Discover more from The Asianometry Newsletter
Visiting the Morris Chang & Chris Miller Semiconductor Forum
Recently I went to a high-profile semiconductor forum held by the Commonwealth business magazine in Taiwan.
Chris Miller, author of the popular book "Chip War" had released his Chinese edition. He was visiting Taiwan for the first time - presumably to promote it.
Morris Chang was prominently featured in this book, so the publisher Commonwealth arranged for the two to meet. I haven’t read the book yet but I love field trips so I went to the event.
In this video, I am going to give some thoughts. I was not taking detailed notes throughout the talk so most of this is just from memory. My memory is spotty so forgive me.
I get the sense that this started out as simply part of the Chip War Asia book tour - Miller is going to Japan and South Korea after this, I think.
But then things escalated when Morris Chang agreed to get involved. Why else would all these foreign diplomats, government officials, semiconductor executives come?
I saw Mark Liu - TSMC’s executive chairman - take a seat in the crowd. In some of the social media, I also noticed TSMC CEO CC Wei.
Then Morris Chang came out alongside the Taiwanese Vice President, with Chris Miller and Morris’s wife Sophie Chang close behind them. Seeing him up close, I really got the sense of his frailty and advanced age.
The Taiwan VP gave the first speech, which covered topics that are important but unrelated to what I have to say here.
Then Chris came out and gave a brief introductory lecture.
He went over some history - starting with the first computers like the ENIAC and how they were used to break codes and guide artillery.
He then briefly touched upon how semiconductor manufacturing moved from the United States to Asia. He showed a picture of a Hong Kong building with the Fairchild Semiconductor logo.
Fairchild had been one of the first semiconductor companies to outsource their IC packaging over to Asia. Previously, the majority of an IC's final cost came from the packaging - the casing surrounding the die.
To bring down those packaging costs, they used something called plastic encapsulation. They hired cheap Hong Kong or South Korean labor to put the die on a ceramic bead and smother the whole thing in plastic.
Miller then showed a Texas Instruments strategic document from 1976 in which Morris Chang proposes the concept of an independent foundry.
Texas Instruments ultimately didn't go with this idea and Miller closed by humorously musing why the "T" in TSMC didn't stand for "Texas" rather than "Taiwan".
He hoped to learn more about this in his discussion with Morris.
With that, Morris got up and got into his seat - with the help of his nurse. And thus began the forum discussion.
First things first. Got to plug the book.
Morris said that he did read the book - in both English and Chinese - and then said the words every author loves to hear: "I wish I could have written it myself". After that, Chang says that one thing he felt that the book had overemphasized was the Taiwanese government's involvement in TSMC's founding.
I listened to Miller's interview on the Verge. And Miller did seem to say as such in that interview. And I quote.
> That’s why the Taiwanese government put up over half the capital in TSMC when it was founded. It was a direct project of the Taiwanese government to make Taiwan more indispensable in electronic supply chains. And it has worked.
Morris Chang somewhat downplays this sentiment in his remarks. He said that the Taiwanese government - with the exception of one man, his friend K.T. Li - did not much believe in him.
So the Taiwanese government contributed in only two ways. First, a financial investment of about 500 million NTD. And second, a transfer of about 100-120 technicians and workers from ITRI. That is about it, no ongoing manufacturing subsidies or the like.
Then when TSMC went public in the early 1990s, the Taiwanese government couldn't sell the stock fast enough! They sold it all the way down from 48% to the 6 or 7% that they have today. Chang had to directly ask the government to stop selling, which to their credit they did.
I felt it pretty ballsy to be saying all of this in front of the ROC Vice President, but old men say whatever they want.
Miller asked him if - at the start - he knew that TSMC would end up being this big. Chang replied that at the start, he had only concentrated on survival. The company's first years were pretty tough and it would not be until the early 1990s that they really got their legs under them.
Pieces of a Puzzle
Chang also addressed Miller's remarks about the independent foundry idea being proposed at Texas Instruments during the mid-1970s. Morris points out that the foundry idea needed a few more pieces at the time for the puzzle to emerge.
Notably it needed the VLSI work of Carver Mead and Lynn Conway - something which I discussed in an earlier video. Their work in modularizing the semiconductor production process into "design" and "manufacture" steps made a foundry possible.
The last big piece of the puzzle that Morris said motivated him to pursue the independent foundry idea had been a conversation he had with Gordon A. Campbell, not the Canadian diplomat. Campbell and Chang knew each other by reputation, and when Chang was at General Instrument he thought to reach out about some sort of OEM collaboration or something like that.
Campbell told Chang that he had subcontracted a fab from Japan and because of that he no longer needed $50 million, just $5 million. After that, Chang knew that the foundry idea would work. There is one last piece to mention, the technicians in Taiwan but that is for later.
One other big thing that Chang closed on had to do with "friendshoring".
Morris - an American citizen - pointed out that the US Secretary of Commerce Gina Raimondo has consistently excluded Taiwan as one of those friends in its "friendshoring" semiconductor initiative.
She has previously said that American dependence on Taiwan for chips is "untenable and unsafe". There is a scenario - not sure how likely it is - that the United States turns its technology export restrictions to Taiwan as well.
Suddenly, companies like Applied Materials and Lam Research can no longer send their best stuff to Taiwan. Chang says it's understandable that you do a few chips in America for national security. But what is the end game here?
Are they trying to get back to the mid-30% global share that America had back in 1990? Because that's no longer America's advantage.
America has a lot of great competitive advantages. For instance, in design. All the best semiconductor design companies are in America - which makes sense because "design should be close to the end user".
But America has no competitive advantage in chip manufacturing. Chang was pretty blunt about this. Before, he has said that America's chip manufacturing costs were 50% higher than they were in Taiwan. Now he says that was an underestimate, and perhaps a far worse underestimate than he underestimated! People laughed at that.
Semiconductors have been one of the most disinflationary products in human history. If their costs were to rise substantially, that would have a big impact on their overall ubiquity. I know that there are some Luddites and YouTube commenters who might think this a great thing, but whatever.
Chang goes on to comment on American work culture. Now let me first preface this. He is well known for being a pretty difficult person to work for. He worked his way up from operations, and there is no forgiveness in operations. He is also a grumpy Asian grandpa. My experience with grumpy Asian grandpas is that they like to murmur.
Anyway, Chang is very hard on the American worker. The last big piece of the puzzle for Taiwan and TSMC are its technicians. He says in Taiwan, he gets amazingly skilled workers even out of 2-year vocational schools.
These guys matter. They are on the front lines. In his last days at Texas Instruments as SVP of Quality and Productivity, TI learned that their new fab in Japan had twice the yield of their Houston fab and asked Chang to go check out why.
Chang goes and learns that technician turnover in Japan was just 3-4%. In Texas that number is 4-5 times higher. Technicians in Asia just work harder and are more dedicated than they are in America.
I’m going to paraphrase him here. If a piece of equipment on the line breaks down at 1 AM, then in America they will fix it by 9 AM the next morning. The technician comes in at 8 AM and fixes it. So for 8 hours the line sits there not doing anything.
But in Taiwan, the technician goes in at 2 AM. He will get a call while he's sleeping and he will get up and get ready to go to the fab to fix the equipment. His wife will wake and ask him why he is up and he will tell her that he needs to go to the fab. And she will go back to sleep without complaint or argument because that is just the culture.
The equipment is expensive and when it stops, the whole line stops. That is the first "No No" of manufacturing, which was ... aaaand then suddenly Sophie his wife interrupted him to say that time was up and Morris had to go.
Few Last things
A few other fun random things that I heard at the forum.
Philips was another one of TSMC's early investors - the second biggest. Chang brought them into the deal at the government's request to validate the concept. In a previous video I talked about Philips’ early role bringing in their IP and technology to TSMC.
Chang didn’t mention that here. Apparently they had little belief in TSMC too, and wanted to get their money out fast. But Morris said he didn’t mind that because he didn’t like having them on the board. That according to him, is why TSMC went public in New York.
He was briefly asked about Intel and their foundry model. Chang responded by quoting one of his big customers Jensen Huang, "TSMC has learned to dance with 400 partners. Intel has always danced alone." I think that says everything.
The event was fun, of course. But there were so many other questions that we did not get to ask him.
Chang mentions that in 2010 he knew that TSMC would be hugely successful. I cannot help but notice that this was after he returned as TSMC CEO - overthrowing Rick Tsai. That would have been fun to hear about. I want to have known what it was like to balance all these smart people and difficult technical decisions. How did he mediate between and manage these huge egos?
In the 2000s Morris had been one of the leading voices lobbying the Taiwanese government to allow TSMC to build fabs on the Mainland. He also blessed Dr. Chiang Shang-yi to leave Taiwan and work for SMIC. Now he says he completely supports the export restrictions on China. Did his opinion change? And if so, why?
He talked about Japan and China. What about South Korea? What was his relationship like with Samsung? I heard a rumor that the Samsung chairman Lee Kun-hee tried to recruit him. Is it true? If so, why didn't he do it?
And so on. Morris Chang is like most grumpy 90-year old men. It is hard to get a hold of him, but when he starts talking then he talks for a while - saying whatever he likes. And the best thing about 90 year old dudes is that they don't give a hoot. Not being able to ask him all the questions we had wanted to ask him feels like a missed opportunity.