Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/the-intel-split.15330/
        )

    [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
)

The Intel Split

soAsian

Active member

author is based in Taiwan. his background is not in the semis industry but he does work in tech industry.


my take away: Intel's edge is an technical debt.

Intel's own tools/process used to give them an edge over everyone else during the early years in the semis industry. Overtime, tool suppliers have catch up with Intel and Intel is falling behind due to their own tools delay/insufficient.

Intel's own tools is losing out to commodity and standardized tools
 
Last edited:
author is based in Taiwan. his background is not in the semis industry but he does work in tech industry.

One minor note, Bob Swan signed the TSMC agreement not Pat Gelsinger. Pat may have renegotiated it or more likely "tried to" renegotiate it but that agreement was Bob's doing. As I have stated before I believe Bob did it to get the Intel manufacturing folks in line and to stop the process delays. A double edged sword for sure.
 
Ben Thompson is really good.
"It’s worse than that, actually: by becoming TSMC’s customer Intel is not only denying itself the scale of its own manufacturing needs, but also giving that scale to TSMC, improving the economics of their competitor in the process."
 
Ben Thompson is really good.
"It’s worse than that, actually: by becoming TSMC’s customer Intel is not only denying itself the scale of its own manufacturing needs, but also giving that scale to TSMC, improving the economics of their competitor in the process."
In reality, Intel can't compete in the scale of the economy game anymore, outsource or not. For example, in 2020 there were 275 million PC sold worldwide. That's a great number and a great year for Intel (and to AMD). But Apple alone sold 206 million units of iPhone in 2020!

Don't forget Apple is only a 20% to 25% customer of TSMC's revenue and most in leading edge nodes. Overall TSMC's production volume is probably 3 to 5 times bigger than Intel's, at least. I don't think Intel can find a way to improve this weakness in the near future.

Instead of fighting the losing battle, Bob Swan's strategy was to use TSMC's scale of economy to buy time and opportunity and avoid any more product delay.

My feeling is that Pat Gelsinger is on the same path now. TSMC/Intel partnership is here to stay.
 
Last edited:
In reality, Intel can't compete in the scale of the economy game anymore, outsource or not. For example, in 2020 there were 275 million PC sold worldwide. That's a great number and a great year for Intel (and to AMD). But Apple alone sold 206 million units of iPhone in 2020!
Aren't you overlooking servers; e.g. those things that make Alexa & co. tick ?
 
From listening to the introduction by CEO Peter Wennink of ASML at this weeks Q4/2021 presentation, it seems to me that the demand for pieces of the semiconductor pie will be so large the coming 10 years (growing to 1 trillion / year) that there is enormous need to grow the pie = for added semiconductor manufacturing capacity.

In that respect I get the feeling that Intel is buying part of that growing capacity at TSMC for their own high-end products, while they are also expanding their own manufacturing capacity, that will need to focus on making stuff for others as well. Otherwise they would not get those huge subsidies from the tax payers (US government and US states) to build the fabs. Intel will need to make chips for others as well in order to keep eg the EV automobile industry and many other parts of manufacturing society running.

The world will need more than 2 major foundries (TSMC, Samsung) located in Asia and will grow to 3 or 4, also in view of the geo-political situation. Intel wants to become that number 3. It will be interesting to see if Europe will be able to get its own major foundry going, or that it will stay dependent on the big (future 3) foundries located in Asia and US.

Europe has IMEC's semi industry R&D in Leuven, Belgium, with nearby ASML in Veldhoven. Would be interesting to see if a big foundry can be build in e.g. Austria where other important new developments take place headed by Infineon.

And for all these foundries there is just a single supplier of lithography that runs the show in these big fabs, ASML. Having the world monopoly with gross margin currently around 53%. ASML wants to grow to 55% in 2025, and it may aim to get even closer to the margins of the chip designers like NVIDIA. As long as they keep the whole world happy providing the lithography tools that are needed for that growing demand of pie. Future will tell.
 
Aren't you overlooking servers; e.g. those things that make Alexa & co. tick ?
In terms of units for server sold worldwide, that's much smaller. The worldwide server market achieved 12.5 million units in 2020. It's a good number but it's not a "huge" number comparing to many other semiconductor products or TSMC's output.
Scale of economy needs volume. Intel knows this is their weakness and to offer foundry service may help them. Although it may lead to more capacity utilization issues.


 

author is based in Taiwan. his background is not in the semis industry but he does work in tech industry.


my take away: Intel's edge is an technical debt.

Intel's own tools/process used to give them an edge over everyone else during the early years in the semis industry. Overtime, tool suppliers have catch up with Intel and Intel is falling behind due to their own tools delay/insufficient.

Intel's own tools is losing out to commodity and standardized tools
I don't understand this, Intel buys all their tools just like everyone else, they don't make their own tools.
 
In reality, Intel can't compete in the scale of the economy game anymore, outsource or not. For example, in 2020 there were 275 million PC sold worldwide. That's a great number and a great year for Intel (and to AMD). But Apple alone sold 206 million units of iPhone in 2020!

Don't forget Apple is only a 20% to 25% customer of TSMC's revenue and most in leading edge nodes. Overall TSMC's production volume is probably 3 to 5 times bigger than Intel's, at least. I don't think Intel can find a way to improve this weakness in the near future.

Instead of fighting the losing battle, Bob Swan's strategy was to use TSMC's scale of economy to buy time and opportunity and avoid any more product delay.

My feeling is that Pat Gelsinger is on the same path now. TSMC/Intel partnership is here to stay.
TSMC makes more wafers than Intel but at the leading edge their volumes are a lot closer than you think. Intel has enough volume, they just need to execute a lot better.
 
Nvidia is designing their own cpu. Cloud companies use their own designs. Intel is being decimated from everywhere.
 
Back
Top