President Wei Zhejia (CC Wei) shows confidence: "There is no way to compete with TSMC"

[Daniel Nenni]

Emphasizing customer trust is the key to success

TSMC President Wei Zhejia (right) was awarded an "Honorary Doctorate" by President Lin Qihong of National Yang-Ming Chiao Tung University. Reporter Pan Junhong/Photography

Economic Daily reporter Zhong Huiling/Reporting from Taipei

TSMC President Wei Zhejia pointed out that TSMC does not compete with customers and win their trust. Perhaps opponents will catch up with the process and technology, but TSMC’s pure wafer foundry model gaining the trust of customers is the key to victory. It is located in South Korea Competitors in California all have products. If they want to compete with TSMC, there is simply no way.

The outside world believes that the opponents hinted by Wei Zhejia are Samsung and Intel.

Yang-Ming Jiaotong University yesterday awarded Wei Zhejia a dual degree of honorary doctorate in engineering and honorary doctorate in science. TSMC founder Zhang Zhongmou and Hon Hai chairman Liu Yangwei were both present to congratulate him. Wei Zhejia said humorously that he felt indescribably happy to get his degree without having to pass the exam this time.

In his speech, Wei Zhejia said that TSMC founder Zhang Zhongmou pioneered the pure wafer foundry model, which allowed many IC product companies to spring up. Under Zhang Zhongmou’s active leadership, TSMC gradually reached the number one position in the world.

Wei Zhejia emphasized that "customer trust" is very important. In terms of technology and manufacturing, opponents may one day catch up with TSMC or be as good as TSMC, although he thinks it is unlikely. However, customer trust is the key to victory, and both powerful opponents have it. If you want to compete with TSMC with your own products, simply put, there is "no way". Wei Zhejia said that TSMC’s values are integrity, commitment, innovation, and customer trust. TSMC will spare no effort to enable customers to succeed in the market.

Citing the work of management expert Jim Collins, Wei Zhejia pointed out that a good company will fail. The first step is for the CEO to "forget who I am", unable to listen to other people's opinions, and start to expand crazily regardless of whether there is enough money, so he told You must "never forget who I am" and "know who I am".

Looking back on those years, Wei Zhejia said that in fact, his first choice was to be a doctor, but he was afraid to see blood. "People need to know what they can do and what they cannot do." His second choice was to evaluate buildings, but he felt that he had no talent. I had no choice but to choose to read the motor. Through self-observation, Wei Zhejia found out where his talents lie, and eventually obtained a PhD in electrical engineering from Yale University in the United States. "When I was studying for a PhD, I knew that my brain was not that strong, but I liked engineering very much, so I went to the industry." Wei Zhejia pointed out that 40 years ago, he did not know that semiconductors would be so important. "I happened to be in this industry and was seen by Morris (Zhang Zhongmou) at the right moment. I was very lucky."

Wei Zhejia said that he learned to listen from Zhang Zhongmou, but so far he has not been able to obtain 100% true information. When Zhang Zhongmou chatted with him, he never interrupted him, but he occasionally interrupted when Zhang Zhongmou was talking.

總裁展現信心 魏哲家：與台積競爭「門都沒有」 | 產業熱點 | 產業 | 經濟日報

台積電總裁魏哲家昨（29）日指出，台積電不與客戶競爭、贏得客戶信任，或許對手會追上製程與技術，但台積電純晶圓代工模式獲得...

money.udn.com

 

[blueone]

I remember when Intel was this arrogant. AT&T too. And IBM. Technology evolves faster now. Apparently Wei needs to reread his copy of Only The Paranoid Survive. All that's missing is a quote saying something like "Intel and Samsung will never be in our windshield".

 

[Daniel Nenni]

Just a reminder, the foundry business started with excess capacity from IDMs. We wrote about this in our book Fables: The Transformation of the Semiconductor Industry. The IDMs of course competed with the fabless companies so it did not end well. Intel and Samsung have virtually spun out the foundry business but they still compete and no matter what they say internal design will always have psychological priority over foundry customers because that is where the big margins are, absolutely. In the US we call this "Fenemies" (friends/enemies) or "coopetition" (cooperating but competing) but the foundation is competition. With TSMC the foundation is trust.

 

[blueone]

With TSMC the foundation is trust.

Having worked in groups that used TSMC as a foundry, I'm not on that page. TSMC was chosen for ease of development (available IP, PDKs, dependable shuttles, customer support), and, of course, execution. These are more easily replicable value propositions than developing fab processes and building and operating fabs. Intel and Samsung may not succeed against TSMC (personally, I think their chances are less than 50%), but we disagree on the importance of the IP trust factor in getting business.

 

[Daniel Nenni]

Having worked in groups that used TSMC as a foundry, I'm not on that page. TSMC was chosen for ease of development (available IP, PDKs, dependable shuttles, customer support), and, of course, execution. These are more easily replicable value propositions than developing fab processes and building and operating fabs. Intel and Samsung may not succeed against TSMC (personally, I think their chances are less than 50%), but we disagree on the importance of the IP trust factor in getting business.

It's not IP trust. This is an example of the trust:

"TSMC was chosen for ease of development (available IP, PDKs, dependable shuttles, customer support), and, of course, execution."

Companies are trusting foundries with their products. That requires a big amount of trust.

 

[cxy11]

https://www.investorschronicle.co.uk/news/2024/03/05/what-tsmc-s-new-plant-tells-us-about-intel-and-the-us-s-ambitions/

 

[blueone]

Companies are trusting foundries with their products. That requires a big amount of trust.

On this point we agree, especially about execution. It occurred to me after I posted that you could be referring to things like production volume competition with internal products, but to be honest it makes little business sense, at least for Intel. I can't imagine the scenario where Intel puts a bottleneck in IFS customer chip production because Intel chips are in short supply, because perhaps leading edge process capacity has greater demand than can be met. Intel is spending many tens of billions of dollars on IFS fab capacity, and it is incomprehensible that Gelsinger (or his successor) would put that at risk to increase quarterly Intel CPU revenue or total margins. All it would take is one major company to walk away from IFS for that reason and much of that IFS investment would get flushed in short order. Many times Intel's decision-making makes my eyes roll, but that behavior is beyond stupid, more like suicidal.

But we digress. I'll be impressed if Intel can just get shuttles working as well as TSMC does. I'm thinking that capability alone will take a few years to perfect.

 

[TBiggs]

The US is a big testing ground.

I would say TSMC isn’t getting a passing grade on any metric.

Intel may have struggled and lost its way with BK and his minions.

There is no corner on chemicals, tools, or physics, innovation that Taiwan owns that ensure it keeps its technology lead over Korea or US.

For Foundry success what is indeed unique is the ability to deliver accurate PDKs and running NTO on target and keeping a multitude of products that behave differently due to density and design sensitivities on target that requires lots of customization and lastly saying yes to ever customers realistic needs. Here TSMC has no rival, but Intel and Samsung can easily do this as well. There is arrogance thru the TSMC from top to bottom these days, same as Intel and IBM before that is certain.

The world is a crazy place and anything is possible and arrogance is even more dangerous than ever.

The one thing TSMC has currently is scale. When you have Apple, Nvidia, AMD, Qualcomm, Intel, Microsoft, Broadcom and others you just need to listen to your collective customers what they want and of course the scale of all those customers enables scale and learning that is unprecedent, but take Xi and POTUS misscalculation to go gone, or a node delay like Intel did, LOL

 

[hist78]

If we don't count luck as a main factor, the business model will decide a business' success.

Using Intel Foundry (IF) as an example:

Assume Intel product/design division and Intel Foundry both want to maintain a 50% gross profit margin. Let's take a look what $20 billion Intel Foundry revenue can do to the whole Intel Corp.:

Case 1. If this $20 billion Intel Foundry revenue is all coming from external non-Intel customers.

Intel Corp revenue: $20 billion
Intel Corp gross profit: $10 billion

Case 2. If this Intel Foundry $20 billion "revenue" is totally coming from the internal sales to the Intel product/design division, Intel Corp. should see $30 billion gross profit and $40 billion of revenue added to its book.

Calculation:
Intel Corp revenue: $40 billion revenue = $20 billion (Intel Foundry revenue) X 2 (in order to maintain a 50% gross profit margin for the Intel Product/Design division)
Intel Corp gross profit: $30 billion = $20 billion (50% of the above $40 billion revenue) + $10 billion (Intel Foundry gross profit)

I skipped many expenses to simplify this analysis.

The result is:

$40 billion revenue vs $20 billion revenue!
or
$30 billion gross profit vs $10 billion gross profit!

A logical Intel Board Director or shareholder will ask Intel CEO the same question: Can Intel Foundry give preferential treatment to the Intel Product/Design division? If not, then what's the point to have the foundry capability inhouse?

Some people may suggest Intel Foundry to build more capacity to satisfy both external and internal customers. This is a very risky gamble. We can see the pains and loss from the past drops of semiconductor demand. The most recent one was just last year. It can kill a company!

To me the Intel IDM 2.0 is a business model of contradictions and conflict of interest.

 

[hskuo]

The US is a big testing ground.

I would say TSMC isn’t getting a passing grade on any metric.

Intel may have struggled and lost its way with BK and his minions.

There is no corner on chemicals, tools, or physics, innovation that Taiwan owns that ensure it keeps its technology lead over Korea or US.

For Foundry success what is indeed unique is the ability to deliver accurate PDKs and running NTO on target and keeping a multitude of products that behave differently due to density and design sensitivities on target that requires lots of customization and lastly saying yes to ever customers realistic needs. Here TSMC has no rival, but Intel and Samsung can easily do this as well. There is arrogance thru the TSMC from top to bottom these days, same as Intel and IBM before that is certain.

The world is a crazy place and anything is possible and arrogance is even more dangerous than ever.

The one thing TSMC has currently is scale. When you have Apple, Nvidia, AMD, Qualcomm, Intel, Microsoft, Broadcom and others you just need to listen to your collective customers what they want and of course the scale of all those customers enables scale and learning that is unprecedent, but take Xi and POTUS misscalculation to go gone, or a node delay like Intel did, LOL

It is not that easy. Foundry business is a team work and takes times to form the right culture. TSMC has executed it very well and has track record for decades. I remembered in somewhere, tsmc executives said there customer 1st silicon success rate (NTO successful) was over 98% after learning for years. FYI. Even the customer service manager is critical.

 

[Maximus]

I remember when Intel was this arrogant. AT&T too. And IBM. Technology evolves faster now. Apparently Wei needs to reread his copy of Only The Paranoid Survive. All that's missing is a quote saying something like "Intel and Samsung will never be in our windshield".

The audiences of this talk are Taiwanese. Recently Taiwanese people feel uncomfortable about (1) CHIP Act and (2) tsmc's oversea expansion that they are afraid the positions of tsmc and/or Taiwan will be weakened.
Burn Lin (the inventor of immersion lithography, formal tsmc's R&D VP) and others just wrote an article titled "How America’s CHIPS Act Hurts Taiwan"

How America’s CHIPS Act Hurts Taiwan | by Chang-Tai Hsieh, Burn Lin and Chintay Shih - Project Syndicate

Chang-Tai Hsieh, Burn Lin and Chintay Shih worry that US inducements are luring the world’s top semiconductor manufacturer down a dead end.

www.project-syndicate.org

I feel CC Wei is not so arrogant that he mentioned many times about how a company would fail in other talks. The reason CC to make the statement was probably to relieve Taiwanese people's concerns.

 

[soAsian]

I remember when Intel was this arrogant. AT&T too. And IBM. Technology evolves faster now. Apparently Wei needs to reread his copy of Only The Paranoid Survive. All that's missing is a quote saying something like "Intel and Samsung will never be in our windshield".

Is he supposed to say we are a third-rate foundry? Come on now, every CEO of any business will say they are the best.

 

[Arthur Hanson]

Trust, integrity, skill set and execution are a very rare combination and hard to maintain at the leading edge. In these factors TSM has mastered for decades and is ingrained in their culture on top of a constant quest for excellence, which attracts top talent and allies.

 

[Xebec]

Is he supposed to say we are a third-rate foundry? Come on now, every CEO of any business will say they are the best.

I agree with this but there’s also a way of communicating ‘we are leading the world, but the hard work that got us here needs to continue if we want to stay that way’.. ‘competition isn’t resting on it’s laurels’, etc.

TSMC will lose at least some of its cheese regardless over the next 5 years to Intel.

 

[Mooredaddy]

I agree with this but there’s also a way of communicating ‘we are leading the world, but the hard work that got us here needs to continue if we want to stay that way’.. ‘competition isn’t resting on it’s laurels’, etc.

TSMC will lose at least some of its cheese regardless over the next 5 years to Intel.

I wouldn't be so sure. Intel is just i9up-[=q2w3e4r53qw1

 

[blueone]

The audiences of this talk are Taiwanese. Recently Taiwanese people feel uncomfortable about (1) CHIP Act and (2) tsmc's oversea expansion that they are afraid the positions of tsmc and/or Taiwan will be weakened.
Burn Lin (the inventor of immersion lithography, formal tsmc's R&D VP) and others just wrote an article titled "How America’s CHIPS Act Hurts Taiwan"

How America’s CHIPS Act Hurts Taiwan | by Chang-Tai Hsieh, Burn Lin and Chintay Shih - Project Syndicate

Chang-Tai Hsieh, Burn Lin and Chintay Shih worry that US inducements are luring the world’s top semiconductor manufacturer down a dead end.

www.project-syndicate.org

I feel CC Wei is not so arrogant that he mentioned many times about how a company would fail in other talks. The reason CC to make the statement was probably to relieve Taiwanese people's concerns.

The article is nationalistic propaganda. The part about TSMC Washington is almost word-for-word a quote from Morris Chang. The arguments in the article are so weak that I'm surprised authors of this caliber put their names on it. For example, the article speculates that due to the CHIPS Act, TSMC may lose its focus on innovation and start to focus on chasing subsidies. That's ridiculous. TSMC is supposedly spending $40B in Arizona, and their CHIPS Act funding is likely to be a few billion at best, and likely spread out over several years. The total manufacturing subsidies for the entire Act are only $39B, and 170 460 companies have applied. Nah, ridiculous is too weak a description.

I wonder... does Taiwan's leadership and Taiwan's people seriously believe that US politicians are simply going to stand by and do nothing about the semiconductor dependency on Taiwan, while the US-China relationship degrades more with every passing month? I think the CHIPS Act is a weak and silly waste of money. If Congress had asked me for a proposed draft of the bill, then Taiwan and TSMC would have had more to worry about.

 

[TBiggs]

Business history is full of companies that dominated and great leadership, vision, culture, strategy, scale and were going to dominate forever; Bell, IBM, GE, Intel, Poloroid, Kodak, Sears, Nokia, BlackBerry, Nortel.…. The list is long and twenty years from now there will be some new ones we think can’t lose.

Some like Microsoft and Apple rediscover dominance to recover themselves, will Intel be the same?

Every failure is so obvious in hindsight. No great company really knows how it will stumble nor is the leadership so terrible it doesn’t try to stop it yet failure happens.

If Intel really need the CHIPS Act to become relevant God help them! Nor in the CHIPs Act directly going to lead to TSMCs fall. It will be some dominos that might be predictable now but not obvious, who has the Trump card , LOL.

 

[blueone]

Is he supposed to say we are a third-rate foundry? Come on now, every CEO of any business will say they are the best.

Of course not. But there is a difference between demonstrating confidence and deep knowledge and boasting. Perhaps it is a cultural thing, but I find boasting like this, and especially what Gelsinger does, unappealing and gives the impression of weakness.

 

[blueone]

Business history is full of companies that dominated and great leadership, vision, culture, strategy, scale and were going to dominate forever; Bell, IBM, GE, Intel, Poloroid, Kodak, Sears, Nokia, BlackBerry, Nortel.…. The list is long and twenty years from now there will be some new ones we think can’t lose.

Some like Microsoft and Apple rediscover dominance to recover themselves, will Intel be the same?

Microsoft recovered itself by getting rid of Steve Ballmer as CEO and putting Satya Nadella in that position. Ballmer thought the world was Windows in clients and servers forever, and caused Microsoft to lose cloud computing leadership to Amazon, and they still haven't caught up. Now Nadella is pushing hard in AI, which is the right thing to do. He's the best thing to happen to Microsoft in a long time. Gelsinger is, IMO, like Ballmer, still living in the past, so I'm much less optimistic, but his strategy for IFS gives me a little hope.

 

[Tanj]

When you have Apple, Nvidia, AMD, Qualcomm, Intel, Microsoft, Broadcom and others you just need to listen to your collective customers what they want

And what they foresee. This is an often overlooked key benefit. All the futures funnel through TSMC at the moment. Any one customer may stumble, but the ones who see the correct path will grow, and TSMC has all paths covered.