We dive into the unbelievable and unlikely history behind the quietest technology giant of them all: the Taiwan Semiconductor Manufacturing Company. Founded in 1987 by the then-56 year old Morris Chang, already a legend in the semiconductor industry by virtue of his meteoric rise and fall at Texas Instruments, TSMC today manufactures nearly all the leading-edge chips for Nvidia, Apple, Broadcom, Qualcomm, AMD, and yes — even Intel. Tune in for an incredible story of innovation, perseverance and lasers. Lots and lots of lasers!Note: this is a remastered version of our original 2021 episode. We don’t often re-release old episodes, but in this case we have a very timely reason for doing so. Stay tuned! :)Sponsors:Many thanks to our fantastic partners:VantaJ.P. Morgan PaymentsStatsigLinks:Episode SourcesCarve Outs:Ted Lasso (Season 1)GreekWho is Michael Ovitz?More Acquired:Get email updates with hints on next episode and follow-ups from recent episodesJoin the SlackSubscribe to ACQ2Check out the latest swag in the ACQ Merch Store!Note: Acquired hosts and guests may hold assets discussed in this episode. This podcast is not investment advice, and is intended for informational and entertainment purposes only. You should do your own research and make your own independent decisions when considering any financial transactions.
Hello, Acquired listeners. We regularly get feedback that this episode on TSMC, the Taiwan Semiconductor Manufacturing Company, is one of the best Acquired episodes ever. And interestingly, it predates our NVIDIA episodes. We did it way back in 2021 when the Acquired audience was about 12% the size of what it is today, which means that the vast majority of you have never heard it.
So we definitely wanted to fix that. Since then, semiconductors have become so much more important in our world, and TSMC has essentially become the only manufacturer of the leading edge chips. They make the primary chip inside every MacBook and iPhone shipped today. They're powering the AI wave, manufacturing all of NVIDIA's chips.
They make the chips for a whole bunch of other fabless companies like Qualcomm, AMD, Broadcom, and hyperscalers like AWS.
And it turns out they even manufacture a lot of chips for Intel, too. Yes. Little known fact. Yeah, TSMC rode the smartphone era to crazy heights, as we all know. And here now in the next AI era here in 2025, it turns out that they are the manufacturing superpower behind all of that, too.
Yep. Well, listeners, without doing too much foreshadowing, now is a very good time for anyone to listen or re-listen to the TSMC episode. So we decided we should go all the way back to the raw audio tracks and remaster this whole thing from scratch for your listening pleasure.
Ben, in fact, I looked it up since we were going back to 2021 when we initially recorded this. TSMC's market cap has doubled since then from $550 billion to over a trillion dollars. And in fact, you were the one that tipped me off to this as we were re-researching here.
They and Saudi Aramco are the only trillion-dollar companies in the world that are not located on the west coast of the United States. Wild.
This is such a crazy stat. It's crazy that the rest are located on the West Coast of the United States, but it really underscores what an extreme outlier TSMC is. So without further ado, the story is truly unbelievable, and we hope you enjoy this presentation of TSMC Remastered.
Is it you? Is it you? Is it you? Sit me down. Say it straight. Another story on the way.
Welcome to Season 9, Episode 3 of Acquired, the podcast about great technology companies and the stories and playbooks behind them. I am Ben Gilbert.
And I'm David Rosenthal.
And we are your hosts. Today's episode is on TSMC, or the Taiwan Semiconductor Company. It is your classic, most people have never heard of it, but it's the ninth largest company in the world episode.
This is wild. Morris Chang founded TSMC at age 56, retired at 74, then came back at age 78, inked the deal to make all of Apple's chips. And yeah, we're going to tell the whole story here. It's wild.
It's nuts. They make literally every chip in every iPhone sold today and soon to be in every Mac sold. If you're excited at all about NVIDIA, AMD, Qualcomm, or even any of the chips that Amazon, Microsoft, Facebook, Apple are making...
All of those chips, or nearly all of them, are actually made by TSMC, along with all the chips in your cars and your smart home devices and fighter jets and everything. Unbelievably, this company that the entire world relies on is on an island that some countries feel is a sovereign nation and the People's Republic of China feels is actually theirs.
So today's episode has it all ascending from startup to tech superpower, an underdog founder, and of course, a good dose of geopolitics.
Indeed.
All right. Well, listeners, it finally felt like the right time to do this episode amidst this global chip shortage that we've got going on that, David, I think I've heard even Ford has paused the production of F-150s because of this. So it is like a massive impact on the world. And we've had TSMC on the agenda to do for like two and a half years now in our little Google Doc.
Totally. Well, I feel like we haven't called it a miniseries, but let's call it a miniseries on semiconductors and, like, silicon. The ARM episode. Yep. Sequoia Part 1. PA Semi.
Yep. Okay, listeners, it is time to jump into the history and facts, and David's going to lead us in that. But as usual, even though we're going to be probably very excited about some companies, less excited about other companies, this show is not investment advice. We may have investments in the companies we discuss.
It's for entertainment and informational purposes only, and you should do all of your own research.
Okay, speaking of, we start in Ningbo, China in July 1931. just about one year after Warren Edward Buffett was born in Omaha, Nebraska. And there are going to be quite a few parallels here as we go through this episode. But in July 1931 in Ningbo, China, our protagonist, Dr. Morris Chang, Order of the Propitious Clouds with Special Grand Cordon,
which is the highest civilian honor that anyone in Taiwan can hold. So he's like a knight of Taiwan. It's the order of propitious clouds. And then I think there's like nine ranks of it. And the highest is special grand cordon.
And he's special grand? He's special.
Yes, he's very special. So he was born then. For those who are unfamiliar with Chinese geography, Ningbo is a small city just a bit south of Shanghai. You know, small. It's about 8 million people.
It's casual. No big deal. China's scale is ridiculous, but certainly wasn't 8 million people when Morris was born in 1931.
No, but I bet it was still probably. pretty big. But yeah, today, 8 million people. Crazy. So Morris's father was a county official and later became a bank manager. So the family moved around a good bit within China as his father was transferring for work. This is pre-People's Republic of China. This is pre-World War II. This is a very different place.
Right. The leadership is not communist.
No, no, no. So his early childhood years were like middle class, not wealthy, but pretty well to do relative to your average Chinese citizen. Then when he was six, the Second Sino-Japanese War breaks out. And Morris and his mom flee the main part of China to Hong Kong. And they go to live in Hong Kong for a few years to escape the air raids and the fighting.
And then on December 8th, 1941, three hours after Pearl Harbor, the Japanese attack and invade Hong Kong. Morris talks about this. He's like, yeah, everybody knows Pearl Harbor, December 7th, 1941. What people don't often talk about is the same thing happened in Hong Kong three hours later on the next day. So they're in Hong Kong. So they flee again back to China.
They end up in Shanghai this time and they stay there for a few years until 1948 after World War II is over. But that's when the Chinese Civil War started. breaks out. That would lead to the Chinese Communist Revolution. And so they flee again back to Hong Kong. So this is crazy. Morris, before he turns 18, he has lived through three major wars.
The Second Sino-Japanese War, World War II, and the Chinese Civil War. So the next year in 1949, which is the same year as the establishment of the PRC, the People's Republic of China, Morris turns 18. And with the help of an uncle that he has in Boston, his life completely changes. He gets accepted to Harvard. So he goes to the US, he goes to college at Harvard. Wow. Talk about a change of fate.
Talk about a change of fate, a change of scene, everything changes. Morris says much later, my reaction entering Harvard was sheer ecstasy, almost disbelief. What a country! The United States was at its peak in its moral leadership and its political leadership in terms of democracy, and it was the richest country in the world.
not to mention stable. I mean, you could say what you want. You could count on the fact that it's likely that 10 years from now, whatever economic structure, political structures exist will continue to exist. If what you want to do and what he ended up doing with his whole life is innovate, having that stability around you and all those structures enable you to do that.
Yeah, like we just take this for granted, but it's a good reminder. At the very least, he's probably not going to have to flee Boston to continue his studies. But he does end up fleeing Harvard, as we'll get into. So Morris loved it. It was like that quote we read. He was so overjoyed to be there. But he realizes he has kind of a new problem in America and at Harvard.
His parents aren't coming over. He's on his own. He's got to support himself and make his own way. And at that time, his race is probably going to limit his opportunities. So as he says, quote, in the early 50s in the United States, there were Chinese laundrymen, Chinese restauranteurs, Chinese engineers and Chinese professors. Those were the only respectable professions for Chinese.
No lawyers, no accountants, no politicians. And what does Harvard churn out? Lawyers. Sort of accountants, maybe. Politicians, yes.
Yeah.
Not a lot of engineers.
Certainly finance professionals.
Certainly finance professionals. As we will see as we will go along, Morris is much more than a finance professional. But Harvard actually didn't have an undergrad engineering program at the time. Huh. That's crazy to think about. If you're really, really focused, you're probably going to go down the street in Cambridge from Harvard to MIT, which Morris does.
So he only spends his freshman year there. And then for his sophomore year, he transfers to MIT so that he can study mechanical engineering. So Morris, our man, has learned the ways of the world in the U.S. He's focused. He starts mechanical engineering a year behind at MIT. He finishes both undergrad and his master's in the remaining three years. And what year is this?
This would have been 1951 when he transferred, fall of 1951.
Okay, so to contextualize what's going on in the tech world right now with quotes around it, because it's not so much a world as a very small continent. I mean, you have all of the post-World War II defense spending that went in, particularly on the West Coast with the innovations from Stanford. But has Fairchild Semiconductor been started yet? Nope, nope, nope, nope.
So maybe Shockley Semiconductor?
Shockley Semiconductor is probably just getting going, but we're probably still in vacuum too. Like Bell Labs. Yeah. To give you a sense, silicon is years away. Transistors are probably just getting going. We're not in the integrated circuit yet. And it's all being done in germanium, not silicon. Wow. So it's like, this is OG. Yep.
So after he gets his master's in the three years, Morris wants to stay and do a PhD, fully complete his technical training. But he ends up failing his qualifying exams twice. They give you two chances to take and he fails twice.
By the way, this is a good time to say, so David and I watched and listened to every footage that Morris has ever spoken that has been released publicly to prepare for this. He is very funny.
Oh, he's great.
The way he talks about this, he says that, unfortunately, the biggest impediment to him going forward was that he failed the qualifying exam. But fortunately for him, they were kind enough to let him take it a second time, which he also failed. And he has this really dry, clever sense of humor.
So in one of the interviews he talks about, one of the Stanford ones, he gets a question from the audience about how did he kick his smoking habit? Oh, yeah. That like the questioner is like, I know you used to smoke. How did you finally stop? And he's like, I never stopped. I still smoke. He's like 94 years old.
And he goes on to make the case for why he's a pipe smoker. And actually, even though smoking is hurtful to his lungs, it's actually beneficial for his mental life. So he's pretty sure it's prolonged his life.
Well, he says he's delved into the data and pipe smokers live longer than non-smokers.
Which I'm sure you can find data to support that. I'm also sure you can find plenty of data to refute that. But yes, this gives you a sense of who Morris is.
Okay, so he's failed his qualifying exams. He's got to go out and get a job and, you know, not as a PhD. He's got to go get a job as a super entry level as an engineer. I mean, he has a master's degree, but still. Okay, so Legend has it. He has a couple job offers. The one he really wants, remember, he's a mechanical engineer. And this is like super early days of technology.
It's not really a thing.
There was electrical engineering at this time. Right, right, right. Yes, he could, but he didn't study electrical engineering.
In terms of where you would want to work, it's not really on anybody's radar screen, especially Morris's, that you're going to go enter the tech industry. So he gets his dream job offer from the Ford Motor Company.
Oh, no way. Yes.
I didn't hear that. And I'm sure this is apocryphal.
But let's repeat the apocryphal story and broadcast it out to hundreds of thousands of people here.
So the legend has it that Ford offers him a salary of $479 a month to go take an entry-level job. And then he has a competing offer from Sylvania's new semiconductor division.
And Sylvania, I know of this company only because my vacuum growing up was made by Sylvania.
Yeah. Oh, well, we can talk much more about Sylvania in one second. This is the competing job offer he's considering. They offer him a salary of $480 a month, $1 more. And legend has it that Morris asked Ford to beat Sylvania's offer. They didn't. And so he took the Sylvania job offer. I'm sure that is a... This is 100% applicable.
Yeah.
But, you know, Morris, he's great. So speaking of Sylvania, do you remember, I'm sure some portion of our audience remembers, but do you, Ben, remember who else started their career in Sylvania's semiconductor division right around this exact same time? We've talked a lot about this and this person on the show. No. Donald T. Valentine. No way.
Yep.
That's right. So he started at Sylvania after Fordham, or maybe it was after the military. He ended up at Shockley. Well, no, then he was at Raytheon, and then he joined Fairchild. Fairchild, okay. Right after the trade race, he left Shockley and started Fairchild.
You're better at remembering these deep details of older episodes than I am. Well, I do a lot of research for this show. Okay. And sometimes research includes past acquired episodes. There you go. So they didn't overlap, Don Valentine and... They were never in the same place.
They were in different locations and different job functions, very different job functions. But they were both, I believe, both at Sylvania. Amazing. At the same time. Crazy. So Don is out... chilling in California, like we were talking about, and falling in love with California. He's playing water polo. He's like, oh my gosh, I'm never going to leave this place. Morris, he's on the grind.
He gets posted as a junior engineer at Sylvania's Ipswich, Massachusetts plant. Not quite the same glamour as Don out in Southern California. So remember, Morris is a mechanical engineer. He doesn't know anything about electrical engineering, but he's working in this new semiconductor division. So after work, he's living in a hotel, by the way. He doesn't even get an apartment.
It's like some company-sponsored hotel. He goes home, back to the hotel from work. And he studies the best textbook that he can find about electrical engineering, which is entitled Electrons and Holes in Semiconductors with Applications to Transistor Electronics, written just recently, a couple years before, in 1950, by by William Shockley.
Oh, wow. Yeah. Shockley and two other guys basically invented the, I'm not sure it was the first transistor, but the first transistor of the type that everything else would then be built upon when they were at Bell Labs not too long before this.
Yeah, not too long at all. I mean, ENIAC was vacuum tubes. And then Shockley invented the transistor. And then in a sec, we're going to talk about the integrated circuit that Bob Noyce and Jack Kilby, who we're going to talk about, co-invented. But anyway, okay, back to this moment in time. So Morris is just studying the Shockley textbook in his hotel room. But he's not a college student.
He doesn't have any teachers. He just has the book. Wow. But he's very resourceful. So he figures out that one of the senior engineers at the plant... is kind of an alcoholic and hits up the hotel bar almost every night. So what Morris does is he comes home from work in the early evening. He studies in his room for a couple hours.
And then later at night, when the older colleague shows up at the bar, Morris goes down to the bar not to drink, but he brings the textbook and he asks the guy, he's like, I don't understand this. I don't understand that. Like, girl, man, he's just like... Buying drinks for his buddy. So great. Incredible. Here's the quote he says later.
He, being the older colleague, didn't solve all my problems, but he solved enough so that I could move ahead. He was my main teacher about electrical engineering. So great. Wow. So this goes on for three years. Morris is like rolling hard. He's burning the candle at both ends, working and at the bar, but not drinking, learning. But as he is learning the industry, coming up to speed,
It becomes pretty clear to him that if he really wants to go places in this new emerging industry, Sylvania, not really the right bus to be on, so to speak. And obviously, Don Valentine figures out the same thing and jumps to Raytheon and then to Fairchild. Morris says the moment when this crystallized for him was there was a talk that a senior manager at Sylvania gave at the plant.
And the quote that the senior manager said that stuck with Morris for the rest of his life was, we at Sylvania cannot make what we can sell and we cannot sell what we can make. Real great position to be in. So Morris was like, damn, I got to get the hell out of here.
That's a signal to move on if you've ever heard one.
Totally. So like Don, Morris leaves Sylvania for greener pastures. However, not to California. Halfway in between. Or to Silicon Valley. Yep. Halfway in between. So we talk... a lot about Fairchild and the Trader's Seat, Silicon Valley, blah, blah, blah, the place to be. Here's the secret. Silicon Valley is all marketing.
The biggest semiconductor company in all types, digital, analog, everything at that time was not in California. It was in Dallas, Texas. It was Texas Instruments.
Which, of course, me, you, many people in our generation know of as the people that made our graphing calculators in high school and college. But of course, at this time, I don't even think they had a consumer division yet. No, no, no, no.
And that's going to come up later. No, TI was the juggernaut. Like, Now, Silicon Valley is Silicon Valley. But then it was, yeah, okay, California, I don't know, West Coast, wherever. TI was the big incumbent. They were the juggernaut. TI actually got its start, I had no idea, before doing the research here.
In the 30s, you're like, how did a technology company and a semiconductor company end up in Dallas, Texas? They started making instruments, Texas instruments, for measuring seismic activity for oil exploration. Whoa. So all the oil companies- Oh, that makes sense about Texas. Yeah. They were like the technology provider to oil companies.
And that's what led them into computing and into digital to power that business. They were huge, not just huge in terms of the company, but they were the technology leader. So Bob Noyce, like I was saying a minute ago, is credited, you know, when he was at Fairchild inventing the integrated circuit and all that. Well, he was the co-inventor.
Simultaneously, it was co-invented by Jack Kilby, who was at TI. And Jack was actually the one who got the Nobel Prize for inventing the integrated circuit. Gordon Moore, who was also at Fairchild and then founder of Intel, along with Noyce, he would coin Moore's law. But Jack has a great quote, too, about the implications of the integrated circuit and semiconductors.
He says, what we didn't realize then, this was a little later, when they were inventing it, was that the integrated circuit would reduce the cost of electronic functions by a factor of a million to one. Nothing had ever done that for anything before. Wow. It's such a great way to frame it too. Like this had never happened in human history.
There was this thing that used to be X expensive in terms of resources. And then magically one day it's a million times cheaper.
Yeah, that's crazy. I didn't realize it was on that scale. This is probably a good time to talk about some definitions because there are some things that we've thrown around already. I think everyone has a general understanding of what these things are, but it's worth understanding more precisely before we move on. The first of which is a transistor.
The best way to think about a transistor is not the tiny little transistor that's on a silicon die today, but think about it as a little encased piece of circuitry with three prongs coming out of it. And those three prongs will save the technical names, basically have an input, an output, and something that controls the input and the output. It's a switch.
It has two purposes, the first of which is being a switch where you can decide that either stuff is going to go through it, stuff being voltage, current, or none, or it rounds to none. And so that way you can decide, hey, this binary piece of equipment is either off, zero, or on, one. Okay, so that's a transistor. Now, a transistor can be made out of lots of different things.
It can take any implementation. Why is everybody talking about silicon? Well, silicon as an element is a semiconductor. It is a metalloid. It has some properties that make it like a metal, like a conductor. It has some properties that make it non-conductive. Imagine trying to move electrical signal through a piece of wood. It's not going to work. But imagine moving it through copper.
It's going to work really well, and you're never going to be able to interrupt it. Well, geez, wouldn't it be great if we had some material, a semiconductor, where we could modify whether current was flowing through it or not? Make it a switch really easily. Yeah.
Exactly. Well, and lots of things are semiconductors. Germanium was the main material for a while, but germanium is expensive and rare.
Silicon is made of sand. I think it's like the second most plentiful, mineable element on Earth.
Yeah, I mean, it's sand, right? Yeah. There's one other major thing, though. So we've been talking about transistors. The IC. Yeah, the IC. The integrated circuit. The integrated circuit. A transistor, it's a switch. Yeah. Before the IC, people were making switches. Like you make one switch at a time, you wire it to another switch.
You know, if you've seen photos of ENIAC and vacuum tubes, literally they're plugging one tube into another. You're still doing that with transistors. When Noyce and Kilby invent the IC, now you can put a lot of switches on one thing. And, you know, fast forward today, the latest processor, you know, the five nanometer processors that TSMC and basically nobody else is churning out.
I don't know, billions, trillions of switches are in like a tiny little... Integrated circuit. Without the integrated circuit, that never would have happened. So this invention, this miraculous invention of the integrated circuit, it happened in 1958. When did Morris Chang join Texas Instruments?
1958. Fascinating.
Coincidence? Yes, totally a coincidence. Absolutely a coincidence.
And again, to peg us in history here, we're still, I think, 10 years before the founding of Intel. Yes, exactly 10 years.
Yeah, Morris obviously wasn't working directly with Jack on inventing the IC. But this gives you a sense. TI, this is the place. This is like Google plus Facebook. Without the world paying attention to them. Yes, and in Texas. So Morris gets assigned as his first project to a sort of problem child within TI called They have entered into a deal with IBM.
IBM is working on their first mainframe computer major project that's going to use transistor logic instead of vacuum tubes, the IBM 7090. And they anticipate so much demand.
for this product usually ibm manufactures everything for all their products themselves but they're like we need more chips than we're going to be able to make ourselves so we need a second source for our chips they turn to ti and they're like hey we can give you all the designs for you know how to do this this chip that we want for our product but
We want you to additionally manufacture some of these in addition to our own line. You might even say almost like a contract manufacturer of chips or like a foundry business almost, you know. Hmm, interesting. But it's not going too well.
So IBM's own plant is churning out transistors with about a 10% yield, which means that of every 100 chips that they turn out of the plant, 90% of them fail and only 10% of them work. That's the first party line. The TI line has about a 0% yield. They're lucky if they're getting any that way. Almost everything coming off the line fails at TI when Morris shows up.
So Morris would say about this later, quote, the supervisor was concerned. The operators were concerned. Everybody was concerned. Yeah. So Morris, remember, he's a mechanical engineer by training. Right. So he starts tinkering. He's like, well, I know this is a mechanical process, chemical and mechanical process creating this stuff.
I'm just going to use my training and optimize it like a good mechanical engineer. So he starts doing some stuff. After about four months, he gets the yields at the TI plant up to 20%. So twice as good as the first party line at IBM. And there's a great profile that was one of the main sources for this episode in IEEE Spectrum. Oh, yeah. Great industry magazine that we'll quote from here.
They write, suddenly, even TI president Pat Haggerty knew Morris's name. IBM thought Chang had just gotten lucky, but when the company, IBM, sent engineers down to talk to him, Morris described the theories he'd been testing and explained why his experimental process worked.
This achievement propelled him into his first managerial job, creating a germanium transistor development department with 20-plus engineers reporting to him. So this is his first big win here in the foundry business. So on the back of all this, TI is like, all right, we got a rising star here. They offered to sponsor him to go finally get his PhD.
They even offered to continue paying his full salary while he's getting his PhD, which they're paying for. All right. So they think like very highly of him. Very, very highly of Morris. I mean, this one probably made the millions doing this.
in 1958. It's funny, I don't know anything about the commercial success of that particular IBM mainframe, but if it's the first one that's transistor-based instead of being vacuum tube-based, I have to imagine it was far more efficient for customers. Customers are probably lining up for it.
I bet there's a lot of demand. And what's Morris making a year? Like $20,000, maybe? You know, how much does it cost to go to Stanford then? Not much. So they're like, sure. So Morris goes to Stanford, but he's now like a pig in mud. He has found his calling. He can't wait to get back to Texas, back to TI. So he finishes his PhD in two and a half years. Wild.
One of the Stanford interviews is with John Hennessy, the president of Stanford at the time. John's like, Morris, tell the students, how did you finish your PhD in two and a half years? Morris is like, I was focused. I didn't do much else. So by 1964, he's done. He's back at TI. And this is right as people have discovered that silicon is way more cost effective and scales up way better.
And if I remember right, the initial attempts at using silicon were that people didn't know how to work with it yet. And so even though it was more abundant and cheaper, there's some particular manufacturing process that you have to do to silicon in order to make it as viable as it became.
Yes. How to dope silicon to make it function and produce it at scale as a semiconductor.
And listeners, this is where you should start to get the idea that, especially today, manufacturing these products involves the most advanced process in human history, consisting of layers of innovation in chemistry, physics, mathematics... It's breakthrough after breakthrough after breakthrough all building on top of each other, which need to all happen in the manufacturing process.
So even here in 1964, we're starting to get into the level of complexity where it's some of the most advanced science ever done being applied in an engineering and manufacturing fashion to get even marginal results at 20% yields off the manufacturing line.
And a little preview to fast forward to today, TSMC, they're a contract manufacturer for silicon. That is what they are. TSMC has 40% operating margins as a contract manufacturer. It's not like this is just, there's no technology or R&D. They are one of the most advanced technology organizations in the whole world. There is so much IP just in the manufacturing.
Take out the design, take out the functions, just making this stuff is so hard. I mean, now it involves lasers. We'll get to it later. It's going to blow your mind how this stuff is done. But anyway, so Morris, he's coming up. He's learning. Literally, as this whole industry is getting developed, he's right there.
So a couple years after he gets back from Stanford, he's still rising through the ranks. In 1967, TI makes him a general manager of one of the divisions within the semiconductor business. And that's where he has his next big breakthrough. And this is on the business side.
So Morris notices what they're doing, setting up these new plants for all these successive new methodologies and processes of manufacturing, you know, at this point, integrated circuits and silicon and pumping out these chips. It's super expensive to do this. Super cost capital intensive.
So what TI and everybody else in the industry did when they would start a new product line that would use a new fab for chips, they charged a lot of money for it because like, man, they put a lot of money into these things. So right off the gate, you want the latest new hotness in the end products that TI is selling. They're going to charge a lot of money for them. Yep.
Morris realized that he's like, That's not actually optimal to do that because as evidenced by his first big win at TI with IBM line, there's a learning curve to getting the yields right and learning how to manufacture a new process. And in the beginning, you're going to have really low yield. And so what you want, ideally from a
fabrication perspective is you want to have a ton of volume from the get-go. As soon as the plan is online, you want to be running at max capacity so that you can, A, learn as fast as possible, get yields up to the profitable levels, and then you want to still be running at max capacity as long as possible because you already spent the fixed cost to make the plan.
Basically, you always want max capacity. So when you started out by pricing so high, you kept demand low and you weren't able to get up to capacity fast enough.
It's almost like they didn't realize the benefit of the potential operating leverage that they had because they were just passing their exact economics onto their customers and saying, you basically have to pay us for us to do all these fixed costs. And then you'll get all the benefits of how cheap it is to stamp it off the press every time.
Whereas what they really should have been doing is saying, we will make an investment. We'll eat the cost of having to spin all this up. But boy, are we going to be super profitable on every chip that comes off the line.
Yep, totally. So Morris is thinking about this. He hires BCG. And they come up with the idea of actually pricing low to start to drive this volume and speed up the yield curve. And then also the side benefit of that is if they're pricing low and everybody else is pricing high, they're going to grab a ton of market share and probably keep that. Paying consultants. I know.
Well, so here's Morris's quote about this. He says, this was in the late 60s and Boston Consulting Group was a very small outfit when we did this. And we used loads of data, a lot of theory, and a lot of effort. The result was so-called learning curve pricing. So start low. and then continually automatically reduce the price every quarter, even when the market did not demand it.
And this was a very successful effort, even though it was somewhat controversial. A lot of people thought we were being foolish. Why would you reduce the price when you didn't have to? But we did it because we believed in it. And indeed, our market share just kept expanding.
That, combined with other strategies, made the TI integrated circuits business the biggest IC business in the world and also the most profitable. This is right when Intel's getting founded, so screw Fairchild, screw National, screw Intel. TI is kicking all of their butts, and it's thanks to Morris.
Interestingly enough, the reason I always thought that Fairchild was so successful in those days was out of all the defense spending and research that was being done at Stanford, the government as a customer. But is Texas Instruments playing in that ecosystem at all? Good question. Probably.
I mean, I think this is a case of the rising tide is floating all boats. Yeah. Fairchild's killing it. Intel's killing it. National's killing it. TI's just killing it bigger than anybody else.
All right, listeners, this is a great time to thank our longtime friend of the show, Vanta, the leading trust management platform. Vanta automates your security reviews and compliance efforts. So frameworks like SOC 2, ISO 27001, GDPR, and HIPAA compliance and monitoring. Vanta automates all of these for your company and makes them fast and simple.
Plus, Vanta is always connecting and monitoring continuously, not just a static point in time.
It's funny. When we first started working with Vanta four years ago, we thought, oh, this is one of those great acquired universe products that lets you focus on just making your beer taste better. In other words, customers can focus on what differentiates their product and outsource things that don't to free up time and resources.
But over the last few years, their product has advanced so much that it's not just Vanta does that for you. It's now actually Vanta does that better than you could on your own. Without a continuous real-time monitoring system, there's no way you could do this, that you could give your vendors and customers this level of confidence and trust.
There's even data now on this too. Between the ability to enable new revenue and the time savings, a new IDC white paper found that Vanta customers achieve $535,000 per year in benefits, and the platform pays for itself in just three months.
Vanta is the one single software pane of glass that connects to all of your services via APIs and eliminates countless hours of manual work for your organization. They now have AI capabilities to make this even more powerful, and they integrate with hundreds of external tools. Plus, they let customers build private integrations with all of their internal systems.
So whether you're a startup or a large enterprise, if your company is ready to go back to making your beer taste better and automate compliance and streamline security reviews like Vanta's 9,000 customers around the globe, including Atlassian, Quora, and Factory, head on over to vanta.com slash acquired and just tell them that Ben and David sent you to get $1,000 of free credit.
That's vanta.com slash acquired.
Thank you, Vanta. So on the back of this, Morris gets promoted to VP at TI, one level below the CEO, running the entire semiconductor business. That happens in 1972. And he becomes the obvious leading candidate to be the next CEO of TI, which he's like, yeah, I want to do that. I'm focused. This is what I love. This is my goal.
This is why I've been going to the bar for three years reading a textbook. Exactly.
But... Might be fair to say history turns on a knife point. Things don't entirely go as planned. Three different viewpoints, as far as I could identify, on what happens next to Morris at TI. He does not become the next CEO, obviously. Viewpoint number one is simply and probably fair that he was just discriminated against because he was ethnically Chinese.
Although at this point, I'm pretty sure already he was an American citizen. But anyway, and he got passed over for I have no evidence for it, but not be at all surprised that that was part of what was going on. Yep. So that's one. Two, second point, which Morris totally acknowledges, TI was a really big company.
The semiconductor division, he had made it probably the most successful and the most fastest rising division within the company. But you mentioned calculators. They were starting to launch the consumer products division at this time. And so in 1978, so six years, he's running the semiconductor division as VP. They move him over to VP of consumer products in 1978.
Because this was a big new strategic initiative and it wasn't going super well. And they're like, oh, Morris is a great manager. He can fix this and turn it around.
Different set of competencies, though. I mean, you need like marketing. Yeah.
Here's Morris's quote on this. Mark Shepard, then chairman and CEO of TI, agreed with the prevailing wisdom at the time that a good manager could manage anything. In this case, I think he was wrong. I found the consumer business to be very different, like you were saying. The customer set, completely different. The market, completely different.
And what you need to get ahead in that business is different too. In the semiconductor business, it's just technology and costs. In consumer, technology helps, but it's also the appeal to consumers, which is a nebulous thing. Not Morris's strong suit, or at least not anything he's trained in. Yeah, that makes total sense.
So in 1983, five years after he gets moved over to take over the consumer business, he hasn't turned it around. It's still struggling. He gets demoted to, quote, head of quality and people effectiveness, which is pretty much a slap in the face. Like this dude built your semiconductor business.
Is this when he says he was put out to pasture? Exactly.
So that's number two. Here's number three. I found some evidence on this. It's unclear to me how much of this is Morris's fault versus his successor's. But while Morris was definitely responsible for making TI Semiconductor a powerhouse... At some point towards either at the end of his tenure running it or under his successor, they totally dropped the ball.
And this is when Silicon Valley in California takes over. So in the mid-70s, the semiconductor industry transitioned over to the metal oxide process, the MOS. You ever hear about MOS, MOS semiconductors? Yeah, the precursor to CMOS. Exactly. So that happened in the 70s. And TI... Again, they had the best engineers. They were well-positioned to lead this transition. They didn't.
And actually, most of the talent within TI that were the ones that led the industry transition to Moz left, including probably most prominently a guy named LJ7, who left and founded a company called MozTech. And then he later became a semiconductor venture capitalist and founded Seven Rosen Ventures, which was one of the early VC firms. So he was a TI guy. And he left.
And, you know, the culture at T.I., as shown by Morris's experience, was not like Trader Estate, Silicon Valley, you leave. It was like, you're a company man, you know, you stay at the company. Right. So Motorola poached a whole bunch of Moz engineers from T.I., And it all kind of fell apart, culminating in the biggest huge loss. Now, this is really history turning on a knife point.
In 1980, so Morris has already transitioned to consumer products. IBM puts out a secret RFP bid proposal for a secret project that they're working on. This is 1980 by a new group based out of Boca Raton, Florida. Do you know what I'm talking about, Ben?
I have no idea.
Some listeners might know what I'm talking about. This is the secret project. This is the RFP to be the microprocessor, the CPU for this secret project. The IBM PC.
Ah, okay. That was at a Boca Raton?
Yeah, it was a secret project, like a Skunk Works division of IBM. Oh, wow. To build the PC, which was a big, you know, IBM was the mainframe company. Right. We're going to build a personal computer. So Skunk Works project and TI, you know, a couple of years earlier under Morris would have been an obvious candidate. Remember, he had the relationship with IBM going all the way back.
TI probably should have been the processor chosen. Instead, of course, it was Intel. I think it was the 8088 that was chosen for that first one. Wow. Boy, did that set things in motion. Well, then the architecture standardizes on x86 and boom, there goes the whole next generation of computing away from TI over to Intel.
The sort of family of IBM with Intel processors and eventually running Microsoft operating systems. Yeah, but then all the IBM clones all running Intel processors. Okay, so this is really where... That is a major loss. In the highway of history, TI accidentally took the off-ramp there.
They did. Now, okay, is that Morris' fault? Is that not Morris? I don't know. Certainly, the culture at TI was, we rotate you around, you're going to fix consumer. He didn't fix consumer, but couldn't. And then this, like the semiconductor powerhouse took an off ramp, as you say. All that, like his career at TI is basically over. So he was the rising star.
Everybody thought he was going to be the next CEO. And at age 52 in 1983, after he stays a couple of years being the head of whatever.
Yeah, he was something staff. Yeah, yeah.
He just resigns. And he's like, wow, well... I guess this is it. My career at TI, 30 years, done. He's still regarded super highly in the industry, though, in the semiconductor industry. So people start calling him with opportunities. And he wants to be a CEO. I mean, that's what's on his mind. He wanted to be CEO of TI. That didn't happen. He wants to be CEO.
But he whittles it down to two opportunities he's going to consider. One is to go to a competitor called General Instrument, which people may have heard of another one of these old chip companies.
It was based in New York in Manhattan in New York City, actually, to go be their COO, the number two there, with an understanding that like, hey, if things go well in a couple of years, you'll replace the CEO, become the CEO there. Or to become a venture capitalist. Really?
Yeah.
No way. I don't know where or how. I couldn't find that out. But he was weighing these two opportunities. The VC idea is going to come back up in a big way in a second. But obviously, he goes with General Instruments, GI. A, his dream is to be CEO. B, he's got this chip on his shoulder from the way TI ended. So, great. So, he goes off to New York. He leaves Texas, lives in Manhattan.
Things are going to work out at GI. The thing, though, is GI had a very different culture than TI. You know, TI was this research, build, develop technology, push the ball forward. GI is this New York based. They were almost like at the time, like a proto tech private equity firm.
Their strategy was they just acquired lots of different semiconductor businesses, either independent companies or divisions from other companies.
Try and integrate them.
No, they would acquire them. They would get these business units into good shape, and then they'd sell them again.
Oh, really?
Yeah. Literally, they were like a financial engineering firm, basically. Definitely not Morris's cup of tea. So he only stays there a year. It's clear that's not a good fit. So he resigns again. So within less than 18 months, he's had two major setbacks. Basically, his dream is over. Here's the quote from him.
He says, after these two setbacks at TI and GI, I did not think that my aspiration to be the CEO of a major US company was in the cards. Well, turns out he was right. He was not going to be the CEO of a major US company. So how do we go from this dude in his mid-50s, former rising star, now washed up, from that to he's in Taiwan, he's CEO of TSMC, I don't think you could ever script this out.
I think this is probably the most unique. Every founding story is unique, but I think this might be the most unique founding story we've had on Acquired so far. So back when Morris was at TI, when he was running the semiconductor business there, he went over to Taiwan a couple times to talk about building a manufacturing plant there.
TI would own and build the manufacturing plant, but outsource to Taiwan. Not like a TSMC style business, like it was a TI plant there. Anyway. He had no connection to Taiwan. Remember, he's Chinese. He's not from Taiwan. People are like, oh, Morris, you went back to Taiwan. He didn't go back to Taiwan.
Yeah, he talks about how Taiwan was a strange land to him when he first got there, that it's not going back. So it's not the land is a strange place to him. But if he is going to call someplace home and return there, is it the People's Republic of China?
Well, I think he would say at this point, it's America. He's been in America.
That's a great point.
He's a US citizen. Well, I don't know what he would say. It's complicated. So anyway, so he had met a bunch of government officials in Taiwan when he was talking about building this plant over there. And that was back in the 70s. Now we're in the 80s, mid 80s. Taiwan at this point, it's a manufacturing nation. You know, they have no IP, no technology there.
Okay, the quote's great. All right, so this is Morris. We had no strength in research and development, or very little anyway. We had no strength in circuit design, IC product design. We had little strength in sales and marketing. And this is, of course, referring to Taiwan as a nation. And we had almost no strength in intellectual property.
The only possible strength in Taiwan that we had, and even that was just a potential one, not an obvious one, was semiconductor manufacturing, wafer manufacturing. And so what kind of company would you create to fit that strength and avoid all the other weaknesses? The answer was a pure play foundry.
Yeah, I mean, that was Taiwan at the time. So to give you a sense, the average gross margin of a Taiwanese company at this point in time in the mid 80s is four to five percent. Zero four to zero five percent gross.
Before you even have overhead operating costs.
Yeah, I mean, it was like, you know, if you grew up around when Ben and I did, you know, sort of born in the 80s in the US, you see made in Taiwan on everything. Like, you know, Barbie dolls, toys, clothes, everything was made in Taiwan. You know, now it's made in China or made in Vietnam or elsewhere. But made in Taiwan was super low-end physical manufacturing stuff.
Yeah, and to way pull forward the seven powers section, as Hamilton Helmer would sort of explain, if your margins, particularly your gross margins, are only 4% or 5%, you're in an industry or a business where all the profits are arbitraged away and everyone's just raced to the bottom on prices and no one's able to build any real enterprise value because everyone's just out competing each other for pure commodity.
Right. I mean, four to five percent gross margins.
People used to hammer on Amazon, I guess, for being like a low gross margin business in the like 40 percent. Anyways, I can't even imagine running a company with that level of gross margins. So the Taiwanese government, though, they want to come up in the world. They're like, this is where we are now. This is not where we want to be. So they knew that technology was the way.
And so they had decided back in the 70s that they would establish an initiative called the Industrial Technology Research Institute, or ITRI. And the goal was for it to become like the Bell Labs of Taiwan to do some tech transfers from the U.S. and elsewhere and home grow some real technology technology.
businesses in Taiwan so that, you know, maybe they can lift businesses out of poverty there, at least.
And so Morris wasn't going to Taiwan to start TSMC. No. He was being recruited to ITRI.
So one of the ministers he had met, a guy named KT Lee, who because of this, he would also become venerated in Taiwanese history. He's known as the father of Taiwan's economic miracle, literally because of this.
Wow.
He recruits Morris to to come over and run E-Tree, be like the head of Bell Labs Taiwan, essentially. And this is like a ridiculous thing for Morris to do. He had been, you know, captain of American semiconductor industry. He was put out to pasture at TI, but at least he was still at TI. And then he was COO at General Instrument. He's going to go over to Taiwan and run like a research park there.
Like what? And like every time someone starts something like this, it doesn't go well. A government top-down innovation mandate from a country that's not a world power tends not to turn into a gigantic economic success.
This is like all the countries and cities and the like that are like, oh, we're going to build the next Silicon Valley in XYZ. And we're going to recruit some former Silicon Valley person to come do that. And it's going to work. Probably not going to work. So everybody tells him not to do this. All his former colleagues, his wife at the time, tell him not to do this.
His marriage was actually falling apart, maybe in part because of this. And, you know, he's had all these experiences. He's like, you know what? I just, I need to change the scene. I got to get out of here. So he takes the job and he figures, you know, it's going to be cushy. This is like a soft landing.
He thinks about this as like the pseudo retirement he's going into.
So here's his quote. By then, I was financially pretty secure. I was not rich, but you also have to realize that the standards of wealth were much lower back in 1985. And he's going to live in Taiwan where, you know, Corporate magnets have 5% gross margins.
But still, in absolute standards, I was financially secure, which meant that I could live according to the way I desire, which was actually pretty modest, for the rest of my life without having to earn a living or a salary. This is retirement.
He also makes a joke, I remember, after that about how, by the way, interest rates were higher back then, so that was much more achievable on less principle. Yeah, totally, totally.
So 1985, he goes over, he takes over as president of E3. It's kind of a culture clash. So this is retirement for Morris, but he's still coming from this hard charging industry. All of the employees of E3 retire. are people in government jobs in Taiwan and government jobs, not even, you know, in like a democracy because Taiwan is like, it's under martial law.
I think that it just ended, you know, this is not the same, you know, this is like jobs for life. You're a government official in a non-democracy type organization. Morris says back then they considered me a foreigner who suddenly became their boss. They were scared of me and they were right to be scared of him.
So there was one thing, though, that the government had done right before Morris showed up, which was they had successfully negotiated one technology transfer license in the semiconductor industry from, did you find out what company this was? This is probably what they were trying to negotiate with TI for.
I do. It's a three-letter acronym. Oh, yeah.
Yep, yep, yep. We haven't talked much about it on this show, but this is another talk about captains of American industry. RCA.
RCA, that's right.
Yeah. So RCA had a semiconductor line and the government in the 70s, the Taiwanese government, had negotiated a tech transfer agreement.
But this is like 10-year-old semiconductor technology, right?
This is not like the latest generation. No. TI and Intel and everybody at Fairchild, they're national. They're leading the way. They're at the bleeding edge of semiconductor manufacturing process. RCA was already at least a generation behind. By the time it actually gets onto the ground in Taiwan, they're two and a half generations behind the leading producers.
So it's like the only thing that you could do with that is... Super low-end stuff. Right. There are some category of goods that don't need a fast or the latest processor. Totally.
Even today, when TSMC or Intel or Samsung or whoever builds a fab... The leading edge fabs, they produce the leading edge stuff for a while and then the new generations come on. They don't shut down the old ones. It's just chips that don't need the same bleeding edge performance.
They keep getting made on the old ones. And often that's automotive or now what we think of as IoT, but the stuff in your smartphone obviously is the... The leading edge. Yeah.
So the government, E3, does actually spin out a company using this old RCA technology that would be called UMC, United Microelectronics Corporation, not a technology leader. It actually does okay in the long run. They would later spin out.
their own chip design business so umc was doing both fabrication for third-party clients and designing some of their own chips with the fab that they created they spin out their chip design business later that becomes media tech oh no way yeah which is a 50 billion dollar company today so like you know the government did pretty good like this is pretty good totally we're doing
And when Morris arrives because of this, he's not starting from a standing start. It's not good, but there's some assets.
They've acquired IP. They've created a company. There's a paved path.
So he gets to work at E3. He's working on all this. He's transforming the organization into a high-performing organization. And then all of a sudden, out of nowhere, KT Lee comes back to him and is like, hey, great. You're running our Bell Labs. You're running E3. Now I want you to start a company. And Morris is like, uh... And KT's like, yeah, yeah, yeah.
I don't want you to have somebody else in Egypt do it. I want you, Morris Chang, to start a new semiconductor company here in Taiwan, and I want you to make it into a global leader. Morris is like, um, okay. He doesn't say this directly. Well, he's got a great quote I'm going to say in a minute. But again, remember, this is not a democracy in Taiwan at this time.
Morris has also won his third job in three years. Yeah, he doesn't need a salary to survive, but this is kind of the end of the rope for him. If he gets fired here at E-Tree, he's legit done. Yeah. Done, done, done. So he kind of doesn't have a choice here. The quote, this is so Morris, so great. He says, it was like in the movie The Godfather. It was an offer I couldn't refuse.
And I do think the implication was go start an Intel or go start an IBM. It wasn't go start the very first pure play foundry. Yeah. Lee had no... I mean, he's the government. He's a minister. Right. He was like, go start a semiconductor company and make it a world leader. Right. Those semiconductor companies, they do really well. So go do that. And that's, of course, when Morris says, okay...
I'm being told I should do this. I have some latitude I can take and some liberties I can take on how I do it. And the quote that I read earlier about evaluating exactly what type of semiconductor company should I start, that's how he sort of informs the business plan.
So, Lee is like, all right, good. We're capisce. We're clear. Come back to me in a week with a business plan. Tell me what you need. And we're gonna make this happen, right? So, we're just like, okay, a week. All right. And then like a day later, Lee supposedly is like, I'm gonna need you to come in on Friday. So, you got like three days. Yeah.
You know, they say necessity is the mother of invention. And yet these three days are what creates the, you know, now ninth most valuable company in the world. Morris comes up with this brilliant idea to create a pure play foundry company to be a contract manufacturer.
Sounds genius today in hindsight, as Steve Jobs would say, you know, it's easy to connect the dots looking backwards. But at that time, was this a good idea, David?
Well, no. The answer is no. You know, like we've said all along, all the chip companies, all the, you know, American and European and, you know, Japanese, all the leading semiconductor companies, they made their own stuff, you know, and there was some sharing of production and some companies were emerging that were borrowing production from the big guys.
There's a great quote right around this time from Jerry Sanders, who is the co-founder and CEO of AMD. And he famously said in the mid-1980s that, quote, real men have fabs.
That's right. Oh, what a quote.
So ironic because in the 2000s, AMD would spin out its fabs and go fabless. Global foundries. Yeah, into global foundries. But yeah, this was not an obvious idea. If you wanted to be a real semiconductor company, you made your own chips. And the idea was like, this isn't like manufacturing Barbie dolls here. This is real technology.
You need to control it soup to nuts. And already at this point in history, I mean, this is an important point to make because I didn't realize this coming in where I thought, wow, Apple really outsources their manufacturing. They outsource some of it to TSMC and some of it to Foxconn. And maybe... some of those people will start to do each other's work. No, this is a completely different thing.
Assembling an iPhone is completely, completely different than taking a brand new design for a next generation chip and manufacturing that chip. One is manufacturing and one is alchemy. The alchemy can only be done by alchemists. I think even here in the late 80s, we're already at the point where it's manufacturing broadly, but yeah, it's not like, well, I got a factory.
No, no, no, no, no, no. Far the opposite of that. We said a minute ago, this is a bad idea. So Morris says, now, however, there was one problem with the pure play foundry model, and it was a fatal problem. It could be a fatal problem, which was, where's the market? He sounds like Don Valentine here. Where's the market? Show me the market.
This whole idea, it was really a solution looking for a problem.
And of course, the solution being that all we have is manufacturing capability here. So let's start a company that just manufactures. And it's like, you're looking around like, okay, who's definitely going to manufacture?
Yeah, Jerry over at AMD is like, real men have fabs.
Yeah.
You know, there are no real startups. I mean, there are startups, but all these startups are building their own fabs. Like, nobody wants to do this. So nonetheless, you know, he has to start a company. He's literally got a gun to his head.
But he does have the core insight here. It's interesting. These companies don't exist yet, but Morris has reason to believe that people will want to start fabless chip companies and that they will need a foundry to fab those chips. And so he says... When I was at TI and General Instrument, I saw a lot of integrated circuit designers wanting to leave and set up their own business.
But the one thing, or the biggest thing, that stopped them from leaving those companies was they couldn't raise enough money to form their own company. Because at the time, as we were just saying, real men, it was thought that every company needed manufacturing, needed their way for manufacturing.
And the most capital-intensive part of a semiconductor company, of an IC company, does the manufacturing. And so I saw those people wanting to leave but being stopped by the lack of ability to raise a lot of money and build a wafer fab.
Totally, right? But those companies, if you build it, they will come. They haven't started yet.
They haven't come yet.
They haven't come yet. So he knows what the long-term market is going to be, but he's got to find the short-term market. He needs some real politic here. So what's that going to be? So he says, well, maybe I can go around to the big guys.
They've been doing, you know, just like my first thing back at TI, they've been doing some line sharing, you know, for either new products that they need excess capacity for or for older products that they need to transition some fabs, but they still need to make components. Maybe I can take some of that off their hands.
And so he goes around and he talks to Intel, he talks to TI, he talks to everybody in the industry and they're like, yeah, he talks to Motorola, like, sure, whatever.
fine and the government had told him we know it's gonna take a lot of money to set up a fab we're good for half of it but you gotta go raise the other half of it and we want you to raise it from like an intel or a ti you know somebody who's gonna be your first customer and that they're gonna be bought in so he does the rounds he goes and talks to everybody he gets meetings with intel he gets meetings with ti they're both like you know morris we like you but no
So he said the last-ditch effort, and he has a meeting with Philips, the Dutch company. They have a semiconductor business. So Morris, he has a great quote about this. He says he would describe Philips as the first rung of the second Raiders in semiconductors, but they were the only interested option. So they put up 28% of the capital. Government puts up 50%.
It ends up being $220 million in total.
The 110 is probably a lot more than what the Taiwanese government thought they were going to be buying here.
And then literally the premier of Taiwan, like the head of the government, has to then go around to all the other business leaders in Taiwan and like strong arm them into investing the rest of it. The other, what is that? 22%, I guess.
Yeah. We also should say, remember that Philips was a Dutch company because that's going to come into play later.
I don't know how that's going to come into play.
Yeah.
Putting a pin in Dutch. Okay. We got a surprise coming. I'm going to be surprised here. We're doing a real time, doing a live. This may be the craziest part about the whole TSMC founding story. I'm 99.9% sure, Ben, you do not know this. Do you know what the pre-money valuation was on TSMC?
No, I couldn't find that anywhere.
It was $0. $0. Morris Chang got no equity in the company.
Zero. So 100% of the company was owned by... The investors.
50% by the government and the other 50% were owned by the investors. Morris got nothing.
And just got to keep his salary. He was a government employee.
Wow. There by the grace of the government. Oh, my God. Isn't that unbelievable? Like, this is so the opposite of Silicon Valley. How is he worth $3 billion today? Well, what he did as TSMC started to work, he basically put all of his money into buying. He bought his own shares in the company.
I don't know if it was privately before they went public on the Taiwan Stock Exchange in 1994 and then the New York Stock Exchange in 1997. But yeah, he put basically all of his excess cash flow into buying TSMC shares.
Oh my God. Isn't that wild? So the government owned 50% of the whole business in...
And you can see their perspective, too. They're like, hey, we hired you to do this. And then we told you to do this. You are our foot soldier. We are the mafia.
Wow. Yeah. Things had really not gone well in his career that he was willing to take that deal.
Yeah. Crazy, right? And OK, before we go on in the TSMC story. We need to have two real quick sidebars. Yeah. So it was 1987. Yep. When TSMC gets officially stood up, they raise the money at a $0 pre-money valuation. Do you know what other company, other big thing happened in 1987? We have covered it on this show in the chip world.
Is this the founding of ARM? Yes, it is.
Yes. Yes. ARM. JV between Apple, Acorn, and VLSI Logic, which was the sort of manufacturing partner. They were an A6 company. That's a whole other sidebar we're not going to get into. But yeah, 1987.
What a year. Brand new, unconventional instruction set architecture that's totally different than the x86 stuff that the whole industry and world seems to have standardized on at this point.
The Annis Mirabelis for the semiconductor industry.
And useless, right? It's in 1987. It's hamstrung. It's very few instructions. PCs are always plugged in. So what do we need a low power chip for?
This thing's pathetic. Real men have fabs and real men use power. Okay, so that's sidebar number one. ARM gets started.
Okay.
Now.
I was wondering, I don't actually know the relationship because obviously today a huge amount of volume of TSMC's manufacturing is making chips for iPhones, which since the outset has used ARM.
Chips that are using all mobile devices, iPhones and Android, all of which are ARM, and lots of servers that are ARM.
So presumably there's some relationship coming between TSMC and ARM. Well, they're really close partners.
I mean, the way now this stuff is so integrated. Yeah. The architecture companies like ARM, the design, the EDA companies like Synopsys, these guys are all deeply, like the engineering is all deeply in bed with one another.
Okay, so you mentioned EDA. I'm going to take your sidebar and I'm going to... You're going to raise me. Yeah, I'm going to raise you one more sidebar. So listeners, we're two clicks out here. So this is a pretty good point to talk about how the value chain went from one company that... created transistors, and then they designed the chip, they manufactured the chip, they marketed the chip.
Here's how the value chain looks today. And I think you've already alluded to, I think in the 80s, it already started to look like this. First, there's EDA. There's electronic design automation. This is the software that professional chip designers use to do their work. So Synopsys, I think Cadence is another big one.
Yeah, Cadence. They're the two leaders. So that's like, I don't know, Excel or like Figma for chip designers.
Yeah, that's what they use. Yeah. So that's category one of four. And of course, as you can imagine, the software to design the chips probably has to be very aware of the manufacturing capability of who's going to be manufacturing the chips. But let's put a pause in that for a second. So then, of course, there's the fabulous chip design companies. So today, think Apple, Nvidia, Qualcomm. Broadcom.
Eventually AMD after they stopped being real men, apparently.
Tons of innovative new startups now, like Cerberus.
PA Semi. Tesla. Before Apple acquired.
Yeah. Oh, PA Semi is coming in a sec.
Okay. Okay. So you've got the EDA companies that are making the software, the Fabless companies that are designing the chips using the software. Then third, there's one company that we have not talked about yet, one component of the value chain. And these are the people that manufacture the machines that go into the factories that the foundries operate.
There's actually one between, well, no, actually I think above EDA, there is one more part of the value chain. There's a fifth, which is IP. So that's all like ARM. Oh, right. Yeah, like architecture, IP. There's actually a ton of companies now that do just straight up IP. And I thought before this episode, these were like, oh, just shell companies that sue one another about IP. It's not that.
Systems are on a chip now. So like everything is on one chip, basically. You need a USB functionality in your chip. You don't need to design it. You just buy some IP off the shelf. So there are companies that do that.
Yep. Okay, so that's a good point. So that's our fifth sort of like IP, they own the instruction set architecture, they kind of create the general rules that you're playing by when you're designing a chip such that whoever's writing the compilers knows what assembly language they're targeting that can then operate on the chip that's going to be designed.
So we covered the EDA, we covered the IP, we covered the Fabless companies. There's somebody before we get to the foundries, which is the equipment manufacturers that sell to TSMC. So more historically, you've got Lam Research, you've got Applied Materials in the US, you've got Tokyo Electron in Japan. But today, I just want to give everyone a taste of this, and then we'll get more to it later.
There's a company that is also Dutch-based called ASML, which was originally ASM Lithography. And lithography is marginally in scope for this episode. There's a whole thing we could do on the magical process that is lithography.
Taking me back to my high school photo lab.
Right? Yeah. And the L, you know, is lithography. So the company was originally called ASM Lithography. They make the most advanced chip manufacturing machines in the world. They're the only company that makes them. They're located still in the Netherlands. Their biggest customer is TSMC. And this is where I want to bring it all the way back around.
And we, of course, will talk about the magic that is these machines later. It was founded in 1984 as a joint venture between Advanced Semiconductor Materials International, ASM Lithography, and Philips. Oh, wow. I did not know that. So that- That's crazy. Is the beginning of the relationship between- TSMC and their equipment provider.
And what a strategic point. I mean, because, well, it's TSMC's insane capital operating cash flow production that enables them to spend CapEx above anybody else that allows them to buy more ASML equipment than anyone else. But that relationship, wow. I mean, these machines, we'll get into it later. It's going to blow your mind what this stuff does.
All right, listeners, this is a great time to thank our big partner here at Acquired, JP Morgan Payments. When we told the team over at JPM that we were remastering our TSMC episode, they smiled and said, oh, we have the perfect thing for this episode that most listeners won't know about, but is super important to how international companies accomplish global trade.
Which, of course, that is like catnip for us. So we were like, well, tell us more.
Yes. J.P. Morgan has a whole trade and working capital group that helps companies manage risk and access liquidity for global trade, in addition to, of course, facilitating the actual payments to drive growth.
Yep. Imagine this scenario. You're a global auto parts manufacturer, and you build up inventory of a key car engine component because it requires a super complicated production process. But then you have to wait for the final demand forecast from the actual car company, your customer, before you can ship units to them.
But in the meantime, you have to secure credit lines and manage your own supply chain.
Sounds complicated. Enter JP Morgan's trade and working capital group. Not only is JP Morgan Payments the leader in all the payments products we've shared over the last year that drive growth for your business, they also have 150 years of global expertise to help businesses navigate issues like this. This legacy dates back to 1799.
So they literally have centuries of leading the industry through times of both economic growth, but also instability.
And as you all know, you can't time the market. But if you intend to be in business for decades, you know you're going to exist through all sorts of different geopolitical, macroeconomic, and supply chain climates, just like TSMC has. So make sure you've got a great payments partner that's also prepared for anything as your business grows.
JP Morgan really is a behind-the-scenes powerhouse that helps all types of businesses and industries facilitate payments, mitigate risks, access liquidity to accelerate cash flow, and so much more.
Our huge thanks to JPMorgan Payments. You can learn more at jpmorgan.com slash acquired or click the link in the show notes and just tell them that Ben and David sent you.
Okay, back to my second sidebar. Also going to be worth it. P.A. Semi. We did an episode. Wave is like episode 20 something.
This is like when Acquired was a very different show when it was actually about small acquisitions. Totally.
Yeah. So I don't know that we actually covered this, but I uncovered in the research for this episode. Do you know the origins of P.A. Semi? No, I don't. Okay, so Arm, my sidebar number one, you know, 1987, also created, right? They're just an IP design company, like I was saying in your sidebar. Um...
It's like Inception over here.
So they just license out the ARM architecture to other companies that then design using the ARM architecture. One of their original licensees was DEC, Digital Equipment Corporation, like OG, way back in the day. Yep. So they took the ARM architecture and... And they tuned it for performance. And they called what they did at Deck, their version of Arm that they created, StrongArm.
And that product line within Deck... would later be acquired by Intel, of all places. Okay. Crazy why Intel acquired an ARM architecture. Right. They're the x86. Yeah, they remarketed it as Xscale. I think they ended up shutting it down. Huh.
So a bunch of the core engineers on the team, like the DEC team that had been working with ARM from back in the day, and they're like, we just got acquired by Intel. What the hell? Like, you know, screw this. We don't want to go work for Intel.
There's no interesting, flourishing alternative architectures at Intel.
Yeah, we're ARM engineers. We can go start our own company.
That's PA Semi?
That's PA Semi.
Ah. And of course, the underpinnings of all of Apple's chips today.
Totally. So the lineage of all of Apple Silicon, probably the most valuable, defensible part of Apple today in terms of technology, was Arm & Deck to Intel to PA Semi to Apple.
Whoa. That's wild. I don't think I ever knew that. So you can trace Apple Silicon all the way back to Apple.
Yeah, because ARM was a JV with Apple. Crazy. With Intel and DEC in the middle. Wow. Okay, so back to TSMC. The short-term market is Morris basically begs all of his old colleagues in the US and European and Japanese semiconductor industries to just give the dregs to TSMC. And it really was the dregs. So here's Morris on what this was.
The IDMs would let us manufacture their wafers only when they didn't have capacity or when they didn't want to manufacture the stuff themselves anymore. Now, when they didn't have the capacity and asked us to do the manufacturing, then as soon as they got the capacity, they would stop giving us orders. So it wasn't a stable market.
So it wasn't actually a thing they wanted to outsource.
No, it was just like they didn't have the capacity, so they needed some extra excess space. But then when they got the capacity online, they took it away. Right. And then the chips that they gave us that they didn't want to make anymore. Well, the reason they didn't want to make it was it was losing money. So they basically were just transferring their losses on producing these chips to TSS.
So how did they get out of this? So Morris continues, the conventional conclusion at the time was that there was no market. That's why the pure play foundry idea was so poorly thought of. What very few people saw, and I can't tell you that I saw, was the rise of the fabless industry. I only hoped for it.
And then as you said, but I had better reasons for hoping for it than the people at Intel, at TI, and Motorola because I was now standing outside. When I was at TI and General Instrument, I saw a lot of these IC designers wanting to leave, start their own businesses, and the constraint was setting up their own fabs.
So like, yes, he saw that at TI, but remember, he had been considering becoming a VC instead of going over to E-Tree. So this is the ultimate end around. He becomes essentially like... the world's best semiconductor VC.
He takes an index out on the whole future innovation and entrepreneurship market in semiconductors by becoming the platform that they're going to build on instead of like going and investing in them. Like he enables all of it. He's like the Y Combinator of semiconductors.
Right. Or in many ways, the Tencent. Tencent, of course, also does direct investing. But the idea that you could get distribution through WeChat, it's kind of like it's not distribution, but it is manufacturing. There is a thing that you have to raise 10 to 20 percent of the capital that you otherwise would have needed to raise if TSMC exists.
Yep. And just like Don Valentine, you know, kind of when he left to go join VC a generation earlier, again, it's not VC, it's TSMC, it's building the platform. But Morris is a hero. So all these engineers, they all look up to him. And he knows a lot of them personally. The ones he doesn't know, like who's not going to take a meeting with Morris Chang, right? He almost ran TI. Totally.
Like he did all this amazing stuff. It's interesting because it's like with the incumbents, of course, because they had it in their DNA to be a manufacturer, of course, they wanted to take the most profitable things and manufacture them in-house. But if you actually are betting on all these startups that will never develop... DNA to be their own manufacturer. They never want to take that back.
Yeah. And so Morris is now going out and evangelizing and he's like, all these great designers, like, we're an option for you now. Like you want to leave, you want to start your own company. You don't need a fab. We'll be your fab. You know, it takes a couple of years. For a couple of years, TSMC has to survive on the dregs from the IDMs, the big guys.
But after a couple of years, these startups get going. You know, little companies like Qualcomm, Broadcom, Marvell, NVIDIA. These are all started with TSMC.
NVIDIA was started in 1993, only ever raised $20 million. $20 million. And never opened their own fab.
I believe 100% with TSMC. Wow. Well, maybe they have other foundries too, but the vast majority of their business, Jensen talks about this, Jensen Huang. It took him actually a little while to get on Morris's radar. But once he did, the vast majority of NVIDIA's chips, TSMC makes them. And NVIDIA is what, like a $350, $400 billion market cap company now? It's wild. And only raised $20 million.
It's like the AWS for chip companies. Yeah. Never would have been possible before. Nope. So this is what's super cool. I don't think Morris saw this. Like this even exceeds his wildest dreams. He was hoping for this fabulous market to take off. But this creates this insane flywheel for TSMC. So the fabless market starts to grow, which they're like seeding and enabling it.
As that happens, TSMC's revenue grows.
And because they have 50% gross margins and 40% operating margins, they can take that profit and buy more advanced machinery.
Build more fabs. Advance the level of their technology. Remember, they were starting from behind on technology. Within about 10 years, they catch up. And then they start to exceed everybody else. So as they push the manufacturing process technology forward, they're manufacturing better chips with smaller process lengths.
They're enabling their customers, which are the Fabless companies, to get better and better performance. As they get better performance, the Fabless companies can address more of the market and more use cases. So their existing customers get bigger and new Fabless customers start producing which gives them more revenue, which repeats the whole cycle. And it goes slowly like any flywheel.
It takes a lot of effort and a lot of time to start turning it. But fast forward to now. So in the early 2000s, when TSMC finally caught up, to the bleeding edge level of technology with other semiconductor companies. There were 22 companies that were at the leading edge. I think it was like, I don't know, let's call it 150 nanometer process or something like that at that point in time.
22 and TSMC finally broke into the pack. They were one of the 22. By the late 2000s, it had gone from 22 down to 14 that were at the leading edge. By the mid 2010s, there are six.
It's basically Samsung and TSMC, right?
Today, there are two. A five nanometer process is the current leading edge. Only TSMC and Samsung. Intel has been trying to get there, but they haven't been able to. They've fallen behind. And the next process is going to be three nanometers. TSMC is going to launch that next year.
Which, by the way, just slipped six months.
Ah, interesting. Well, Samsung has already slipped to 2024. Whoa. So very likely in the next process, it's just going to be TSMC.
Which means that you will see that on an Apple slide somewhere announcing the next iPhone, talking about how it's a three nanometer process. They'll take all the credit for it. And TSMC is totally fine with that because their job is not to market. It's to empower their customers.
This flywheel, it's just unreal what happens here. They run the table on the whole industry.
It is interesting. The industry went from vertical to horizontally integrated, where the very best products in the market became horizontally integrated. And it's interesting how I'm trying to figure out what drove that. Because at some point, I guess there's a couple components to it. One is the speed at which Moore's law happens makes it such that you can't be good at everything.
You can't be good at everything from EDA to making the manufacturing equipment to running the manufacturing process to designing the chips. You're not going to write your own instruction set architecture. People did need to break into best of class.
Morris has got this great quote about this that I have in here. So he says, the semiconductor business is like a treadmill that speeds up all the time. If you can't keep up, you fall off. And that's Moore's law. From 22 down to two down to one, even when their competitors are only doing the one thing the TSMC has done, like if you fall behind by a step, you're toast.
Right. And it's because there's this big part of it that you're talking about that hasn't come up on other episodes because we tend not to talk about companies that require a lot of manufacturing prowess. But in order to stay on that treadmill, the number of tens of billions of dollars that you need to be spending into CapEx is going up.
So you need to be enormously profitable so you can build the factories for the next generation.
Yeah. I mean, well, there's two things. So yes, that is a hundred percent true. And the scale of this now, I mean, TSMC just announced they're going to spend a hundred billion dollars in CapEx over the next three years, $30 billion this year, 60 over the next two. And I bet that keeps going up. So that's a lot of billions.
You might even say like, this is so strategically important and people are talking about this. Certainly China's talking about this. The Governments might need to come in with a bazooka of money and create other options because almost all their manufacturing is in Taiwan. It's in this strategically geopolitically challenged location. We need to re-onshore some of this in the US.
China, of course, wants their own. You can't just spend the money and do this. US government could come in and say, we're going to spend a trillion dollars this year to do this. They can't do it because they're We're going to get to powers later, but there's this marriage of scale economies and process power, the TSMC, like in this industry.
There is no amount of money you could spend to catch up next year. You can't because the engineering is so hard and the learning curve takes decades to get to this point. I was listening to a podcast, a Bloomberg Oddlots podcast about this, where they were talking about this and their reporter who covers TSMC is great.
China, you know, they asked the question like the, well, China could just spend a billion dollars and do this, create their own fabs. And they're doing this.
What's the company called? SMIC? SMIC, yeah. Because basically, TSMC seems to have picked a side in the US. And so with a little bit of prodding, I'm sure, from various presidential administrations over the last five years. Yeah.
The guy who covers TSMC was like, they can do that and they are doing that, but they wouldn't know what to do with it. And it's not because they're dumb.
It's the hardest thing in the world to do this stuff. To make the equipment that ASML does and to manufacture the way that TSMC does, it is the hardest thing to do in the world.
Yeah. Anybody else could get all the same equipment from ASML.
Actually, that's not true.
Oh, no, I'm saying even if you could, you wouldn't know what to do with it.
Right.
And it's not because you're dumb. It's they're only like a small number of people in the world that can operate this stuff.
All right, I'm jumping out of my seat here. So I'm going to do the ASML thing now. So the reason that some people can't get their hands on the ASML equipment is because the Netherlands did not renew their trade agreement with China. Also, likely it has been reported that probably that is because of U.S. prodding to say, hey, these pieces of equipment you're making seem pretty specialized.
You're the only people in the world who can do it. And it makes the most cutting edge semiconductor manufacturing technology. Maybe let's not sell that to SMIC in China. And so they're not doing that. Now, you might say, oh, come on, how hard can this stuff be? Well, these machines... Tell us, Ben, what these machines do.
Well, first of all, they cost $200 million for a machine that makes the chips. And that's going to go up to like $300 million. And by the way, on a lot of this, we have a lot of thank yous for John Bathgate and Britton Johns from the episode of The Knowledge Project that they went on to talk about a lot of this stuff. It takes four 747s to ship one of these machines.
So you buy one, your TSMC, you buy one and it arrives. And of course, the 747s, then there's a crew of ASML employees on site, not only to assemble it, but then to help you run it. So like you mentioned, these companies are deeply integrated with each other to pull this off.
Okay, so what does running it mean?
What do these machines do? Okay, so it becomes exponentially harder to manufacture chips the more dense they are. So David, you mentioned that 150 nanometers or so from several years back, and we know now that the M1s are made on this 5 nanometer process. Well, the wavelength of white light, of regular light, is 193 nanometers.
Ooh, that seems like a problem.
Well, it's certainly wide. But, you know, we're humans. We come up with clever solutions. We can solve this. And so you shoot it through a lens. And maybe you shoot it through some water. Like a laser. Well, not yet. But even that really only gets us to like 11 nanometers.
So how the heck are we supposed to make these chips where the transistors are ostensibly only five nanometers apart when what we've done to date, shooting through lenses and shooting through water, gets us to 11 nanometers? Well, okay, so this is crazy. You have to create a plasma. So what they do, and this is called extreme ultraviolet light, or EUV. This is a process that is just wild.
On one side of the machine, you drop molten tin. On the other side of the machine, you then hit it with a highly specialized laser. You perfectly pulse them. It explodes into a plasma, which creates extreme ultraviolet light. Now, of course, this is hard enough to do, as you can imagine how that might work. But you actually have to do that 50,000 times per second.
Yeah, and what I read is that the accuracy with which that laser needs to hit the drop of molten tin is more precise than the calculations to send the Apollo missions to the moon. And you got to do that 50,000 times a second.
Unbelievable. Now, of course, think a little bit more about this. Well, wait a minute. That wavelength is so small. We're going, you know, shy of 11 nanometers here. We're going to 5 nanometers, 3 millimeters, that actually it is absorbed by all known mirrors, which we're used to reflecting light. But they don't reflect this light because the wavelength is so small.
So part of this process involves reflecting it a bunch of times, like 20 or something, before etching the silicon. So, what do we do? Well, ASML actually needed to invent a new type of mirror to do this, and they also needed a contract with a German company to make this special type of laser, which is the only known company in the world capable of making it. This This is crazy hard stuff.
They only make 50 of these machines per year or so. They used to have competitors, like Nikon used to compete with ASML on this, but it's too hard. They gave up. That's how hard extreme ultraviolet lithography is.
And of course, we haven't talked a lot about this, and I think it's outside the scope of the show, but just to overly simplify, lithography is kind of the process of taking that silicon wafer and etching a design on it. And if we want to do that in smaller and smaller ways, we got to do with more and more specialized equipment.
And at the end of the day, if you want to make the M2, the M3, you know, the A18X bionic, whatever it's going to be called. Bionic. There is no other way to make it than this extreme cutting edge alchemy.
It truly is alchemy. So, you know, Ben.
So you're a government. You want to throw $100 billion?
I was going to say, you know, like Acquired's doing really well. Like we're on a tear here. You know, we've got power. We've got brand power. We've got network, you know, economies. We've got our community. Like we're doing well. We should invest in this. We should have this opportunity. We should compete with TSMC. Screw the governments. We'll do it. We've got a couple hundred million dollars.
We'll buy this stuff. You know, you have a CS degree. You know, you're the more technical one of it. You can run this stuff, right? When we get the shipments from ASML, you can make this happen.
Wouldn't know the first thing to do. Even if we could invest the cash, even if we could build the facility, even if we could buy the machines, which, by the way, that's going to be hard because there's 50-some on backorder. So I can't even get it for a couple years.
TSMC has ordered out all of them for years.
It takes people who have done the most advanced manufacturing in the world ever in history in order to know how to do the next version of it.
And this is why TSMC has 40% operating margins.
It's crazy.
Totally crazy. I'm just in awe of this. Completely. Okay, so a little while back, before we get totally geeked out on that, which was awesome, you said something like, how do we get this flywheel effect? It's great, but how do we really get from TSMC started taking the dregs from the IDMs, then the Fabless companies came along. How do we get from there to now? Yeah.
There's another really important chapter here.
And you're going to flash us forward from like 93, 95 to like 2010-ish. Is that what's about to happen?
Well, first, we'll stop in 2005. So 2005, you know, things are going well. Better than Morris ever imagined. These fabulous companies are getting started. NVIDIA is killing it. I mean, I was making gaming PCs at the time. I wanted those NVIDIA GPUs.
But NVIDIA wasn't a top 20 stock in the world.
I mean, Intel was like, NVIDIA, come on. Right. Real men have fabs. Okay, maybe we're beyond that part, but they were making GPUs. NVIDIA's stock tracked whether they won the next Sony contract for the next PlayStation or the Xbox.
That was the market for GPUs. That was the market.
Right. Great market, but it's not what we're talking about here.
It's not about machine learning. It's not about crypto. It's like, is the next PlayStation going to include your chip or not?
Totally. Totally. But still, great for TSMC. It's awesome. 2005, Morris is 74 years old. He's like, all right, I did it. I've been buying TSMC stock with my own money. It's done well enough. I don't really need to work anyway. I'm going to call it. I'm going to retire. Ready to retire. Ready to ride off into the sunset. He hands the reins of TSMC over to his longtime lieutenant, Rick Sy.
And he retires. He spends a couple years. He's just chilling. I don't know what he's doing. He loves literature. He's reading all sorts of stuff. He's on his second marriage, which he credits his second wife for really reinvigorating him and inspiring him. Then it's summer of 2009.
By the way, that's right around the time that people were starting to speculate that EUV might work.
All this had been kind of an idea to this point. There had been science projects before. Oh, cool. I didn't realize that. Oh, well, this is going to make what happens even more sense. The financial crisis had happened in 2008. And chaos everywhere. We've talked about it lots on this show. Surprise press conference, TSMC, summer 2009. They announced that Morris is returning to lead TSMC as CEO.
Rick is out. Morris is coming back for the third act of his career. I don't even know what number he's wearing. He's not 45 because that was the second act. He's like Jordan. He's beyond Jordan at this point. Coming back, he's going to be CEO again at age 78. Rick would actually have a second act himself. Do you know what Rick is doing now? No. Rick is CEO of MediaTek, which spun out of UMC.
So he's doing fine. Rick's doing great. But Morris comes back. Why does Morris come back?
But this is heralded as like kind of a botched transition, right? Yeah. Well, there's a lot of stuff going on. Like from Morris to Rick, people kind of viewed it as like, you didn't really do a great job bringing in the next CEO of the company.
Maybe. I don't know enough to say. Okay. I think maybe, but also there's a lot going on at this moment in time. So the financial crisis, that's like a crisis that's affecting everybody. So that's one thing. But the other thing, so in the press release, there's a quote from Morris. He says, one, this move will not affect TSMC's fighting spirit and is likely to spur greater intensity.
But two, he says that he sees, quote, golden opportunities ahead. What are these golden opportunities that he's referring to? It's 2009. Mobile.
Right. The smartphone. 2007, in July, the iPhone comes out. 2008, the iPhone 3G comes out with the App Store for the first time, the SDK, all these developers building for it. But of course... And Android comes out in 2008? Yep. Apple had, to this point, while building this operating system, the scaled-down version of OS X, its Unix, but... They weren't designing their own chips.
They just used an off-the-shelf Samsung chip. They got it right with saying like, hey, we got to use ARM in these things because we need a really low power device. So they've done actual God's work and magic to be able to bring a PC x86 operating system, create a sub operating system from it. Computer in your hand. Totally, that runs on ARM.
Yep.
Miracle. But of course, it's an off-the-shelf Samsung processor. Totally.
Well, even that's great for TSMC. Like, you know, Intel's not making that. Okay, so that's one. We're going to talk more about that in a sec.
But we should say, and Samsung also fabbed it because Samsung is both a chip designer and a manufacturer.
But the point is, on mobile, the previous whole paradigm of computing and silicon and everything was PC. It was like stuff plugged into a wall. It was Intel. It was x86. And like, yeah, TSMC could now access some of that because AMD went fabless. But come on. But now all of the leading companies that are going to make silicon for design are ARM companies, Qualcomm, Broadcom, MediaTek, Apple.
Who all are fabless. All are fabless. Okay, so that's a big opportunity. And guess who knows all of those people? Morris. Morris.
And we should say too, 2009 was an interesting tipping point because if you'll remember back to the 2007 introduction of the iPhone, Steve Jobs has a slide where he says their hope, their goal is to get 1% of the existing smartphone market. So Apple had no notion. I mean, Google had no notion of how big smartphones were about to become.
In 2009, we're starting to see, I think the iPhone 4 came out. We're starting to see a ton of different OEMs making Android phones. You're moving into this era where everyone's looking at each other going, oh, this might actually be the next computing paradigm.
Yeah, yeah, yeah. Well, that was half of the next computing paradigm. Remember back when this is when I started in BC, there were two waves that everybody was talking about. Mobile and social? Mobile and on the consumer side, everything's shifting to mobile. That was what happened. Bring your own device? Well, sort of. You're on the right track. What happened in the enterprise? The cloud.
The cloud. Right, so you got mobile and you got cloud. And it's so simplistic, but those are the two things that drove trillions of dollars of market cap over the next decade. Well, what's the cloud? So at first, the cloud is good for Intel, right? x86, you're putting CPUs in the cloud.
Amazon's buying lots of... Dude, the cloud's the best thing that ever happened for Intel. Totally. Those are incredible server architecture.
It was the best thing that ever happened to Intel. But as the cloud progressed... and computing workloads progressed, the CPU became a lot less important. Like, AI started becoming a thing. CPU, like, yeah, maybe you need some of that. Maybe you'll use Intel. Maybe you'll use ARM, like, whatever.
But what really matters... Well, the majority of cloud workloads are still on CPUs today.
Yeah, okay, fine, fine.
But you're right, future-looking...
Why is NVIDIA now a $400, $304 billion market cap company? It's not because of the PlayStation.
And it's bigger than Intel, right? And NVIDIA's 2x Intel's market cap, something like that? Yeah. It's the cloud.
The notion of chips that are really good at parallelized processing, which is GPUs, and matrix multiplication effectively, vector math, versus the CPU, which are sort of these general purpose workhorses built for the operating system that runs on your computer, super good for serial processing,
Of course, there's like multi-core, you know, there's 64 cores on a CPU now, so they're good at parallelization too. But all this stuff, especially machine learning, is GPUs.
It's GPUs and it's specialized, like the Tesla Dojo stuff. That's not x86.
Oh, yeah, we're in this. I mean, the other thing that Foundry's enabled, the Fabless era enabled, is the custom chip. Like everybody's building custom chips for all sorts of things.
Yep. So you got these two big golden opportunities that are coming online. And Morris is like, I got this.
And we should say, we should clarify too, I think Tesla uses Samsung. Oh, interesting. I didn't know that. Not TSMC, or at least for part of it. And I think they actually even fabbed their chips in Austin in the U.S., Really? Yeah. I can't imagine that's going to last.
This is like the beginning of the like what everyone's sort of hoping for in the US is this like return to American manufacturing of chips.
They're going to have to go to TSMC though in the next generation because you want three nanometer.
It depends. I mean, it depends what the workloads are.
Yeah, I guess so. It depends what you need. Well, anyway, point is Intel's dominance is over and the index on all that's going to take over is TSMC. Yep. And Morris riding back in. He comes in. He gets these deals done. So like the Apple deal. 2012. Morris Chang, 78, 80 years old.
And I think the Apple rep on that was Jeff Williams, the classic Tim Cook's Tim Cook.
That's right.
I think there was something where it was even like one went over to the other's house for dinner or something, and it was like a living room conversation to ink the deal for, hey, we bought this company, PA Semi. We've been designing our own chip architecture in-house. We're going to launch, I think it was the A4 chip.
Yeah, I think that was the first one.
And it was Apple basically saying, we think a lot of people are going to buy a lot of iPhones in the future. And we are competing head to head with Samsung because they're a company that is not clear on strategy. They have a consumer angle here with the Galaxy phones. They think they're also kind of a foundry. And Jobs hated Samsung famously. What did he call them?
He called them some derogatory term.
Well, there's been a few interesting things. There was Steve Jobs saying he was going to wage thermonuclear war.
That was on Google, right?
I think that was Google.
But he had some like, oh, Samsung, they're just like coffee, like something that like really put them in.
It was about the lawsuits. It was like when they kept stealing Apple's designs. Yeah. And then there was something else where someone, this is later, but Tim Cook read the quote on stage about it being a toxic hell stew.
Well, it doesn't get any better than that.
No. But Bloomberg reported that it was a really big risk for both companies, both Apple and TSMC. Apple was relying on a company that was then seen as an also-ran. And the quote is, I think this is actually Jeff Williams, if we were to bet heavily on TSMC, there would be no backup plan.
And for TSMC, it meant an initial investment of $9 billion, fabs are expensive to build, and devoting 6,000 employees to building a dedicated plant for Apple in just 11 months. It took several years before it even began producing the chips. So that was in 2010. And then I think 2012 was the launch of the A4, designed by Apple, built on the PA semi-acquisition, and of course fabbed by TSMC.
And I think it wasn't until the iPhone 6, which was what, 2020? 14, 2013, something like that. That they were solely TSMC? I think so. And that was like the huge hit product. Because remember the 6 was when they first increased the screen size and those things flew off the shelves.
I'm pretty sure some iPhones had Samsung fabbed A4s and 5s in them and some had TSMC fabbed ones.
But I think by the 6, all iPhones were huge winners, but I think the 6 was like... Mega, mega winner. And I think that was all TSMC.
Huh. $9 billion of manufacturing capacity just for a deal with one company paid off.
That was a bet the farm deal. And kind of like something only Morris could do. Totally.
I mean, it really speaks to founder gravitas. Yeah. Even if he had no equity as a founder.
No equity that he didn't buy. So after getting that deal done in 2013, Morris steps down as CEO again, but he stays on as chairman. And then finally, once it all plays out and TSMC is on top, in June of 2018, Morris retires, presumably for real? He even stepped down from the chairman role this time. Yeah, fully retires from chairman at age 86. Oh my God. Crazy. Wow. So that was 2018, so...
I mean, let's talk about now. So 2020, TSMC, we alluded to this, operating profit of $20 billion on $48 billion of revenue. They took 17 of the $20 billion in operating profit and plowed it all back into CapEx last year in 2020. Beginning of this year, January 2021, they give guidance that they will raise CapEx from 17 last year in 2020 to 25 to 28 billion in 2021.
In April of this year, 2021, they raised it again to $30 billion forecast for the year and $100 billion over the next three years. That's like the real shot across the bow that everybody wakes up, the financial markets wake up and they're like, holy crap. TSMC has cornered the market. Even Samsung's not going to be able to keep up with this. It's wild.
All right, listeners, it is time to talk about one of our favorite companies, Statsig. It's funny, David, Statsig has gone from this little startup when we first started working with them a couple of years ago to this total powerhouse now.
I know, it's wild. I was looking it up and they have added all these customers since we started working together. OpenAI, Figma, Atlassian, Vercel, Notion, tons more. At this point, if there's a growth stage tech company out there, there's a pretty good chance they're using StatSig.
Yep. So listeners, if you are unfamiliar with Statsig, they basically took what was the standard product infrastructure at every big tech company, and they built it as a standalone company. This includes advanced experimentation tools, A-B testing, feature flags, product analytics, session replays, and more.
So if you're building the next great software company, this sort of infrastructure is essential because it allows your product and engineering teams to release things quickly, measure the impact of them, and track progress over time.
Totally. So, I mean, as we've talked about on the show forever at companies like Facebook or Netflix, data was just a part of how everything was built, which contributed to all the crazy bottoms up organic growth that they had. Now with Statsig, you can get that from day one at your startup.
And today they're not only trusted by startups, but also by more mature enterprises like Bloomberg and Microsoft and Electronic Arts. Turns out that a single system for data driven product decisions is useful at any scale.
Yeah, and by the way, the scale they're operating at is completely insane. They process over 2 trillion events per day now. By the way, David, this is updated. The last I checked it was 1 trillion, and then this morning I pulled it up, 2 trillion. And they handle releases to billions of end users.
If you're listening to this podcast and you've used software in the last few years, there is a very good chance you've been a part of many experiments orchestrated by Statsig.
Yeah, it's just awesome. And as they've gone upmarket, they've also started to offer some interesting deployment models like being able to run the whole thing natively inside your existing data warehouse or just using Statsig's fully hosted solution.
If you want to leverage Statsig to grow your business, there are a bunch of great ways to get started. Statsig has a very generous free tier for small companies, a startup program with a billion free events, that's $50,000 in value, and significant discounts for enterprise customers. To get started, go to statsig.com slash acquired, and just tell them that Ben and David sent you.
Thank you, Statsig.
So more on today, David, TSMC today.
Well, okay. So speaking of data, I think this is the data point that really kind of says everything. So since the first IPO in Taiwan in 1994, TSMC has had compound annual revenue growth of 17.4% for 27 years. Revenue growth. 17.4% compounded for 27 years. Now the IRR, the equivalent on valuation on market cap. So it was 4 billion market cap at the Taiwan IPO in 1994. Today it is 550 billion.
So that is a 19.9% IRR starting from a $4 billion base over the last 27 years. So 20% IRR over 27 years. Incredible by any means. Starting from a $4 billion base, it is now currently, as we record, the ninth most valuable company in the world.
And I think other than Saudi Aramco, it is the only company in the top 10 that we haven't done on Acquired.
Ooh, interesting. Yeah, the U.S. oil companies are no longer in the top 10. But Berkshire is... That might be foreshadowing some future episodes this season.
I mean, they're in hallowed grounds at this point. The other thing, just talking about financials today. So crazy that they grew 31% in revenue from 2019 to 2020. They doubled their CAGR from 2019 to 2020. Yeah, it's nuts. Talk about accelerating growth. Okay, so in 2020, their adjusted net income was $17 billion. How are they going to go spend $100 billion over three years?
Is that going to be out of profits of each of those years, or do you think they're doing some kind of financing?
I don't know. I actually don't know if they've done any financing. I'm quite confident they'll make enough profit to fund it organically because... big news just in the past week. They started this a little earlier in the year, but now they're really doing it. They're getting away from Morris' second big innovation of reducing prices.
In fact, I think they're going to raise prices this year by 20%.
So the first announcement a couple months ago was they're not going to cut prices. And then they just announced they're going to raise prices. Nobody's ever done this since the pre-Morris days. Pricing power in action. Totally. I mean, what a clearer picture of how they have taken a commodity business And turned it into, I mean, this has got to be one of the biggest moats of all time.
Totally. I mean, they've got $28 billion of cash and cash equivalents on the balance sheet. And they're going to use that and all the cash that they generate from their operations to plow directly back in to making sure that everybody else is five plus years behind. Unbelievable. The other thing is that they already are the largest.
They have over 50% of the market for foundries, for all contract manufacturing of chips. And like 95 plus percent of the profit. Correct. I thought where you're going with that, it is also true that they have 90% market share on the current generation, like the leading edge chips.
In the five nanometer, Samsung has like five, 10% market share and TSMC has 90 plus percent.
In many ways... Going to 100. They're the Apple of semiconductors. They don't have all the market share, but they have all the most profitable market share.
Yeah, exactly. They are the iPhone of semiconductors. Like, you can still buy previous generation, you know, worse technology from... And for plenty of... You know, on the Odd Lots podcast, this was actually... They talked about sort of like the bear case going forward for TSMC.
And one potential one is that, oh, well, the processing power is so good that you're not going to need the leading edge anymore. I find that a really weak argument. Like, you always need the leading edge. You think Tesla doesn't want the leading edge?
Totally.
You think Apple doesn't want the leading edge?
Software will always match the complexity on the most advanced hardware it can run on. Totally. Which is why, like, I love when people are like, Apple's slowing down my computer. I'm like, yes, I'm sure that's what's happening.
They wrote special code that they're putting on there to make the consumer... No, it's because every piece of software just always assumes that it has the most advanced processor on Earth, and it always gets to developers... Sure, they test on two- and three-year-old equipment, but no one's making sure that the six- and seven-year-old laptops are as performant.
Software designed for the current generation is hardware.
You think that Google and Amazon are going to be like, now we're good.
Right. Hell no. It actually is worth touching on. There's one other interesting bit about this 5 nanometer process, which first of all is a marketing name at this point. What it used to or originally referred to was the length on the gate on the transistor. At this point, it's not exactly 5 nanometers, and the additional performance is not going to come from making smaller gates. Right.
Here's the interesting thing, though. You actually can't put these transistors much closer to each other. So if you think about silicon atoms that are between the transistors, you can only fit five of them in a nanometer. So in a three nanometer process, sure as marketing speak, but right, like at some point you cannot subdivide silicon anymore.
So either we need to change the substrate or the innovations are sort of going to come from elsewhere. Right.
Which has always been the case. Like Moore's law was technically the doubling of the number of transistors on an integrated circuit.
Now it comes from multi-core. It comes from all the other advancements of figuring out how to make chips do more stuff faster.
Yep. That I think is going to keep going. And I think it's going to keep being expensive and getting more expensive. And I think TSMC is the only company that's going to be able to keep up at the leading edge.
Yeah. Do you know, David, about Moore's second law?
Ooh, no, I don't.
So everyone knows about Moore's law, but there's this second one, which is also known as Rock's law, after Arthur Rock. Arthur Rock. Yeah. Yeah. Oh, gee. It states that the cost of a semiconductor chip fabrication plant doubles every four years.
Hmm.
So with fabs today costing 15, $20 billion, I don't know that that's proven exactly true, but it's certain.
Well, shoot. If we just look at TSMC's CapEx forecasting, they're going from 17 to 30. Yeah. To 60 over two years. So that's way faster than four years.
So the interesting thing is when you combine these two things, the Moore's Law and Moore's Second Law, it implies that the leading company, that most profitable company, will become a monopoly.
Winner take all.
There you go. And it's fascinating that both of these laws aren't actually in conflict because Moore's Law is about... effectively, when you really look at it from a financial perspective, operating expenses when producing at scale. And Rock's Law is about the upfront capital expenditures to enable all that production. So it's everything we talk about on the show.
It's being able to pile investment into fixed cost as much as possible at huge scale in order to realize the benefits of making as many of the thing as humanly possible at global scale. And TSMC, interestingly, is the most perfect example of this. And I say interestingly because we almost always talk about operating leverage and scale in the context of software on the internet. Software, yeah.
This is how venture capital started. Because actually, manufacturing chips, the operating leverage that comes from huge amount of fixed costs into foundries to make chips and then hopefully be very profitable, 50% gross margin on those chips, venture capital financing was built for... that for semiconductors.
And it just so happened to work just as well or even better with software on the internet. Even better in the notion that gross margins of software can be 80% to 90%, not 50%. But I would back that down because it doesn't have the sort of moat defensibility characteristics that being able to plow your CapEx into manufacturing capability does.
Yeah. Yeah. Should we do power now? Absolutely.
Let's do it. Let's do it.
So for folks new to the show, this is one of the discussion topics we do for every episode as we go through Hamilton Helmer's excellent Seven Powers.
The best business theory book.
We've had Hamilton on the show. He's amazing. Go read the book if you haven't. He identifies seven powers, essentially sources of defensibility, which he defines as long-term differential profit margins versus your competitors, as we've been talking about on the whole show.
The seven that he identifies are counter-positioning, scale economies, switching costs, network economies, process power, branding, and cornered resources. And we almost always talk about network economies. We talk about counter-positioning on this show.
Sometimes we talk about branding.
Sometimes we talk about branding.
I think we're talking about none of those this time.
Yeah, we sometimes talk about scale economies, which we're going to definitely talk about here. But I think we're going to have our first process power if I'm going to forecast. But let's start. Let's go down the list. Counter-positioning.
I mean, like maybe you could. When they were starting, and in particular, would the incumbents have started with the exact business model? No, because their profit center was the integration that all the margin you get of integrating design and manufacturing.
And by saying, nope, we're going to be a pure play manufacturer, TSMC theoretically was saying, no, we're going to take less gross margin and we're just going to make it up in volume. I'm actually not sure it played out that way. I think they have more. Do you know what Intel's gross margins are?
I actually don't know. I would suspect they're higher, but I don't know. Yeah, there was counter-positioning here.
I don't think I said this when we were going through it, but before TSMC and the PurePlay Foundry model, if you were either a Fabless company, one of the very, very few, or you were another IDM and you were trying to get some excess capacity, you rent it from another IDM, most of the IDMs would be like, okay, you strong-armed them, you got a great strategic relationship, they'll give you some capacity.
But they also demanded the right to market your products under their brand too. So like, which obviously TSMC wasn't going to. So yeah, there was counter-positioning. Like the IDMs, no way they were going to do what TSMC was going to do. Right.
Huh. Okay. Scale economies. Absolutely. That is the biggest. It's one of the top two with process power, in my opinion.
Switching costs. Well, it's funny. Now there are huge switching. You can't switch off TSMC.
No, unless you're going to stop being on the leading edge. If you're going to change from being a phone company to an automotive company, you can switch off of them.
Well, I think it's even deeper than that. Again, we haven't gone. Listeners probably think we've gone deep technically on this episode. We haven't even scratched the surface. But yes, if you want the leading edge, now you got to be TSMC. But you got to be so integrated with TSMC to do this.
Say you want to switch to Global Foundries or one of the other competitors out there, of which there are a few. You can't just call up Global Foundries and be like, hey, I'm porting over. Expect my business on Monday. It takes years because you're so deeply integrated with the process. So yeah, big switching costs.
Network economies, it's not really worth talking about.
Not in the traditional sense. You know, this is not Facebook here.
And certainly none of TSMC's customers really benefit from other customers being on it.
No, I do think there actually is... I don't think Hamilton captures this in his... And I don't know if you would consider this one, but there is like an ecosystem aspect here because the EDM companies and the IP companies are so deeply integrated with TSMC. If you want to be using ARM, for instance, or, you know, they're kind of the best integrated with.
Now, I don't think that's network economies. That is kind of like this ecosystem thing. TSMC actually has a name for this. They call it like the open innovation something or other. It's some corporate name, but it means this.
I do wonder if it's actually worse for a lot of people that Apple is a TSMC customer because who else has access to the five nanometer process right now?
They're going to take as much as they can. Yeah. Good point.
Process power. Yes. I think other than Pixar, this is the first time we've really, although we weren't doing seven power string.
Yeah. To me, this is the clearest example I could ever imagine of process power.
It takes all 40 years of processing. TSMC's history to have arrived at where they are today. And even if 10 people left and tried to start the next TSMC, to be able to create what they've created at this point from scratch, virtually impossible.
All of their IP, all of their people, all of their know-how, all of their relationships with ASML and the like, yeah, no amount of money can replicate it.
I think the only thing that will unseat TSMC is a complete paradigm shift. Yeah. Something where, like what mobile did to desktop. If there's something where the compute required in the future is unable to be provided by anything that TSMC is good at today.
If all the crazy laser molten tin ASML stuff we were talking about If all of a sudden there's discovered a new, either different or way cheaper way of quantum computing. Yeah. Way to do this. Then that kind of could reset the playing field. Yeah.
Yeah. Yeah. Totally. But even little shifts, I bet they'd be fine. If everyone figured out that like, hey, silicon's not the best substrate and we can figure out a better substrate. Yeah.
If there were like an AWS moment, which is funny because TSMC is the AWS equivalent, but where something happened that just made it way cheaper than it used to be, you could now get access to the technology and the know-how orders of magnitude cheaper than it is now. That would take away a big part of their power.
But I don't see that happening. No. Absent a paradigm shift, this is TSMC's to lose. They're pretty much in the groove. So I think we should skip branding and corner resource for now. It's not really worth talking about.
Literally, they're antithetical to branding. It's Apple's brand. It's not TSMC.
So this is, I think, a good time to enter our geopolitics discussion. Because I was thinking about the other way that TSMC could fail would be that China decides the moment is right to go and assert our force and take over Taiwan.
Depending on how you see it, either annex Taiwan or assert its as always claimed sovereignty in Taiwan.
Yes, actually start enforcing what has been right the whole time, I think as they would sort of say. Yeah.
If they're speaking in my casual tone in English from America, then doing all this business with the West, I have to imagine that assuming that it didn't start a full war, like an actual world war, which it may, then of course they would start using all the TSMC manufacturing capacity for all the Chinese customers. And Huawei's been a TSMC customer for a long time. And is currently cut off. Yeah.
So how do you capture that in power? What is the power? I mean, that's a risk.
That's like a bear case.
Right. Let's not get too specific on this, but maybe in a general sense, how do you capture the power that a company has that comes from a regulatory environment? Where would that get classified under? That they have a lot of room to be operating safely?
Um, maybe cornered resource, I guess. So you're saying this is like an anti-power for this.
This is a weakness. Exactly. Yeah. Exactly.
Yeah.
I suppose all that matters is things that you have that your direct competitors don't. So I, in this strange straw man that I'm putting together, it would really be about what if you were located in a country that none of your competitors were also domiciled in and being there gave you some special ability to be more profitable than others.
Which they had in the beginning with the government of Taiwan. Basically, the mafia boss was like, this is happening. We are going to strong arm all the business leaders in the country to investing in this. We're going to make sure that this happens.
Yeah. Okay, well, let's put a pin in that because you're right. It turns out that it's actually not a perfect power discussion. But the geopolitics thing is interesting.
Well, I think it's the bear case.
Right. That to me, absent an enormous computing paradigm shift, is the way that TSMC has an enormous risk in the business.
Yeah, totally. Which does make it kind of surprising that they haven't diversified their geographical operations very much.
So this is interesting. So they're facing a lot of pressure for this. They are spending, I think... $12 billion this year to start a plant in Arizona, which will not be the three nanometer. I don't even think it'll be the five nanometer. It's not their most advanced manufacturing. I think the US is subsidizing in a big way.
I think that's part of the Biden administration's most recent bill to try and bring some semiconductor manufacturing here. But they're also starting a fab in Japan that came out on their last earnings call.
So they're doing some... And they have operations in China, I believe, too.
Yep, they're doing some diversification, but I don't think it's for this reason. I think it's because they're basically getting free money to open fabs in other places. And Morris has even made comments like, I don't think it makes any business sense for us to have the leading edge in those countries, even though those countries want us to have them there.
I think it makes sense based on the ecosystem that we've created in Taiwan to keep operating it here. Yeah. So the question is if it directly helps, let's take the U.S., for example, the U.S. 's prowess as a semiconductor manufacturing force in the world to have TSMC's Arizona plant, or if it's really just indirect and the idea is like, let's try this as a first stab.
We'll get more people in the U.S. familiar with doing this again in case we need to restore this. Yes.
Yeah. I mean, this is a scary, scary future to contemplate and hope to God it doesn't happen. But really the thought exercise here is what would happen if China annexes Taiwan tomorrow?
Which is scary for a number of reasons, the smallest of which is this corporate takeover. It's scary for a lot of people, their lives. Yeah, scary.
But I wouldn't say it's the smallest, like everything. Imagine if we didn't have access to leading edge semiconductors. Yeah. That's... everything. What part of our lives do not run on semiconductors?
Right. Ford can't make F-150s right now.
Basically, all of our technological progress would stop. Yeah, you're right. So I think the question is, and I don't know enough to answer this, what would happen? Would it be possible to airlift the process power that TSMC has physically out of Taiwan to somewhere else? You get all the people, ASML now sends the stuff somewhere else, you airlift everybody out, there's an evacuation.
Does the process power come with it or not?
I don't know. That's a good question. I mean, if the Toyota production system is an example where Toyota tried to, there was that factory that joint venture with, was it GM?
Yeah, the Numi plant that's now the Tesla plant in Fremont.
With Toyota trying to replicate their process somewhere else didn't work.
Now it wasn't under threat of war.
Right, this one would need to. It's actually a good question. If you think about the U.S. 's strategic defensive weaknesses, what's more important, having onshore semiconductor capability to continue to advance technology in the nation or Boeing? our ability to build, which we've always held up as this example of the U.S. needs that to stay U.S.
owned, to stay operating, stay profitable, to stay prosperous, because it is a matter of the U.S. way of life that we're able to protect.
Well, Boeing needs semiconductors.
That's a great point. So we're now outside our depth, but is it actually more important to have cutting edge semiconductor capability here than airplanes or any of the other sort of defense supply chain?
And, you know, maybe the answer here is like Korea. Same situation exists in Korea with Samsung, right? Like North Korea is right there. I've been there. I've been to North Korea. Like I went to the DMZ. Like it's so weird. It's like an amusement park. Weird. It's super, super weird and bizarre. But yeah, North Korea is right there.
You know, maybe it's the same like China's right there, but this isn't actually going to happen. But I don't know. It feels in the last year like the risk of it actually happening has ratcheted up quite a bit.
I think so. I mean, it's like globalization as a whole. It is in the best interest of everyone to continue to share resources, to continue to entangle everything until somebody decides that it's not and that we have a big problem. And hopefully for lots of reasons, it just continues to be okay that TSMC is located on an island that is of disputed claim. Yeah.
Maybe the best thing that could happen is, my carve out a while back was the book by the Harvard, chair of the Harvard Astrophysics Department about Oumuamua that he postulates was an alien spaceship. Maybe if we discover that aliens are real, that's going to be the uniting force. All these conflicts seem pretty petty. I wouldn't use that as an investment thesis, though.
No. Okay, before we get into playbook and just hit some things that I think we missed during the narrative, or at least didn't put a fine enough point on in the narrative, I have a what would have happened otherwise that I want to hit. We haven't done this in a while. No, we haven't. And I'll just read this as a direct quote from Bloomberg.
And there were some awesome sources for this episode, all of which are linked in the show notes. In the mid-2000s, as Apple Inc. was preparing for the release of its new smartphone, Steve Jobs approached then-CEO of Intel, Ottolini, about providing the chips for the iPhone. Intel already sold iPhone, the processors that ran on its Macs.
We need to add video to Acquired so that everybody can see the look on my face right now. I'm just like, literally, I got like fists in the air. I'm so happy.
And remember, Adelini was the guy that Jobs brought out on stage during the Intel transition. They were burying the PowerPC to say, this is the future. This is the partnership. So, okay, but Jobs made what Ottolini considered a lowball offer, and Apple awarded the contract to Samsung.
It later began designing the chips itself, eventually outsourced production to TSMC, a contract manufacturer in Taiwan that had been found, blah, blah, blah, blah, blah. So what could have been? Apple went to Intel and said, do you want this contract? Because they were partners on the back. Totally. And apparently it was less about the fact that, I'm sorry, you want to use ARM? What?
No, we're the x86 company. And it was more about, we felt it was a lowball offer.
Biggest strategic error of all time. All right, I'm going to postulate a playbook theme and I'm put forward as a post. More than a playbook theme. What's the, you know, like in geometry, there's like laws, right, that are proved, but then there's like postulates that are like, you can't prove them, but our fundamental understanding of the universe doesn't work if they don't work.
Whatever that is, axioms, I don't know whatever it is. I'm going to put one of those out there. Please never make strategic decisions based on economics. Hmm. This is prime example. Like the number of, and we talk about this all the time on this show, VCs passing over valuation on something, you know, Andreessen getting cold feet about a $300 million valuation on Uber.
Yep.
Right. This Intel move passing on being partnering with Apple.
And maybe more specifically than economics, because like you could imagine that you would want to pass on this if Intel didn't get any of the upside from the deal. Yeah. Assuming that the structure is right, then passing because a number is too low in the structure.
Or Ford Motor Company not hiring Morris Chang over $1, you know, like whatever. Like it's just, humans are so prone to cutting off their noses despite their faces.
So we already have the Rosenthal doctrine of never bet against the internet, but now we have the Rosenthal postulate, which is never make strategic decisions based on pricing.
Not economics, but pricing.
Hmm. I like it. I need to add a new section to the acquired website. All right. All right. Next on Playbook is another one on Intel fading. So it takes a very long time to become irrelevant. So despite Intel's stock price being, I think TSMC is like two and a half X Intel stock price. As a matter of fact, ASML is actually larger than Intel by market cap now.
They are the sole source provider of one thing in the value chain to mostly one company, and they're bigger than Intel now.
I'd be fascinated. Okay, so public markets investors who are listening, shoot us a DM in Slack or post in general.
AcquiredFM at gmail.com.
Or AcquiredFM at whatever channel works for you or Twitter, whatever. I'd be super curious if you are along this thesis that we're sort of laying out on the show, how are you playing it between TSMC and...
ASML, which is now Europe's most valuable company.
Right. I mean, probably you just invest in both, but how do you think about that?
Right. And what's the up and comer that's kind of speculative at this point, but could be another puzzle piece here.
Are you also shorting Intel through all that? Like, what do you do here?
All right, so my point on Intel is it takes a long time to become irrelevant. They still control 80% of the computer processor market, and they have an even bigger share in servers. So despite everything we're saying... It's still huge. Workloads running on CPUs that are in computers and on the cloud, pretty big business.
Yeah. The majority of workloads that are happening in the cloud is not Tesla Dojo. It's, I don't know, some company that's not a tech company somewhere in the world running their Outlook server on Office 365. Yep. It doesn't need five nanometer process.
Two other Intel things. One is that indecision has been very tough on the company. Bob Swan, who was the former CEO, started to prepare to outsource manufacturing of Intel designed ships to TSMC. I think even like two years ago, this was the plan. They finally decided, thrown in the towel, you know, Intel is the greatest ship manufacturing company in the world.
Real men are sensitive. They talk about their feelings.
Bob Swan is no longer the CEO of Intel, and now in a complete reversal, their new CEO, Pat Gelsinger, wants to turn Intel into a foundry themselves by which other fabulous companies can contract with Intel to build. Maybe that's right, but if so, they got to figure out, and I think they're thinking about this the right way because they said it's going to be a fully separate autonomous division.
They got to run that like a completely separate independent company of the rest of Intel. And if so, I don't actually know why Intel owns it.
Yeah. I mean, well, A, let's look at AMD here, right? They did this. They spun out their manufacturing into GlobalFoundries.
Which has been good for GlobalFoundries and AMD. Like GlobalFoundries is getting ready to IPO.
Yep. Yeah. That's probably the right strategic decision, but it's not going so well. I mean, like it's going fine. It's not TSMC. Right. It's going better probably than if they had not done that, but they're not a winner here. Like TSMC is the winner. Yeah.
Yeah. I guess the playbook theme there is indecision is paralyzing. I mean, this company has spun its wheels one direction or the other, and all it's done is make itself deeper in the mud.
Oh, I just looked up. I was trying to remember this. Gelsinger was the VMware CEO. He started his career at Intel and then EMC and then EMC owned the majority of VMware. They became the CEO of VMware. Yeah. And he was the outside candidate to replace Bomber as Microsoft CEO. Yeah, yeah, yeah.
You know, I hear he's really revered in the organization, that people think he's really going to make some good change there. We'll see. The last thing on Intel, and it's funny this is not the Intel episode, but there's a thing that happened here that is very similar to the fact that Kodak developed... the digital camera first in their lab. They knew it.
They knew this was the future, and they didn't commercialize it because it's impossible to counterposition yourself because of the innovator's dilemma. Intel actually saw extreme UV lithography, EUV, first. So Intel was the biggest early investor in EUV, committing more than $4 billion to it in 2012.
Whoa.
It was slower than its main rivals, and this is from the Wall Street Journal, in adopting the technology and skeptical about whether it would work. Eventually, Intel calculated that it was a sure bet to try and improve existing ways of handling lithography. And of course, where we are today, EUV completely enabled the next generation of chips to be built that existing ways couldn't.
What a great argument and example for why you need startups. Right? Right? Like... Totally. Like, yeah, Intel was there. They invested in it. They saw it and they were like.
They put $4 billion in. And I think even to this day, there is not a shipping Intel chip that was manufactured by Intel using UV.
Wow. That's crazy.
You're right. It is the most perfect, pure example of the innovator's dilemma in action.
That's why you need startups.
Yep. All right. My next one is that if you're only looking at the outcomes that happened, you cannot reverse engineer what the probability that it would happen is. And this is a very abstract way of me saying the strategy of if you build it, they will come that Morris implemented. is a bad strategy, and it also worked.
Right. This is what Sequoia and Don Valentine hated. They would never invest in developing a market. That was rule number one. We invest when the market already exists, not when we need to develop it.
And this is the classic problem. This is the knock. Up here in Seattle, there's a lot of people spinning out of Microsoft, starting companies. Classically, people coming out of Microsoft would always want to build platforms because Microsoft was the platform company, and they would always... have a small, too small of an understanding of the market of people that wanted that platform today.
And they assumed if you build it, they would come. Morris was that exact problem. And yet, if something is going to be true 10% of the time and fail 90% of the time, one out of 10 times, it's going to work. And it may have been the case.
I guess what I'm saying here is if you're starting a startup, it's impossible to know if this was actually a good strategy or if it was a bad strategy that probabilistically just happened to work.
I mean, this is the thing about startups, right? There are all these rules, but they can all be broken. There is no formula. Yep, totally. All right. Other playbook themes? I just have one more that, again, we talked about a bunch in the episode, but I want to highlight and actually add one spin on here. You know, the Jeff Bezos quote about AWS.
As a startup, anything that doesn't make your beer taste better. The analogy back to German beer factories and outsourcing electricity generation.
Outsource things that aren't your core competency. Right.
Focus on what makes the beer, your beer, whatever that is, proverbially, taste better. And everything that is not that, like finance and accounting, outsource, etc. Double underscore that. But this is obvious, so obvious. But obviously, Bezos didn't say it directly. And thus, I think we don't highlight it enough.
The counterpoint to that is anytime you see something that lots of people, lots of companies are doing that is not making their beer taste better, that is a massive opportunity to go build a platform company. That is how you build a platform company. Grading. All right. So we were thinking for grading. Look, we could grade like, I don't know, Taiwan's decision to do this.
To own 50% of the company at the outset.
Whatever, like, you know, A plus, you know, not interesting. So we had the thought, experiment, we'll try this for this episode. Rather than like letter grading this, we'll ask a question. Where does TSMC belong in the pantheon of great technology companies of all time? Is it Fang level? Is it like top five? Is it top 10? Is it like top 20? Like, where is this?
What is the right context in which we should be placing TSMC, this whole story, the company, the power, all of it?
So interesting, because it really does raise this question of value chain. We talked about the five-part value chain that exists today for making chips. And so it's interesting because you could say, well, it belongs wherever Intel belonged circa 2000. Or you could say, well, the set of products that TSMC manufactures have 100x the scale that Intel in 2000 had.
If you think about it, all this stuff that everyone's all excited about, every time someone talks about the next wave of computing and they're like, machine learning, or they're like...
crypto or they're like 5g and anything they tell you is something tsmc makes that enables it all when mark andreessen says software is eating the world it's only eating the world because tsmc has made it so freaking cheap to manufacture silicon and then you can run whatever you want on that silicon and it's the cost of compute asymptotically approaches zero because tsmc tsmc tsmc
So how much do we ascribe to them versus ASML? How much do we ascribe to them versus the entire landscape of talented chip designers out there, including the like 600 chip designers at Apple working on the Apple Silicon? It's hard to disambiguate that. So where does it belong? I mean, it's probably the most successful and important B2B hardware company of all time.
I think we can safely say at this point it surpasses Intel. I mean, gosh, right? Like that's a big statement to say, right? Like Intel, Silicon Valley, the traders say like all of it. Moore's law.
But in compounding, all the value shows up at the end. So it is true that the value that TSMC will create in the world over the next year, two years, three years is probably more than the entire silicon industry leading up to this point combined.
I mean, hell, they grew 30% last year. At an already unimaginable scale.
Right.
Intel's not doing that.
Right.
Okay, I think we can say it's above Intel.
I probably wouldn't say it is above Facebook, Amazon, Apple, Microsoft, Google in terms of pure value creation in the world.
I mean, devil's advocate, you could argue that none of the innovative things those companies are doing now happens without TSMC.
Yeah, unless the foundry model and the fabless model was inevitable.
Yeah, maybe somebody else would have done it. Yeah, maybe. But they didn't. They didn't. Norris did. I mean, guys, the thing that's really just beaten me over the head in this episode, we've probably beaten all of you over the head with, or at least I have, is, you know, look, there's the geopolitical risk with being in Taiwan. Other than that, I don't know that there is a stronger moat
that any company has in the entire world than TSMC. Compare it to all the FANG companies and Microsoft. Those are very, very strong moats. But we've seen all of those, you know, they've changed, they're new companies, they've emerged, they've, you know, Microsoft fell and then now it came back with new strategy and Facebook's not that old and Google's not that old. TSMC is impenetrable.
Yeah, their business model and the costs required to compete are such that they have... It's like bulletproof. It's everything but bulletproof. Yeah, totally. Sadly.
Sadly, yeah. So, you know, I don't know. Maybe we're exaggerating because we're, like, so deep in it. Like, we always go native on these episodes.
Right. The only way it could be more valuable is if the company had an army. It's like people talk about the U.S. dollar is backed up by the full faith of the U.S. government, which implies guns. Yep. And so it's only because everybody's currently playing by the rules that any business gets to stay in business. And so... This one just happens to be a little bit more at risk than other ones.
All right, so I think we can safely say top 10.
I think the question is, is it top five?
Well, Defensible is an interesting question. So in 30 years, will TSMC be a huge company?
Well, they've got this dynamic going right now with this flywheel where structurally nobody can catch them. Something unforeseen has to change.
But something unforeseen will change because it always changes. Right, right, right. Yes, yes. True. Who's had the most similar dynamic in the past?
Standard Oil.
Either been successful or unsuccessful. Standard Oil is a good one.
I mean, that's very different style, but same sort of dynamic with Standard Oil, right? Was they crowded out structurally how they were set up. Nobody else could compete. And the rich kept getting richer.
And they kind of still exist. That's the best.
That is the best part. They kind of still exist. Yeah.
All right. I'm with you. I'll go top 10, but probably not top five.
What I'm wrestling with is how much of it is just marketing. And I don't mean marketing in a bad way, but intentionally TSMC rides under the radar. They intentionally have no brand. The brand is the customers. They want the customers to succeed. So we don't hear all the time about them like we do the fan companies.
Yeah, we will start to. I think anybody who tunes into this episode probably saw... the name of the episode and then thought, I should tune into that because I've seen more about this thing recently that I previously didn't know about. Yeah, I should learn about this.
Kind of like we did when we were like, we should do this episode.
It's finally time.
Yeah.
Yeah.
All right. Well, that's where I want to leave it. All right. I'll put a stake in the ground. I'm going to say, I think I'm with you. Top 10, not top five yet. But, you know, maybe we need to revisit this.
I will definitely say it's the most successful B2B hardware company ever. And the question is, is it the most successful B2B company ever? I'd say it's probably just competing with Microsoft there.
Yeah. I mean, and again, maybe even like across all industries, right? Like look at, I mean, shoot, semiconductors run everything. Yeah. And they run semiconductors.
Semiconductors are the new oil, David.
Okay, okay, enough, enough, enough. We got to bring this one home.
Carve-outs?
Carve-outs, let's do it. I've got two. Jenny and I were just down in Santa Barbara for a couple weeks, rented an Airbnb down there. It was so great. We did that last year. Hopefully this becomes an annual thing in the summer. Yeah. Escape the freezing San Francisco summers. So while we were down, we don't watch a lot of TV usually, but, you know, it was like...
Change of scene, summertime in a new place. We're like, all right, we'll watch some TV together at night. So this is like, for the percentage of you out there who are living under a rock like me with TV, we've watched now most of Ted Lasso season one. Because we heard season two was terrible, but that made me think, well.
It was terrible. It is terrible. But season one's great.
That made me think, oh, if people are this upset about season two, that means season one was really good. It's so good. If you haven't watched it, we're on episode eight now. So we're not quite done. Welcome to the club. So good. Love it.
And then the other TV show we watched, this was Jenny's suggestion, old school throwback show called Greek, which aired in the mid 2000s and is about Greek sorority and fraternity life on a fictional university. And it's just so good. It's like one of those heartwarming period pieces, but it was right from when we were in college. So yeah, it's fun.
Nice. All right. David watching TV, who knows what could change in the world?
Right. Maybe TSMC's mode isn't as deep as we thought.
All right, well, mine is a book that has been recommended to me for two or three years now, and I finally got around to reading, and it was awesome. It's called Who Is Michael Ovitz? And if you've read Shoe Dog, and you've read The Ride of a Lifetime, and you've read, what's the Ford one? An American icon, these iconoclastic CEO founder business books.
There's a Sam Walton one, Built in America, I think. Yes. Yeah. This one needs to be on your list, especially if you've enjoyed any movies or TV shows that were put together in the last, well, let's be specific.
Or our two-part Andreessen Horowitz series.
Totally. From like 1975 to 2000, Michael Ovitz put everything together. And it is this wonderfully written book about an unbelievable business story, the strategy behind it, The way that with Creative Artist Agency, they just completely upended the entire industry in Hollywood and did it really without ever talking to the press and were very tight-lipped about it.
For some Hollywood outsider, I found the book really wonderful, really compelling. I also think I previously had only read The Ride of a Lifetime and watched the Disney Plus special about sort of the history of Disney and Disneyland. I had a one-sided view of Michael based on just his short tenure at Disney.
Yeah. I was going to say, yeah, what a great connection with, you know, Acquired and Disney and Tiger. Totally.
And what kicked it off was doing the Andreessen episodes and hearing about how they based it on CAA. So especially if you like those episodes, or if you like the Disney episodes, or if you're a movie fan, or if you like these classic CEO business stories, Who is Michael Ovitz was just an awesome read.
So cool. Well, and you know, like all the media that we grew up on, probably even more so because we were kids, but it was the adult movies and the kids movies too. But you know, when you're a kid and like the adult movie that you really want to see, but you're too young to see.
And just all these such classics like Goodfellas, like that was just my previous carve out or Jurassic Park or just everything that they packaged. It's cool to hear how it came to be.
Super cool.
I don't think we told you at the beginning, but you can join our Slack, acquire.fm slash Slack. Come hang out with other talented, smart, good-looking people like yourselves. And with that, listeners. feel free to share the show with a friend. Shout out from the social media hilltops. Yeah.
You always, sometimes you say, but I'll chime in here too. Like seriously, you know, it's funny. Podcasting is this weird thing, right? There's no viral loop. It's not like, you know, you can share, please share it from social media. We love that. That's great. If you love this episode, you think it's interesting. You think what we do is cool here. But really the way this goes is word of mouth.
That is it. People tell their friends, they listened to this episode, they thought it was cool. They think that their friends would really enjoy and learn from listening to this too.
Share a thing you liked, share a thing you disagree with us on, whatever it is.
If you feel that way, please do that. If you don't feel that way, get in touch with us and tell us why.
All right, listeners, we will see you next time.
We'll see you next time.
Who got the truth? Is it you, is it you, is it you who got the truth now?