Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/will-intel-have-enough-euv-for-7nm.14191/
        )

    [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
)

Will Intel have enough EUV for 7nm?

Daniel Nenni

Admin
Staff member
It sure does not seem like it. Big win for TSMC 3N?

ASML EUV Shipments.jpg

Logic Advanced Node Production Map TSMC Samsung Intel 2021.jpg


 
Under ASML 2012 co-investor program, Intel invested $4.1 billion, TSMC invested about $1.4 billion, and Samsung invested just less than $1 billion in ASML to accelerate EUV development.

Why Intel invested the largest amount of money among the big three semiconductor companies and had all the insider information yet ended up to buy the least number of EUV?

It's clearly a leadership and management failure by Intel themselves.


 
Last edited:
The numbers of EUV deliveries need to be updated post ASML Q1 results. The company has stated, 2 units of EUV were taken back (likely from INTC's fab 28) and will be resold to someone else. besides, the total delivery in 2021 will increase to 55 units, and will grow a bit more afterwards.
 
TSM estimated 3nm capacity in 2023 is eye popping

Agreed. It incudes Intel using TSMC N3, which I think will be bigger than expected given the delays in Intel 7nm and the lack of EUV systems. I really don't see how Intel comes back from outsourcing to TSMC. AMD is winning and Intel 7nm is no longer competitive. Samsung 3nm is having a rough start. TSMC winning again.
 
Agreed. It incudes Intel using TSMC N3, which I think will be bigger than expected given the delays in Intel 7nm and the lack of EUV systems. I really don't see how Intel comes back from outsourcing to TSMC. AMD is winning and Intel 7nm is no longer competitive. Samsung 3nm is having a rough start. TSMC winning again.
exactly. Intel won't have enought EUV for all product lines in 2024. Intel needs TSMC 3nm to protect it's market share for server, desktop and laptop.
 
Last edited:
At IEDM last December Intel said EUV did not arrive in time for 10nm so they had to reengineer it for DUV. That was the excuse why 10nm was late. I couldn't believe it when I heard it. I actually had to wait for the replay to listen again.
 
At IEDM last December Intel said EUV did not arrive in time for 10nm so they had to reengineer it for DUV. That was the excuse why 10nm was late. I couldn't believe it when I heard it. I actually had to wait for the replay to listen again.
Asian companies such as Samsung and SMIC . Tend to be more willing to place risk order for CAPEX even when they are not ready yet.
US companies are more cautious.
It could be Intel committed EUV too late to ASML and need to wait for a long time.
 
Asian companies such as Samsung and SMIC . Tend to be more willing to place risk order for CAPEX even when they are not ready yet.
US companies are more cautious.
It could be Intel committed EUV too late to ASML and need to wait for a long time.

That is exactly what happened. Bad supply chain management. I was also told that TSMC and ASML fully collaborate. Intel and ASML do not fully collaborate.
 
Intel has a new development center in Israel and lithography companies are expanding their offerings. It's only a matter of time until there's a competitor to asml. Israel likes to include a lot of ai in their manufacturing equipment something asml doesn't really do.
 
Intel has a new development center in Israel and lithography companies are expanding their offerings. It's only a matter of time until there's a competitor to asml. Israel likes to include a lot of ai in their manufacturing equipment something asml doesn't really do.
It may take 10+years even if it works.
 
It may take 10+years even if it works.
If the new development center philosophy works it can be done in less than 3 years. I admit I'm writing a science fiction story. Intel employees 50k people indirectly in Israel and they're probably not willing to let that business leave.
 
  • Like
Reactions: VCT
At IEDM last December Intel said EUV did not arrive in time for 10nm so they had to reengineer it for DUV. That was the excuse why 10nm was late. I couldn't believe it when I heard it. I actually had to wait for the replay to listen again.
That reminded me to look at their paper. It seemed to be self-conflicting, saying EUV was good to avoid multipatterning, but they still need multipatterning with EUV in the end. It was not very clear, but this was not a full SPIE paper. It did present the first snapshot of the 36 nm pitch M1:
Intel 36 nm pitch M1.jpg
To be fair, imaging the gaps between the line ends in a single exposure would not look so good, particularly with the adjacent line end gaps just slightly shifted.
 
At IEDM last December Intel said EUV did not arrive in time for 10nm so they had to reengineer it for DUV. That was the excuse why 10nm was late. I couldn't believe it when I heard it. I actually had to wait for the replay to listen again.
I guess one of the thing Intel won't say is that EUV machines and the R&D resources associated with them are very expensive. In a short term it won't help Intel's profit margin and stock price. Intel would rather spend money on stock buy back program than on those expensive EUV machines.
 
Last edited:
That is exactly what happened. Bad supply chain management. I was also told that TSMC and ASML fully collaborate. Intel and ASML do not fully collaborate.
Another problem with Intel's monopolistic culture. Intel wants to control it's suppliers, not collaborate with them. In the past they had that kind of power, today AMSL is actually the larger company by market cap and Intel is not even close to it's most important customer.

So Intel goes to AMSL and drags them through onerous negotiations trying to strong arm them, but AMSL is the MUCH stronger party in these types of negotiations and essentially says "we'll just sell these machines to TSMC instead, they are the better partner to us".
 
Another problem with Intel's monopolistic culture. Intel wants to control it's suppliers, not collaborate with them. In the past they had that kind of power, today AMSL is actually the larger company by market cap and Intel is not even close to it's most important customer.

So Intel goes to AMSL and drags them through onerous negotiations trying to strong arm them, but AMSL is the MUCH stronger party in these types of negotiations and essentially says "we'll just sell these machines to TSMC instead, they are the better partner to us".

100% true. Intel does not collaborate like TSMC does. Bullying no longer works inside the semiconductor ecosystem thanks in great part to TSMC. That's why the frenemy or co-opetition strategies no longer work. Hopefully Pat Gelsinger thoroughly understands this before it is too late, absolutely.
 
Back
Top