Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/why-is-intel-reserving-tsmc-n3-given-its-own-node-roadmap.14669/
        )

    [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] => 2020771
            [XFI] => 1050170
        )

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

Why is Intel reserving TSMC N3 given its own node roadmap?

BlueNode

New member
Hi all -- Intel's decision to use TSMC's N3 process would make sense if Intel was stopping its own development, and was exiting manufacturing. But it's not, and its own comparable nodes are scheduled for the same timeframe as N3.

What Intel was calling its 7nm, now Intel 4 or something, is due in early 2023 (this is after various delays). TSMC's N3 was delayed to late 2022 last I saw. What's the point of developing products in a completely different node when you have your own process? Did Intel say anything about the specific product categories they would use N3 for? It seems like a lot of redundant engineering effort.

Thanks.
 

Jumper

New member
More capacity = more money at this point. Maybe it will change and the market will normalize in some way but Intel wants to capture as much as they can. And they can do that only with more capacity.
 

Daniel Nenni

Admin
Staff member
Hi all -- Intel's decision to use TSMC's N3 process would make sense if Intel was stopping its own development, and was exiting manufacturing. But it's not, and its own comparable nodes are scheduled for the same timeframe as N3.

What Intel was calling its 7nm, now Intel 4 or something, is due in early 2023 (this is after various delays). TSMC's N3 was delayed to late 2022 last I saw. What's the point of developing products in a completely different node when you have your own process? Did Intel say anything about the specific product categories they would use N3 for? It seems like a lot of redundant engineering effort.

Thanks.

TSMC N3 is not late. That is fake news. Apple will have their version of TSMC N3 in the 2022 iProducts, absolutely. N3 risk production is Q4 2021 and high volume production in 2H 2022. This is the standard Apple schedule.

CC Wei said that N3 was a longer development cycle than N5 but he did not say it was late. "Our N3 technology will use FinFET transistor structure to deliver the best technology maturity, performance and cost. Our N3 technology development is on track with good progress. N3 risk production is scheduled in 2021 and volume production is targeted in second half of 2022."

 

Daniel Nenni

Admin
Staff member
Bob Swan, former Intel CEO, signed the N3 deal with TSMC to hedge a bet against Intel getting their process development house in order. Bob did not have the technical chops to motivate his team so he put their feet to the fire with the TSMC deal. It was a great move by Bob in my opinion. It certainly motivated them but it really confused Wall Street which brought back Pat Gelsinger. Pat took a different approach. He used his 30+ years of Intel experience and brought back his manufacturing cronies to get the Intel manufacturing house in order. He still has TSMC as a fall back but if his new process roadmap does not pan out he will probably be job shopping anyway.
 
Last edited:

benb

Active member
I don’t know what the thought was with N3, but it probably wasn’t to restrict AMD or Nvidia from
”getting” the capa.

One thought you have as a responsible chip company is to have a disaster continuity plan. Intel has always had at least two fabs, one domestic and one foreign, produce a node. Maybe N3, due to the weirdness lately of Intel ticks and tocks, wasn’t going to have a foreign or a domestic fab available, so they were like, light bulb, let’s let TSMC be part of the plan, or even all of the plan.

On the other hand, extending the Virtual Factory to TSMC is very un-Copy Exact! which is a keystone of Intel culture. Maybe it indicates an evolution of Intel culture.
 

Portland

Active member
So you are suggesting that had Intel chosen David Perlmutter for CEO instead of BK Intel would be fabless?

It was interesting comments on working with Steve jobs and the struggles with architecture. He didn't hold back and what he said was controversial enough that i'm being vague with it. He's saying it is different than some fan or analyst.

He was/is a brilliant and successful engineer but that doesn't make a great CEO alone.

It's such a different world today. Back then Intel was king and not a company struggling against tsmc ecosystem. We're see machinery like Goss being replaced by landa and I'm wondering if asml is next. Let's be honest no one takes people with Henry Kissinger voice seriously but today you can train yourself out of it. There's more opportunities available.
 
Last edited:

VCT

Member
Intel will have very limited production capacity for Intel 4 in 2023.
If Intel can not have 3nm support from TSMC. Intel market share will decrease a lot in 2023 and 2024.

It takes two years for Intel to provide 10nm Desktop CPU after 10 nm laptop CPU launched.
 

soAsian

Member
Bob Swan, former Intel CEO, signed the N3 deal with TSMC to hedge a bet against Intel getting their process development house in order. Bob did not have the technical chops to motivate his team so he put their feet to the fire with the TSMC deal. It was a great move by Bob in my opinion. It certainly motivated them but it really confused Wall Street which brought back Pat Gelsinger. Pat took a different approach. He used his 30+ years of Intel experience and brought back his manufacturing cronies to get the Intel manufacturing house in order. He still has TSMC as a fall back but if his new process roadmap does not pan out he will probably be job shopping anyway.
lol. kind of harsh words there.
 
Top