Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/intel-ceo-tan-reconsidering-fate-of-chipmakers-new-manufacturing-tech-cfo-says.24680/
        )

    [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] => 2031070
            [XFI] => 1060170
        )

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

Intel CEO Tan reconsidering fate of chipmaker's new manufacturing tech, CFO says

Daniel Nenni

Founder
Staff member
1772695345613.png


March 4 (Reuters) - Intel CEO Lip-Bu Tan is now starting to recognize its 18A manufacturing technology as a potential offering for ‌external clients after relegating it largely to internal use last ‌year, Chief Financial Officer David Zinsner said on Wednesday during a tech conference in San Francisco.

This could mark a reversal from a major facet of Tan's turnaround strategy set out last year, when he said he believes Intel's so-called 18A manufacturing process — in which his predecessor Pat Gelsinger had deeply invested — ‌could generate a reasonable ⁠return only if it is used for Intel's own products.

Shares of the company were up about 6% amid a ⁠broader uptick across chip stocks.

"While Lip-Bu was ... thinking that we probably should focus on 14A as a foundry node and make 18A really just an internal node, now that we've got seen some real progress there, I think ‌he's now starting to recognize that this is actually a good node to offer to external customers as well," Zinsner said at the Morgan Stanley Technology, Media and Telecom conference on Wednesday.

Reuters has reported that only a small percentage of the chips printed via 18A have been ‌good enough to make available to customers. Intel has said its yields, or the number of good chips per silicon wafer, are improving monthly. Weak yields also routinely ‌pressure margins.

Since his appointment as CEO, Tan has made big changes to Intel. Last year, Intel cut roughly 20% of its workforce as Tan reshaped the company's strategy to tackle artificial intelligence.

Tan has also vowed ‌to continue to operate Intel's factories and pursue new customers for its next-generation manufacturing tech called 14A.

 
Il semblerait que le PDG d'Intel, Lip-Bu Tan, soit en train de revoir sa position concernant la technologie de fabrication 18A. L'année dernière, il affirmait que ce procédé devait rester interne et être utilisé uniquement pour les puces Intel, car c'est de là que proviendrait la rentabilité. Désormais, il semble envisager de le vendre à des clients externes.

Le directeur financier a mentionné lors d'une conférence que des progrès significatifs https://leon-casinos.fr/ avaient été constatés avec la technologie 18A, suffisamment pour que Tan reconsidère sa position. Ce qui est intéressant, car il y a peu, Reuters signalait des problèmes de rendement : seul un faible pourcentage de puces était utilisable. Intel affirme cependant que ces chiffres s'améliorent chaque mois.

L'action a bondi d'environ 6 % suite à cette annonce, même si le secteur des semi-conducteurs connaissait déjà une bonne journée.

Depuis sa prise de fonction, Tan a mené une politique assez agressive : il a supprimé près de 20 % des effectifs l'année dernière, misant fortement sur l'intelligence artificielle. Il semblerait qu'il soit désormais disposé à confier un rôle plus important à 18A dans l'activité de fonderie, au lieu de la gérer uniquement en interne.

Il est encore trop tôt pour tirer des conclusions, mais la situation est à suivre. Si les rendements s'améliorent réellement, proposer 18A comme option pour les clients externes pourrait changer la donne.
 
Last edited:
Exactly. GO INTEL!
Intel offered it heavily to any and all takers in 2021-2024. Why were there not takers? what has changed? Why is Intel still planning to have 25-30% of its silicon external in 2029

18A planned peak is in 2030. by the time any external customers ramp, 18A will be pretty mature
 
Intel offered it heavily to any and all takers in 2021-2024. Why were there not takers? what has changed? Why is Intel still planning to have 25-30% of its silicon external in 2029

18A planned peak is in 2030. by the time any external customers ramp, 18A will be pretty mature
Maybe because after recent process screw-ups at Intel people wanted to see a real process in real production and yielding before signing up to it?
 
Maybe because after recent process screw-ups at Intel people wanted to see a real process in real production and yielding before signing up to it?
This is true. but since 2025, do you think that IFS has improved its reputation with constraints, yield, costs, and Intel continuing to outsource to TSMC going forward?

Which brings up a question: Why do you think intel choose to use N2 Nova lake for 90% of Nova lake wafers (CPU die)?
 
This is true. but since 2025, do you think that IFS has improved its reputation with constraints, yield, costs, and Intel continuing to outsource to TSMC going forward?

Which brings up a question: Why do you think intel choose to use N2 Nova lake for 90% of Nova lake wafers (CPU die)?
Since 2025 Intel has now got Panther Lake into mass production in 18A, which gives customers confidence that they can actually deliver what they promised -- unlike some previous processes like 10nm where thy tried to do a lot of new stuff and fell over.

Exact yields and costs are a minor concern (and can be argued about/improved later) compared to having a process that can't actually be mass-produced and can destroy your business -- so this will make customers much more willing to go with Intel, at least in 18A -- which seems to be the gist of the announcement.

This may also have increased customer confidence in 14A, though the proof of the pudding is still in the eating.

Why did Intel (product division) choose N2 for Nova Lake? Maybe they had exactly the same concerns as everyone else about whether Intel Foundry could deliver, but knew that TSMC would... ;-)
 
Last edited:
Why did Intel (product division) choose N2 for Nova Lake? Maybe they had exactly the same concerns as everyone else about whether Intel Foundry could deliver, but knew that TSMC would... ;-)
Exactly!

Side note: perhaps we should wait until Fab 52 is loaded to half of planned production capacity before we say 18A is ramped and healthy
 
Exactly!

Side note: perhaps we should wait until Fab 52 is loaded to half of planned production capacity before we say 18A is ramped and healthy
 
Why did Intel (product division) choose N2 for Nova Lake? Maybe they had exactly the same concerns as everyone else about whether Intel Foundry could deliver, but knew that TSMC would... ;-)
It's pretty simple N2 is outright better on PPA according to Intel products and they need that to compete so they went with N2 for high end so the main compute tile went to TSMC
 
But Intel claim 14A has better PPA than N2... ;-)

(yes I know, it's too late for Nova Lake...)
14A is like 27 risk production anyway. I have only heard that Intel rates 18A somewhat near N3P in terms of PPA and 18AP is like close to N2 only in perf/watt but Nova Lake Is N2P
 
14A is like 27 risk production anyway. I have only heard that Intel rates 18A somewhat near N3P in terms of PPA and 18AP is like close to N2 only in perf/watt but Nova Lake Is N2P
Is N2P one of the derivatives of N2 that actually has loosened design rules and is less dense than N2 for manufacturability reasons?

Proof will be in the pudding once we actually see an 18AP product but I think someone here said “the proof is in the pudding that’s being eaten” and that’s N2 😆
 
A twitter TSMC guy said N2.

The mobile SoC companies (Apple, MediaTek, Qcom, etc... ) first use N2. HPC companies (Intel, Nvidia, AMD, Google, etc...) Use N2P. N2X is an optimized version of N2P.

1772812847451.png


 
Back
Top