Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/intel-manufacturing-business-suffers-setback-as-broadcom-18a-tests-disappoint.20913/page-4
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Intel Manufacturing Business Suffers Setback as Broadcom 18A Tests Disappoint

lefty said:
I don't think people read Intel's statement correctly:

It says primarily using external partners. It doesn't say they are no 20A Arrow lake CPU tiles. Officially Intel said that all CPU Arrow lake tiles are on 20A - therefore they had to correct this statement and now they say most Arrow lake CPU tiles are TSMC. In other words, nothing has changed - one or two SKUs will have 20A and the rest will be TSMC N3.
Click to expand...
Intel said Arrow lake was 20A. They have said this for 3 years. CPU tile on 20A. Arrow lake is the only 20A product. Arrow lake is BSP + GAA.
Then it leaked out that Some arrow lake skus will be on TSMC. Intel knew this was true internally but did not comment
Pat also said Lunar lake and Arrow lake would use TSMC. People assumed he was talking about SOC chip. He was not

Now Intel announced 20A is not being used on initial skus. The rumor says some skus are 20A? Which does INTEL say is using 20A.

Question for Investor relations "will any products use 20A from Intel" ... I think it is cancelled but I might be wrong.

N3 and 20A are very different processes with very different layouts. I can't see how you could efficiently support both
 
Xebec said:
An apparent update from Reuters (found on https://x.com/Mojo_flyin/status/1832055416345579629)

“Both Intel and Broadcom responded to Reuters request for comment. Broadcom shared that it has been testing Intel’s foundry Services’ products, and these tests remain ongoing with no final conclusions having been reached.”

“Intel stressed that 18A remains on track for high-volume production in 2025”.
Click to expand...
In related news [dateline Oct 2023] "Intel stressed that 20A remains on track for arrow lake high-volume production in 2024"
 
lefty said:
well the full statement is "Arrow Lake processor family will be built primarily using external partners and packaged by Intel Foundry."
They don't explicitly say there are no 20A SKUs
Click to expand...

Ultimately, the statement you quoted, is meaningless, since it's a forward-looking statement, and as such, is not required to convey reality.
 
SemiFunctional said:
Intel 4 &3 use EUV. What's not cost effective is that nobody is buying intel 4 & 3.
Click to expand...
I stand corrected. I was referring to this slide from Intel which shows cost for Intel 3 as being higher than TSMC. It was my (incorrect) assumption that the cost delta was because EUV had not been implemented. I am still not sure I'm convinced that running Intel 3 and not ramping 18A is a cost effective decision. If this data is accurate then moving to 18A returns all the money Intel is paying TSMC to their pockets at a roughly equivalent wafer cost. Paying yourself to do the manufacturing has to be better than paying someone else. I believe that outsourcing to TSMC when you have a relatively cost effective solution that has comparable to better performance is very short sighted.
1725671725908.jpeg
 
Artificer60 said:
I stand corrected. I was referring to this slide from Intel which shows cost for Intel 3 as being higher than TSMC. It was my (incorrect) assumption that the cost delta was because EUV had not been implemented. I am still not sure I'm convinced that running Intel 3 and not ramping 18A is a cost effective decision. If this data is accurate then moving to 18A returns all the money Intel is paying TSMC to their pockets at a roughly equivalent wafer cost. Paying yourself to do the manufacturing has to be better than paying someone else. I believe that outsourcing to TSMC when you have a relatively cost effective solution that has comparable to better performance is very short sighted.
View attachment 2256
Click to expand...
The problem is whether you can sustain large capital investments. What will happen after 18A, given the majority of the market goes to TSMC (assuming there is no geo-conflict)? Instead, Intel can focus on delivering better products. It seems to be a less risky approach with better rewards.
 
Artificer60 said:
I stand corrected. I was referring to this slide from Intel which shows cost for Intel 3 as being higher than TSMC. It was my (incorrect) assumption that the cost delta was because EUV had not been implemented. I am still not sure I'm convinced that running Intel 3 and not ramping 18A is a cost effective decision. If this data is accurate then moving to 18A returns all the money Intel is paying TSMC to their pockets at a roughly equivalent wafer cost. Paying yourself to do the manufacturing has to be better than paying someone else. I believe that outsourcing to TSMC when you have a relatively cost effective solution that has comparable to better performance is very short sighted.
View attachment 2256
Click to expand...
Actually, what I learned from this chart is advanced package is the only goose with golden eggs now and can lay more eggs.
 
Artificer60 said:
I stand corrected. I was referring to this slide from Intel which shows cost for Intel 3 as being higher than TSMC. It was my (incorrect) assumption that the cost delta was because EUV had not been implemented. I am still not sure I'm convinced that running Intel 3 and not ramping 18A is a cost effective decision. If this data is accurate then moving to 18A returns all the money Intel is paying TSMC to their pockets at a roughly equivalent wafer cost. Paying yourself to do the manufacturing has to be better than paying someone else. I believe that outsourcing to TSMC when you have a relatively cost effective solution that has comparable to better performance is very short sighted.
View attachment 2256
Click to expand...

How much trust can we put in Intel's PowerPoint slides, especially those related to cost advantages and disadvantages?
 
lefty said:
No they didn't. They said primarily external foundry. They did not say 100% external
Click to expand...

Please re-read what I said and what you just wrote:

I said "Now Intel announced 20A is not being used for initial SKUs" . this is 100% true . I did not say 100% external.

You can parse out "primarily" if you wish. you can parse out what "external" means. I am saying initial skus are not on on 20A. This is true

There will be much, much bigger shoes to drop in the next two months... All of the options being worked on are major changes. I wouldn't worry about re-interpreting Intel's statements from last week (just an opinion)
 
lefty said:
Semiaccurate is claiming that 18A has been delayed for more than one year:
www.semiaccurate.com

What is really going on with Intel’s 18a process?

Intel is in deep trouble, the 18a process is broken, and it is all down hill from here, right?
www.semiaccurate.com www.semiaccurate.com
Click to expand...
He is not claiming that. He is listing all the rumors floating around now about Intel\Intel 18A and he is saying all these rumors are false (he shot them down after investigating).

"With that in mind, let us tell you what we found out about the 18a rumors that are swirling around of late. The first rumor is that 18a is significantly delayed, more than a year. Next up is that the yields, to use a term of art, suck. They go on to various versions of performance problems for those parts that do yield. ‘Because’ of all this, companies like Broadcom are publicly trashing Intel up and down.
Then there is the flood of executive exoduses (exodi?) that have been well publicized. Throw in slashed budgets, cut dividends, plagues of emus, stock delisting ‘leak’ rumors, and more just for fun. It is all real, all pointing to a failing company with no chance of anything, right?

Nope. Other than the dividends, budget cuts, and executive departures, weeks of hunting this all down ended up shooting it all down."
 
dkr1986 said:
He is not claiming that. He is listing all the rumors floating around now about Intel\Intel 18A and he is saying all these rumors are false (he shot them down after investigating).

"With that in mind, let us tell you what we found out about the 18a rumors that are swirling around of late. The first rumor is that 18a is significantly delayed, more than a year. Next up is that the yields, to use a term of art, suck. They go on to various versions of performance problems for those parts that do yield. ‘Because’ of all this, companies like Broadcom are publicly trashing Intel up and down.
Then there is the flood of executive exoduses (exodi?) that have been well publicized. Throw in slashed budgets, cut dividends, plagues of emus, stock delisting ‘leak’ rumors, and more just for fun. It is all real, all pointing to a failing company with no chance of anything, right?

Nope. Other than the dividends, budget cuts, and executive departures, weeks of hunting this all down ended up shooting it all down."
Click to expand...

Exactly how did Charlie Demerjian debunk any of the aforementioned rumors, without comments from the various companies, when companies do not comment on rumors?
 
C-4 said:
Intel, and Intel employees who are reading this, I am cheering for you. Each and every one of you. Go prove the rumors wrong. Go 18A!
Click to expand...
Now I see intel becomes focusing on 18A and suspending some manufacturing expansion plans. Although there are still bumps ahead, go for intel. Go for Foundry Competition!
 
Fred Chen said:
The Broadcom response to 18A seems like a bad followup to Qualcomm's previous response (or lack thereof) for 20A: https://semiwiki.com/forum/index.ph...-a-potential-takeover.21046/page-3#post-75227 Last year industry analyst Ming-Chi Kuo wrote that Qualcomm possibly stopped its work on 20A, opting for TSMC or even Samsung instead: https://medium.com/@mingchikuo/qual...e-meaning-that-intel-18a-r-d-and-bc29ea2493d1
Click to expand...
Was Intel 20A planned as a foundry offering and not an internal only node?

I thought Intel’s strategy was 4 for internal, 3 for foundry + internal, 20A for internal, and 18A for foundry+internal?
 
You must log in or register to reply here.
Back
Top