Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/samsung-foundry-strategy-meeting-key-finding-the-reason-google-missed-out.23063/
        )

    [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] => 2021770
            [XFI] => 1050270
        )

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

Samsung Foundry, Strategy Meeting Key 'Finding the Reason Google Missed Out'

Daniel Nenni

Admin
Staff member
Deals Lost to TSMC, Focused Analysis on Galaxy, 3 Nano, and IP Issues

Samsung Electronics is in the midst of an inspection to restore its competitiveness in semiconductor foundry. Various discussions are taking place to solve fundamental problems. In particular, it is said that they are intensively analyzing the recent case in which Google transferred the mass production of application processors (APs) that had been entrusted to Samsung Electronics to TSMC.

According to industry sources on the 18th, Samsung Electronics' Device Solutions (DS) division is holding a global strategy meeting chaired by Vice Chairman Jeon Young-hyun. It is reported that strengthening foundry capabilities is included among the various agenda items.

The foundry division has been struggling for a long time. As competition for advanced processes below 5 nanometers (nm) intensified, it was expected to form a two-way system with TSMC, but the gap is widening.

According to market research firm TrendForce's compilation of global foundry market shares in the first quarter of this year, Samsung Electronics' share decreased to 7.7% compared to the fourth quarter of last year (8.1%). During the same period, TSMC increased from 67.1% to 67.6%. Samsung Electronics is now in a position where it has to worry about the pursuit of Chinese companies such as SMIC (6.0%) and UMC (4.7%).

In terms of performance, deficits are continuing. Accordingly, Samsung Global Research has begun management diagnosis of the foundry division following the system LSI division. The goal is to present a direction by analyzing overseas cases and strategic success factors.

20250618150403216.jpg


*Google AP 'Tensor'

The 'Google incident' that shocked Samsung Electronics and the world last month is also a subject of analysis. Samsung Electronics and Google have been collaborating in the AP sector for a long time. Google used Qualcomm APs in its smartphones in the past, but developed its own AP 'Tensor' through the system-on-chip (SoC) project 'Whitechapel' with Samsung Electronics for several years. It is known that the chip was modeled after Samsung Electronics' AP 'Exynos'.

Since then, Samsung Electronics has been in charge of mass production of Google's Pixel series APs. They have continued to collaborate over the generations, and the alliance between the two companies has continued even in processes below 5nm. The problem is that Samsung Electronics' yield (the ratio of good products among finished products) has decreased as it approaches 3nm.

In addition, the lack of semiconductor design assets (IP) also played a role. This means that Samsung Electronics' software (SW) resources were not enough in a situation where Google's desired performance and functions were diversifying.

In addition, the fact that Google is releasing foldable smartphones in addition to bar-type smartphones and the competition with Samsung Electronics' mobile business is intensifying seems to have had some influence. This is similar to the reason Apple shifted its AP manufacturing from Samsung Electronics to TSMC in the past.

After going through a series of processes, Google has entrusted the AP of the Pixel 10, scheduled to be released in the second half of this year, to TSMC. This product is made through TSMC's 3nm process. The industry predicts that TSMC will maintain its exclusive position until the Pixel 14.

A semiconductor industry insider explained, "Losing Google is a case that shows the complex problems of Samsung Foundry at once," and "I understand that there are many discussions and concerns going on internally as well."

20250618150433176.jpg


*Samsung US Foundry Factory

Samsung Electronics has been having trouble attracting advanced process customers since it gave up Qualcomm, Nvidia, and others to TSMC. The gap in orders is even bigger as the System LSI Division, which was a reliable cornerstone of the Foundry Division, is going through a difficult time. As a result, it is understood that the recovery of the operating rate of the leading line is being delayed.

Samsung Electronics plans to establish a contingency plan to lead the foundry recovery through successive management diagnoses and global strategy meetings. The spin-off of the foundry and the transfer of part of the System LSI Division are being discussed.

It is also working to enhance its core competitiveness by strengthening collaboration with partners such as Synopsys. It is also pushing to expand the application of advanced processes beyond artificial intelligence (AI) to automobiles, robots, etc.

It is also struggling to proceed as planned with the 2nm process scheduled for mass production in the second half of this year. Since there is a lack of 3nm references, the first meaningful 2nm customer is likely to be the System LSI Division. The AP 'Exynos 2600', which is being developed targeting the 'Galaxy S26' series early next year, is a representative example. The analysis is that if the product shows good performance, it can target big tech companies that are highly dependent on TSMC.

 
The myopia of the corporate world never ceases to amaze me. I believe Samsung's biggest problem is as plain as the nose on Vice Chairman Jeon Young-hyun's face. Their yield rates are rubbish. If they had reasonable yields, they would have not been hemorrhaging customers at this rate. TSMC is a tough competitor and a very good foundry, so they may still lose a deal here and there, but I have to believe the not-TSMC crowd would continue to throw them enough bones to keep them in the game.

Similarly Intel is going to have to prove to potential customers that their process yields well, their progress on process development stays on track and that they can deliver orders on time. Delivering specific orders on time is very different from delivering a handful of products against a quarterly projection.
 
Deals Lost to TSMC, Focused Analysis on Galaxy, 3 Nano, and IP Issues

Samsung Electronics is in the midst of an inspection to restore its competitiveness in semiconductor foundry. Various discussions are taking place to solve fundamental problems. In particular, it is said that they are intensively analyzing the recent case in which Google transferred the mass production of application processors (APs) that had been entrusted to Samsung Electronics to TSMC.

According to industry sources on the 18th, Samsung Electronics' Device Solutions (DS) division is holding a global strategy meeting chaired by Vice Chairman Jeon Young-hyun. It is reported that strengthening foundry capabilities is included among the various agenda items.

The foundry division has been struggling for a long time. As competition for advanced processes below 5 nanometers (nm) intensified, it was expected to form a two-way system with TSMC, but the gap is widening.

According to market research firm TrendForce's compilation of global foundry market shares in the first quarter of this year, Samsung Electronics' share decreased to 7.7% compared to the fourth quarter of last year (8.1%). During the same period, TSMC increased from 67.1% to 67.6%. Samsung Electronics is now in a position where it has to worry about the pursuit of Chinese companies such as SMIC (6.0%) and UMC (4.7%).

In terms of performance, deficits are continuing. Accordingly, Samsung Global Research has begun management diagnosis of the foundry division following the system LSI division. The goal is to present a direction by analyzing overseas cases and strategic success factors.

20250618150403216.jpg


*Google AP 'Tensor'

The 'Google incident' that shocked Samsung Electronics and the world last month is also a subject of analysis. Samsung Electronics and Google have been collaborating in the AP sector for a long time. Google used Qualcomm APs in its smartphones in the past, but developed its own AP 'Tensor' through the system-on-chip (SoC) project 'Whitechapel' with Samsung Electronics for several years. It is known that the chip was modeled after Samsung Electronics' AP 'Exynos'.

Since then, Samsung Electronics has been in charge of mass production of Google's Pixel series APs. They have continued to collaborate over the generations, and the alliance between the two companies has continued even in processes below 5nm. The problem is that Samsung Electronics' yield (the ratio of good products among finished products) has decreased as it approaches 3nm.

In addition, the lack of semiconductor design assets (IP) also played a role. This means that Samsung Electronics' software (SW) resources were not enough in a situation where Google's desired performance and functions were diversifying.

In addition, the fact that Google is releasing foldable smartphones in addition to bar-type smartphones and the competition with Samsung Electronics' mobile business is intensifying seems to have had some influence. This is similar to the reason Apple shifted its AP manufacturing from Samsung Electronics to TSMC in the past.

After going through a series of processes, Google has entrusted the AP of the Pixel 10, scheduled to be released in the second half of this year, to TSMC. This product is made through TSMC's 3nm process. The industry predicts that TSMC will maintain its exclusive position until the Pixel 14.

A semiconductor industry insider explained, "Losing Google is a case that shows the complex problems of Samsung Foundry at once," and "I understand that there are many discussions and concerns going on internally as well."

20250618150433176.jpg


*Samsung US Foundry Factory

Samsung Electronics has been having trouble attracting advanced process customers since it gave up Qualcomm, Nvidia, and others to TSMC. The gap in orders is even bigger as the System LSI Division, which was a reliable cornerstone of the Foundry Division, is going through a difficult time. As a result, it is understood that the recovery of the operating rate of the leading line is being delayed.

Samsung Electronics plans to establish a contingency plan to lead the foundry recovery through successive management diagnoses and global strategy meetings. The spin-off of the foundry and the transfer of part of the System LSI Division are being discussed.

It is also working to enhance its core competitiveness by strengthening collaboration with partners such as Synopsys. It is also pushing to expand the application of advanced processes beyond artificial intelligence (AI) to automobiles, robots, etc.

It is also struggling to proceed as planned with the 2nm process scheduled for mass production in the second half of this year. Since there is a lack of 3nm references, the first meaningful 2nm customer is likely to be the System LSI Division. The AP 'Exynos 2600', which is being developed targeting the 'Galaxy S26' series early next year, is a representative example. The analysis is that if the product shows good performance, it can target big tech companies that are highly dependent on TSMC.


"In particular, it is said that they are intensively analyzing the recent case in which Google transferred the mass production of application processors (APs) that had been entrusted to Samsung Electronics to TSMC."


I thought everyone already knew that Google has been moving to TSMC for its Tensor mobile processors since late 2023 or 2024. Shouldn’t Samsung’s analysis have been done last year?
 
I just realized it's missing scale and labels on Y axis what's with the foundry and missing scale and y axis even on their D0 the scale is missing 🤣
 
Back
Top