Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/gelsinger-to-beat-tsm-in-two-years.19379/
        )

    [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] => 2021370
            [XFI] => 1050270
        )

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

Gelsinger to beat TSM in two years

Arthur Hanson

Well-known member
In Barron's, Gelsinger feels Intel can beat TSM in two years. I have my doubts about this, unless he is hoping China invades Taiwan. Any thoughts or comments on this appreciated. I feel TSM's relationships with AMD and Nvidia among others are key.

 
Last edited:
TSMC is the technology leader and the highest volume foundry. Intel Is trying to take the technology lead and wants to add foundry customers. Aspirations are good.

Intel needs to deliver products to end customers on 18A, then needs to add external foundry customers so that their external revenue is higher than SMIC, GF, UMC. We presented the forecast for this and.... IMHO..... it will take 3+ years. Until then, Intel will continue to buy 5x more silicon from TSMC than it sells to external foundry customers.

milestone 1: is intel selling 18A as a final product to end customers? Milestone 2: when will Intel have 25K wafers per month of external fab customers?

Intel has documented these milestone dates.... we will see if they hit them
 
TSMC is the technology leader and the highest volume foundry. Intel Is trying to take the technology lead and wants to add foundry customers. Aspirations are good.

Intel needs to deliver products to end customers on 18A, then needs to add external foundry customers so that their external revenue is higher than SMIC, GF, UMC. We presented the forecast for this and.... IMHO..... it will take 3+ years. Until then, Intel will continue to buy 5x more silicon from TSMC than it sells to external foundry customers.

milestone 1: is intel selling 18A as a final product to end customers? Milestone 2: when will Intel have 25K wafers per month of external fab customers?

Intel has documented these milestone dates.... we will see if they hit them

I agree. One whale foundry customer might break the ice and get business rolling along.
 
correct me if I'm wrong but TSMC don't talk about their customers.

Intel bad mouth Nvidia. I don't know if that's going to win them a customer or not.
 
correct me if I'm wrong but TSMC don't talk about their customers.
Intel bad mouth Nvidia. I don't know if that's going to win them a customer or not.

100% true but TSMC (CC Wei) does respond to critics now, Intel included. Maybe Intel will not be a TSMC customer for long?
 
I really wonder what’s next in 2026-2028. Intel has unquestionably lost the battle for 3nm. 20A/18A are likely to be very impressive but at most they’ll only be the beginning of Intel taking/holding the crown. The process after 18A is probably a lot more decisive for whale hunting.
 
I really wonder what’s next in 2026-2028. Intel has unquestionably lost the battle for 3nm. 20A/18A are likely to be very impressive but at most they’ll only be the beginning of Intel taking/holding the crown. The process after 18A is probably a lot more decisive for whale hunting.
I get the vibe that most of the fabless industry is of the same mind. Chip designers and IP guys also seem like they enjoy talking about their work/gossiping. Based on this, my understanding of the industry, and my understanding of the mindset of chip designers; I would guess that most fabless firms will be watching 18A from the sidelines. IMO 18A is a make or break node for IFS; but in a building trust and credibility as foundry to rival TSMC kind of way.

correct me if I'm wrong but TSMC don't talk about their customers.
You would be correct. It is not general foundry practice to talk about their customers or their products. Even for painfully obvious things like Apple and TSMC. As far as I know TSMC had never said anything resembling "...our lead customer Apple will ramp N3 in 2023". TSMC always talked in terms like "our lead N3 product". Popping the Apple using *insert node name here* bottle was reserved for Apple to do on their own schedule.
 
Did Intel ever announce their whale? or primary customers? or prepayment amount? Don't need a name..... but the rumor is that it is not a firm commitment of anything (normally you have take or pay clauses and required minimum wafers or percentage of wafers). It is rumored to be a MOU with refundable prepayment. Didn't they say there would be more info before the end of the year?
 
Did Intel ever announce their whale? or primary customers? or prepayment amount? Don't need a name..... but the rumor is that it is not a firm commitment of anything (normally you have take or pay clauses and required minimum wafers or percentage of wafers). It is rumored to be a MOU with refundable prepayment. Didn't they say there would be more info before the end of the year?

I remember Intel did say it will announce a major IFS customer win by the end of 2023. I think it's better for Intel to follow the typical foundry industry practice by not talking about its custom names at all.
 
I really wonder what’s next in 2026-2028. Intel has unquestionably lost the battle for 3nm. 20A/18A are likely to be very impressive but at most they’ll only be the beginning of Intel taking/holding the crown. The process after 18A is probably a lot more decisive for whale hunting.

18A is an important step in gaining customer trust. Foundries must deliver on their word, you than thank TSMC for that and their "Trusted Foundry" sound byte. It really does play. Samasung in on the other end of that spectrum. Hopefully Intel will be up next to TSMC rather than next to Samsung.
 
Back
Top