Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/could-samsung-ever-be-serious-competition-to-tsm.20821/
        )

    [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
)

Could Samsung ever be serious competition to TSM?

Arthur Hanson

Well-known member
As the number two in the fab business, is there any way Samsung could come close to or maybe even pass TSM or is it game over and TSM will dominate for at least the next decade? Any chance TSM will become a major player in high band width memory and pose a threat to one of Samsungs business lines? Any thoughts or comments appreciated.
 
Samsung has made several attempts to surpass TSMC by taking shortcuts, such as being the first to adopt EUV for 7nm and leading with GAA technology. However, these efforts have not been successful. Looking ahead to 2030, the only significant breakthrough on the horizon is backside power rail/delivery, and it appears that Samsung is not leading in its adoption.

TSMC's current position can be attributed to various factors, but one key element is their strong risk management. Every year, TSMC seems to implement multiple strategies to ensure the successful launch of the iPhone, often building in significant redundancy in R&D. This approach instills strong confidence in their customers, who tend to prefer consistent 10-15% efficiency improvements annually over risky, leapfrog advancements.

As for HBM, I don’t believe TSMC will be a major player in the near future. TSMC has formed a strategic partnership with Hynix, and when CC Wei introduced Foundry 2.0, he specifically mentioned that it does not include memory.
 
Samsung has made several attempts to surpass TSMC by taking shortcuts, such as being the first to adopt EUV for 7nm and leading with GAA technology. However, these efforts have not been successful. Looking ahead to 2030, the only significant breakthrough on the horizon is backside power rail/delivery, and it appears that Samsung is not leading in its adoption.

TSMC's current position can be attributed to various factors, but one key element is their strong risk management. Every year, TSMC seems to implement multiple strategies to ensure the successful launch of the iPhone, often building in significant redundancy in R&D. This approach instills strong confidence in their customers, who tend to prefer consistent 10-15% efficiency improvements annually over risky, leapfrog advancements.

As for HBM, I don’t believe TSMC will be a major player in the near future. TSMC has formed a strategic partnership with Hynix, and when CC Wei introduced Foundry 2.0, he specifically mentioned that it does not include memory.
na, it's all about resources. SK and tw both are small regions with 30 m ppls. it can't support too many R&D projects. samsung already put most of it's resources in memory. tw also failed in memory. so fair for both side.
 
na, it's all about resources. SK and tw both are small regions with 30 m ppls. it can't support too many R&D projects. samsung already put most of it's resources in memory. tw also failed in memory. so fair for both side.
Does the same issue wrt to lack of resources apply to the US and its inability to produce competitive logic technology? The US has more people but the median quality is much lower.
 
Does the same issue wrt to lack of resources apply to the US and its inability to produce competitive logic technology? The US has more people but the median quality is much lower.
Maybe it has more to do with BOD capability.

Mark Liu once was a manager in a US fab, and never got a chance to enter the vp level. He quit and became the ceo of tsmc later. Did you learn anything here? something wrong with the US system.
https://www.crunchbase.com/person/mark-liu-2

if they had kept Liu and promoted him to be the ceo. there is no tsmc competition now. tsmc could just be a 2nd tier company poses no threat, like GF.
 
As the number two in the fab business, is there any way Samsung could come close to or maybe even pass TSM or is it game over and TSM will dominate for at least the next decade? Any chance TSM will become a major player in high band width memory and pose a threat to one of Samsungs business lines? Any thoughts or comments appreciated.
Samsung has been trying how long and failed, nope not going to happen, but hope is a dangerous strategy.
 
Does the same issue wrt to lack of resources apply to the US and its inability to produce competitive logic technology? The US has more people but the median quality is much lower.
not a lack of talent, resource, or ability to work hard! It was and is a failure of leadership.

Had Paul said yes to Steve and made it a priority TSMC wouldn’t have leadership, but that x86 myopic focus from the BoD down to CEO is the issue, actually it is still an issue with IDM2.0
 
not a lack of talent, resource, or ability to work hard! It was and is a failure of leadership.

Had Paul said yes to Steve and made it a priority TSMC wouldn’t have leadership, but that x86 myopic focus from the BoD down to CEO is the issue, actually it is still an issue with IDM2.0
It's very difficult for bean counters to understand the semiconductor industry, or impossible. The trend is that founders would train their sons or grandsons to inherit the business and pass down the know-how, as seen with companies like Qcom and Nvidia. Unfortunately, this didn't happen for INTC.
 
It's very difficult for bean counters to understand the semiconductor industry, or impossible. The trend is that founders would train their sons or grandsons to inherit the business and pass down the know-how, as seen with companies like Qcom and Nvidia. Unfortunately, this didn't happen for INTC.
I think after Grove it has been a downhill
 
not a lack of talent, resource, or ability to work hard! It was and is a failure of leadership.

Had Paul said yes to Steve and made it a priority TSMC wouldn’t have leadership, but that x86 myopic focus from the BoD down to CEO is the issue, actually it is still an issue with IDM2.0
This is not correct. If Paul had said yes, Intel would have been bankrupt.... Intel was not and is not good at what Apple wanted. Intel needs to focus on what they do best and avoid what they are not good at. the Kansas City Chiefs cannot win a NBA championship no matter how hard they try and how good their leadership is. I am painfully aware of exactly what Intel could and could not do in Mobile and with Apple at that time.

Side note: you do realize that Apple chose Samsung to manufacture their chips .... not TSMC until later?

Which brings us back to the Original question. Samsung can compete with TSMC IF they execute and wait for opportunity.
 
Last edited:
This is not correct. If Paul had said yes, Intel would have been bankrupt.... Intel was not and is not good at what Apple wanted. Intel needs to focus on what they do best and avoid what they are not good at. the Kansas City Chiefs cannot win a NBA championship no matter how hard they try and how good their leadership is. I am painfully aware of exactly what Intel could and could not do in Mobile and with Apple at that time.

Side note: you do realize that Apple chose Samsung to manufacture their chips .... not TSMC until later?

Which brings us back to the Original question. Samsung can compete with TSMC IF they execute and wait for opportunity.
As someone who has never been on the inside you sure seem to be the ideal armchair quarterback.

At that time Intel had tons of money as technology lead. It could easily have afforded to cater to Apple. If they put as much into that customer as they did to their internal than who knows. Of course if they delivered like Samsung which was a possible outcome than of course no difference. Samsung logic focus and culture is poisoned by their memory culture a leadership. You could also argue that Intel is equally poisoned by their multi decade x86 captive market and relationship with their product group.

The challenge is can the zebras transform their stripes, a very hard transformation far deeper than the catchy words from leadership of 5N4Y and IDM2.0 blah blah blah.

Time for Intel Foundry to build trust with all the people they are dancing by committing and delivering what they commit and having transparency.
 
Last edited:
This is not correct. If Paul had said yes, Intel would have been bankrupt.... Intel was not and is not good at what Apple wanted. Intel needs to focus on what they do best and avoid what they are not good at. the Kansas City Chiefs cannot win a NBA championship no matter how hard they try and how good their leadership is. I am painfully aware of exactly what Intel could and could not do in Mobile and with Apple at that time.

Side note: you do realize that Apple chose Samsung to manufacture their chips .... not TSMC until later?

Which brings us back to the Original question. Samsung can compete with TSMC IF they execute and wait for opportunity.
To make it clear, Intel did not say they wouldn't produce Apple chips. They said they would do it -- but only at a price higher than Apple would pay. Intel later said they underestimated Apple's volumes -- and so their assumptions about learning costs down over time was wrong.

Apple was a really tough customer then. (and probably they still are). I know one company that lost Apple business because they didn't want to bid on the next generation product.
 
To make it clear, Intel did not say they wouldn't produce Apple chips. They said they would do it -- but only at a price higher than Apple would pay. Intel later said they underestimated Apple's volumes -- and so their assumptions about learning costs down over time was wrong.

Apple was a really tough customer then. (and probably they still are). I know one company that lost Apple business because they didn't want to bid on the next generation product.
100% correct on pricing. But the only way it would have broken even was with unrealistic cost reductions. It was a good call.

I have consulted for two companies that FIRED apple as a customer due to massive losses. LOL
 
I think after Grove it has been a downhill
Craig Barrett wasn't as good for Intel as Andy Grove, but he was at least in charge when Intel bought the ARM stuff to create XScale. (Intel sold it in 2006 under Otellini - another shortsighted mistake imo). Craig also had to greenlight the Core 2 and probably Pentium M programs which really helped cement Intel for a decade+. Engineering in general seemed to remain solid under Craig's leadership, and Intel did serious public branding campaigns under him.

Also Itanium started development under Andy :).
 
Craig Barrett wasn't as good for Intel as Andy Grove, but he was at least in charge when Intel bought the ARM stuff to create XScale. (Intel sold it in 2006 under Otellini - another shortsighted mistake imo). Craig also had to greenlight the Core 2 and probably Pentium M programs which really helped cement Intel for a decade+. Engineering in general seemed to remain solid under Craig's leadership, and Intel did serious public branding campaigns under him.

Also Itanium started development under Andy :).
If i remember pat hated Itanium he said x86 is good yeah i forgot him the guy who created copy exactly after that it's been disastrous decision
 
If i remember pat hated Itanium he said x86 is good yeah i forgot him the guy who created copy exactly after that it's been disastrous decision
CE! was Craig, but it is a good thing not a bad thing. Global manufacturing doesn't work without CE and it is a staple tool across the industry when transferring in already in production technologies to a new fab. Also I think people fundamentally don't really understand what that the point of CE! is structured change control and variability reduction (a prerequisite for statistical process control), and not just copying exactly and never changing anything. I don't know if CB was a Edwards Deming fan, but alot of the work CB did to re-invent intel quality and then reinvigorating intel's manufacturing align very well with Deming's 14 points.
 
Last edited:
Back
Top