Discover more from The Asianometry Newsletter
TSMC, Briefly Explained
Taiwan’s Most Important Company
You can watch the video for this script below:
Author’s note: TSMC’s market cap is heading towards $500 billion. Per Bloomberg as of this writing, it is sitting at $471 billion. That is a pretty amazing achievement. If you look at all the biggest market cap public stocks (which excludes some Chinese companies like Tencent and State Grid, etc) then TSMC sits at 11th place. The next largest Taiwanese company appears to be Foxconn, ranked at 409th place.
(Off topic but how about Kweichow Moutai, a liquor company being ranked 19th place?!)
I have noticed that a lot of people find this channel through my work in TSMC and semiconductors. I feel then that it would be a good time to bring this script out of the archives as a brief narrative of TSMC’s birth and rise to prominence.
Want to help support the Asianometry YouTube channel? Check out the Patreon and help me pay for all this coffee I spend at the cafes I go to write.
TSMC or Taiwan Semiconductor Manufacturing Company is arguably Taiwan’s crown jewel. It bears the pride of the nation and is the most valuable company that the island has ever created. Foxconn might make more revenue and gets more headlines but most of its factories are in Mainland China.
The founder Terry Gou is also a shifty character with an iffy reputation and close ties with the Communist Party. Regardless of his political stances, he is a shifty character who is known for scouring the world cutting (and ditching) subsidy deals with governments desperate for the jobs his company provides.
Author’s note: If you are interested in learning more about Foxconn, you can watch this video that I did a while ago.
The majority of TSMC’s manufacturing facilities are in Taiwan. The company has stayed in Taiwan and it proudly bears the island’s name in its own. Its founder and first CEO Morris Chang is revered in Taiwan.
This article explains what TSMC does and why it is such a big deal.
TSMC From the Beginning
In 1985, a 50s-something former Texas Instruments executive named Dr. Morris Chang was looking for a new job. He had just quit a short stint as COO and President of General Instrument — a small semiconductor manufacturer. He had been disappointed by the company’s unwillingness to invest in R&D.
It was then that a Taiwanese government official named K. T. Li brought Dr. Chang, born in Ningbo near Shanghai, over to Taiwan. He was asked to serve as president of ITRI, a government research institute studying new technologies. He streamlined the government organization, making it compete for private contracts and even firing a few people for incompetence. This would earn him hate-mail from Taiwanese legislators. A year or so later, Li came again with another idea: The Taiwanese government wanted a semiconductor company that can compete with the established semiconductor giants like National Semiconductor, Intel, and AMD. Why not be its CEO? That company would be TSMC.
TSMC started life as a Hsinchu-based corporate spinoff from ITRI. It was jointly owned by the Taiwan government (which continues to own the largest single portion today with some 6%), electronics company Phillips, and private investors — $220 million of startup funding. Founder Morris Chang is today retired and a billionaire (just barely). Unlike with Jeff Bezos or Bill Gates it is not because he owns founding stock. He owns just about 1% of TSMC.
ITRI — TSMC’s research institute parent, had licensed an outdated processing technology which it had later improved on its own to a newer generation. The organization though was still 2 generations behind its competitors. In an industry where either you are at the latest generation or you suck, this is a big disadvantage.
(Author’s note: If you want to know what happens to those who cannot rise fast enough, watch this video about Singapore’s TSMC)
Adopting the traditional semiconductor manufacturing business model to compete head to head against the likes of Intel would guarantee automatic failure. Chang needed to find a model that could work around this shortcoming.
He would recall a friend who approached him a few years ago looking for money to buy equipment for manufacturing some chip designs he had made. That friend would eventually find a manufacturer willing to “rent” him the capacity at a fraction of the cost. This got Morris thinking.
Around then, there were about 50 semiconductor companies called “fabless semiconductor”. They designed special purpose chips that were brought to life by the big semiconductor giants like Texas Instruments, Intel or IBM. Those giants drove tough bargains. They gave their fabless customers lower priority space and demanded rights to the design. So if a fabless company made a hit, the giant would come out with a competing chip under their own label. But what could the fabless companies do? No one could afford $50-100 million dollars (this is before VC was a big thing) to buy this expensive semiconductor manufacturing equipment.
Until now. TSMC opened its doors with a controversial premise: You guys design and sell your chips. We handle all the manufacturing and will not compete with you. TSMC realized the crucial thing about this sort of industry — that it was a service-oriented business. Despite the fact that TSMC’s technologies were far behind the market leaders, it can collect customers on the basis of customer service, trust, and collaboration. TSMC succeeded when its customers did.
So TSMC grew. It grew by helping fabless customers like Broadcom, Nvidia, and others grow. It invested immense amounts of money in R&D to investigate the latest techniques and in servicing its customers. Being able to work with all different types of customers helped it see the most common types of challenges that customers would face.
Apple is TSMC’s largest and most important contract — all it took was employing 6,000 people 24/7 and investing $9 billion of their own money for a brand new factory oil Tainan in just 11 months. Today, Apple contracts the entirety of TSMC’s leading edge process capacity, 5 nanometers or N5, down in Tainan for their iPhone, iPad and Mac products.
Author’s note: If you want to know more about the relationship between Apple and TSMC, you can watch the video below
Morris Chang retired in June 2018 after some 30 years as TSMC’s CEO. Today, the company is the acknowledged market leader in the foundry industry. It had started out with technology far behind the rest but is now either number one or two, depending on who you are asking.
The semiconductor foundry industry at the leading edge used to be quite crowded with plenty of choice. But over time things got more and more expensive and people started dropping out. The latest to do so was GlobalFoundries, the leading player with American facilities.
This leaves just TSMC, Intel, and Samsung Foundry as the last competitors in the industry capable of doing work at the leading edge. This does not mean that TSMC has a monopoly. Samsung Foundry has quickly emerged as one of TSMC’s toughest current competitors and a critical second source for fabless customers looking for leading edge node capacity.
But the market is indeed tightening. This is to be expected for any maturing industry as players consolidate. This has been happening for a while. But the question has gotten more complicated with US sanctions and the US-Chinese trade war getting involved. The Chinese startup foundries are moving fast up the value chain and have billions of government support behind them. We will all need to keep a close eye on the sector in the coming future.
If you know anyone who might be interested in reading this, share it with them