in quotes since only one of them appears to be dedicated to 14nm? Intel's problem is that their fabs are a moving target. They don't build a 22nm fab and that is it, but they are constantly convert existing fabs. In the last couple years they botched that process several times, converting to the non-working 10nm too early, then reverting that etc.


You are completely correct. Nobody can save Intel from themselves, they have to do that on their own either way.
According to Wikipedia there are 12 Fabs for logic wafers (excluding RP1). I think 3 of those are doing 10nm. That leaves up to 9 doing 14nm. I would think one or two of those are already being converted to 7nm.

@Kocicak Several of the D series fabs are 'Mega' fabs as are the fabs in Chandler. I don’t have the numbers, but I’m very confident that the total acres Intel has dedicated to advanced node production is higher than TSMC (allot of their production is dedicated to smallish SoCs, so lots of dice on fewer wafers). This is just a guess since I haven’t found an aggregated list for either company.
 

JasonLD

Senior member
Aug 22, 2017
253
203
86
It'd be better to fab the server wafers at TSMC then to lose all the market share to AMD and ARM.
Well, not going to expect that scenario unless Intel also completely manages to mess up 7nm. I don't expect AMD and ARM to gain 30% of market share in 3-4 years even assuming Intel keeps on struggling.
I understand Intel's struggles with 10nm brings pessimistic expectations of their future nodes, but Intel also happens to be the only company capable of coming back.
 

jpiniero

Diamond Member
Oct 1, 2010
7,924
1,206
126
I understand Intel's struggles with 10nm brings pessimistic expectations of their future nodes, but Intel also happens to be the only company capable of coming back.
The scary thing is that (as Ajay mentioned) they are spending a ton on buybacks. They might simply not have the funds to "come back" since they are spending it all on propping up the stock. A single fab is like a couple billion.
 

JasonLD

Senior member
Aug 22, 2017
253
203
86
The scary thing is that (as Ajay mentioned) they are spending a ton on buybacks. They might simply not have the funds to "come back" since they are spending it all on propping up the stock. A single fab is like a couple billion.
Well, fab expansions should be part of R&D/Operating budget, they don't usually pull that expense out of cash-on-hand when they are long term expenses. That is something I wouldn't worry about unless they are bleeding red.
 

A///

Senior member
Feb 24, 2017
201
75
71
In theory, Intel could adapt their 14nm and maybe even 10nm to another fab's node. This isn't a simple processor and does take a decent amount of time, not to mention a lot of money. I feel it would be like the Sandy Bridge Costa Rica debacle.

So they already have TSMC's knowledge then as well as all of apples, amds and qualcomms, so why would they care about intels?
And if they did as the person presumes, then China's offerings in the mobile market would be as good as Apple's, except they're not. I love Android, but even I know the quality differences are still day and night. Samsung is close, but Android is no iOS. In the same light you could argue China may already have its tentacles in US based businesses with high yield IP.
 

moinmoin

Golden Member
Jun 1, 2017
1,657
1,594
106
Well, not going to expect that scenario unless Intel also completely manages to mess up 7nm.
7nm most certainly isn't doing well. Process node development happens concurrently, and when the extend of the issues with 10nm became clear (which for me was in spring 2018 with the launch of Cannon Lake) I fully expected it to be quickly replaced by 7nm. But that has not happened, instead Intel's dry spell with funny filler products pretending to be real deals continues.
 

JasonLD

Senior member
Aug 22, 2017
253
203
86
7nm most certainly isn't doing well. Process node development happens concurrently, and when the extend of the issues with 10nm became clear (which for me was in spring 2018 with the launch of Cannon Lake) I fully expected it to be quickly replaced by 7nm. But that has not happened, instead Intel's dry spell with funny filler products pretending to be real deals continues.
I don't think you can just jump to the next node like previous node never existed. Pretty sure there are parts of previous node design that will carry over to the next one. if 10nm gets delayed, all future nodes will inevitably gets delayed.

Regardless, Intel wouldn't be in a bad position if they manages to move their consumer/Server CPU lineup to 7nm in 2022. So all the talk about Intel spinning off their fabs or fabbing their CPUs on TSMC is bit premature at this point.
 

Ajay

Diamond Member
Jan 8, 2001
7,064
2,364
136
7nm most certainly isn't doing well. Process node development happens concurrently, and when the extend of the issues with 10nm became clear (which for me was in spring 2018 with the launch of Cannon Lake) I fully expected it to be quickly replaced by 7nm. But that has not happened, instead Intel's dry spell with funny filler products pretending to be real deals continues.
I don’t know how you can be so sure of that. 7nm doesn’t use SAQP, which is one of the biggest problems with 10nm DUV. I believe by now, that Intel has learned what it needed to from other aspects of the 10nm node. Node implementation is incrementally dependent with multiple teams in different stages of research and development. If spending on EUV equipment doesn’t go way up over the next 18 months - that would be very bad news.
 

chrisjames61

Senior member
Dec 31, 2013
596
306
136
In theory, Intel could adapt their 14nm and maybe even 10nm to another fab's node. This isn't a simple processor and does take a decent amount of time, not to mention a lot of money. I feel it would be like the Sandy Bridge Costa Rica debacle.


And if they did as the person presumes, then China's offerings in the mobile market would be as good as Apple's, except they're not. I love Android, but even I know the quality differences are still day and night. Samsung is close, but Android is no iOS. In the same light you could argue China may already have its tentacles in US based businesses with high yield IP.

All the huge factories that produce high tech goods for western companies on mainland China are run by either Americans or Europeans for a reason. The Chinese just don't have the knowledge to produce these things on their own. They can make cheap, inferior knock-offs but that is it.
 

ASK THE COMMUNITY