Complexity Investing & Semiconductors (with NZS Capital)
01:42:19 Share

Okay. Excitement, fun, brevity. I love it. For brevity is the soul of wit.

Welcome to this special episode of Acquired, the podcast about great technology companies and the stories and playbooks behind them. I'm Ben Gilbert, and I'm the co-founder and managing director of Seattle-based Pioneer Square Labs and our venture fund, PSL Ventures. And I'm David Rosenthal, and I am an angel investor based in San Francisco.

And we are your hosts. On our TSMC episode, one of the 65 sources that we used was an episode of The Knowledge Project with Brinton Johns and John Bathgate. Brinton and John are public equities investors at a hedge fund called NZS Capital, and they spend a lot of their time researching semis. It was packed so full of great content that I actually watched it twice to make sure that I understood everything. It was so good.

It was awesome. And then in a wild coincidence, the very next week, even before we shipped the TSMC episode, David and I were at Capital Camp, great event organized by Patrick O'Shaughnessy and Brent Beshore. And we ran into Brinton in person. And I was like, I recognize that guy. What do I recognize him from? It was like the Spider-Man gift. It's like, you, no, you. Very, very much so. So after nerding out the whole rest of the event on TSMC, geopolitics, semis,

We decided to have Brenton and John on Acquired. And on this episode, we actually didn't even get into semiconductors for the first hour since it was so fascinating to hear about their investment principles at NZS. And I've said this many times on the show, but yet again, new frameworks that have totally changed the way that I think about the world. It's frameworks all the way down, Ben. It is.

Okay, listeners, now is a great time to thank one of our big partners here at Acquired, ServiceNow. Yes, ServiceNow is the AI platform for business transformation, helping automate processes, improve service delivery, and increase efficiency. 85% of the Fortune 500 runs on them, and they have quickly joined the Microsofts and the NVIDIAS as one of the most important enterprise technology vendors in the world.

And just like them, ServiceNow has AI baked in everywhere in their platform. They're also a major partner of both Microsoft and NVIDIA. I was at NVIDIA's GTC earlier this year, and Jensen brought up ServiceNow and their partnership many times throughout the keynote.

So why is ServiceNow so important to both NVIDIA and Microsoft, companies we've explored deeply in the last year on the show? Well, AI in the real world is only as good as the bedrock platform it's built into. So whether you're looking for AI to supercharge developers in IT, empower and streamline customer service, or enable HR to deliver better employee experiences, ServiceNow is the platform that can make it possible.

Interestingly, employees can not only get answers to their questions, but they're offered actions that they can take immediately. For example, smarter self-service for changing 401k contributions directly through AI-powered chat, or developers building apps faster with AI-powered code generation, or service agents that can use AI to notify you of a product that needs replacement before people even chat with you.

With ServiceNow's platform, your business can put AI to work today. It's pretty incredible that ServiceNow built AI directly into their platform. So all the integration work to prepare for it that otherwise would have taken you years is already done. So if you want to learn more about the ServiceNow platform and how it can turbocharge the time to deploy AI for your business, go over to servicenow.com slash acquired. And when you get in touch, just tell them Ben and David sent you. Thanks, ServiceNow.

Now, as always, this is not investment advice. We almost certainly hold stocks that we talk about on this episode. So do your own research, make your own decisions, but love the frameworks that we dive into with Brenton and John. So without further ado, on to our conversation.

All right. Well, listeners, we want to introduce you to NZS Capital, their investing philosophy, and partially because we think they're a fascinating firm, similar to Honam of Altos or Hamilton Helmer of Strategy Capital, but also the deeper David and I have dove down the rabbit hole the last week of reading all the papers they've published, we feel like we've gotten a lot smarter. And so we basically just want to expose the world to more and more of that. So Brenton and John, welcome.

Thanks for having us. Thanks for having us. Well, first, let's start with complexity theory, which is a concept that your whole firm is based on and you've published a 47-page paper on that is just A, full of fun little graphics, but B, I think probably has five to eight kind of mind-blowing concepts in them. And the first one is you come right out and admit that you don't know the future. What's going on with that?

You would be bad marketers as VCs. Right. He claims to know the future. No, I mean, look, the whole investing philosophy comes out of a lot of pain. Right. So we are investors for a long time. We are wrong a lot. Like all investors are wrong a lot on a consistent basis. And we are just looking for a better way to think about things.

Somebody suggested this book to me called The Origin of Wealth by Eric Beinhacker. And sort of serendipitously around the same time, someone suggested Complexity by Mitch Waldrop to Brad. And we both read those books and Origin of Wealth was a slog. I think it took me six months to really get through it and then swap books and started thinking about sort of a different philosophy.

I'd heard a little bit about complexity theory and the Santa Fe Institute, which I want to get into. I think Bill Gurley talks about this fairly frequently and Michael Malvison, and that's how I kind of originally got turned onto it. But tell us a little bit more about what is it? Because it's not at all about investing. It's about the world. That's right. Yeah. In fact, I think it was Bill Gurley that recommended complexity to Brad.

Complex adaptive systems are all around us, right? That's what governs the world. That's how the world works. We don't know how the future is going to unfold because the system is interacting together and it creates what's called emergent behavior. And emergent behavior makes predicting useless in most cases. And we can have guidelines and heuristics and those are all helpful. But as far as exact outcomes and what's going to happen in the future, those are a lot more difficult.

Santa Fe Institute started with a group of scientists from the Los Alamos National Labs, and they came together and they were mostly physicists and they started talking to economists. It was sort of hard sciences and soft sciences. And the physicists were like, hey, economists, guys, you guys seem really smart, but you know, your theories, they don't work.

All your math doesn't work. So what's up with that? With our math, it's extremely precise. In fact, when the math is off just a little bit, Einstein's like, oh, your math is off. Pluto should really be here. And he comes up with the theory of relativity, right? We literally made the atomic bomb. It works. Yeah, it works. And so they started coming together around this idea of complexity. What is complexity? How do we define complexity? Where does it sit? And because...

we are living in this complex adaptive system, how do we think about the future? How do we think about life? How do we think about going forward? And for us, this sparked an interest in biological systems. And we found this sort of biology vein much more interesting than the traditional economics vein and much more applicable to investing than the traditional economics vein. Well, that's so cool. And am I right that y'all actually went to the Santa Fe Institute and took courses there? Like, how deep did you go in this?

We did. Like a lot of rabbit holes, we go down very deep. We quickly became members of the Santa Fe Institute. They call it the action group. It's this group of non-scientists that are allowed to sit in on a lot of the science. And so then we took this complexity course over a weekend, Brad and I did at Stanford. And that was just a ton of fun, actually. John and Joe, the two other investors on our teams, they took a longer course. They actually had to do real work. Brad and I didn't have to do homework.

But we just learned so much. And I remember sitting outside of this cafe in Palo Alto with Brad, and we'd just sort of been at this course with Debra Moore, this lady that teaches at Stanford who studies ants. And I thought, man, this concept of resilience is really fascinating. You know, it's really more about resilience than it is about predicting the future. And it's about adaptability. Biology doesn't really care that much about the future.

They care about adapting to this wide range of futures. My bees don't really care if it's going to snow tomorrow. They can adapt to snow. They've learned how to do that over millions of years. And what if we looked at companies like that? And so then, of course, we kept reading, we kept writing. This was probably 2011, 2012. And in 2013, we published this long paper that you referenced, which is super geeky, but it's got a lot of pictures because that's the way we think

You've got the back to the future DeLorean in there. It's got the DeLorean. Like, what more could you want, right? We're really hoping for a DeLorean for the office. That's our dream office furniture. On the note of ants, this is probably the first and best example of an extreme version of resilience in an organization. Can you share the insight you had there?

Yeah. So we attended this class by Deborah Gordon, and she has been studying this group of ants for 30 years in New Mexico. They obsess over this group of ants, right? And they know what every ant is doing at all times. And what they found was really fascinating. They found that about half the ants in the colony weren't doing anything. They were just sort of sitting around, and then they had half the ants doing these defined jobs. And that's very counterintuitive. We think of ants as sort of the ultimate productivity machines.

But it turns out ants aren't optimized around productivity. They're optimized around longevity. They're optimized around resilience, around living as long as possible. Let's say it that way. So that was really insightful for us. We thought, man, all these companies are optimized around productivity and Wall Street only makes it worse because we're obsessed over quarterly earnings.

And so what if companies were really optimized around this long-term thinking? Of course, we see that with lots of companies. Most of them tend to be run by founders because founders have a lot of skin in the game and they think long-term. But there are CEOs that think that way also. We know that the average tenure of a CEO in S&P 500 is less than five years. So they're not optimized like ants are. They're trying to get a lot of return really quickly. But companies that take this long-term view are so much more interesting.

When I read that in your paper, the thing that hit me over the head, I was like, oh, this is Warren and Charlie's, you know, laziness bordering on sloth. That's exactly it. The goal is not productivity. The goal is long-term steady returns and resilience. I think Warren and Charlie grokked this very early. And, you know, there's a lot of science behind it, math, but they don't need that. They're so good with folksy wisdom.

So then the thing that hit me, of course, I'm like, well, a company is not going to have half their employees sitting around doing nothing. But just as a fun thought experiment, what if a company only was growing at half the growth rate of a high growth company? But, you know, it's a marathon on a sprint. They could do that over 40, 50 years instead of thinking in these five and 10 year time horizons.

Do you have any good anecdotes on, I know you have this firm belief that hypergrowth is bad and actually slow, very long-term compounding growth is the real holy grail?

Yeah. I mean, I think what we're generally looking for is we kind of use Groupon as an example, whether that's fair or unfair of like, we're not looking for the next company to hit X revenue run rate in the shortest period of time or whatever it is. Like we're really looking for this durable, resilient growth. And I think that the way you framed it, Ben, that if a company does have employees that aren't driving the, I guess, level of growth that you'd be seeing in a hyper growth firm, that's okay as long as it's hyper durable. And so you could look at

I mean, look at Danaher, kind of like some of the classic iconic growth companies that have compounded for decades. Yeah.

And that's generally where you see the compounding. It says, obviously, you compounded 10% to 12% a year or in the teens, but you can do it for 10, 20, 30, 40 years. You can just get tremendous value creation. So I guess some companies in the portfolio we admire that do that would be someone like Texas Instruments, where they have a really decentralized culture and they actually push responsibility and decision-making down into the deeper parts of the organization. And so the CEO is not really a manager. He's a capital allocator. And he almost has to think more like an investor and a

and a portfolio manager than an operator. And so I think that's what we really look for is companies that can provide this durable growth. Again, it's very Buffett-like. We're hopefully finding companies where you can just set it and forget it, and they're going to put up moderate to healthy growth for 10, 20, 30 years. And in our framework, which we can talk about around resilience and optionality, that's what we're looking for in the resilient bucket of portfolios, these companies that can really compound it at a healthy rate for a very long time.

Yeah. So you guys have these two concepts and then one kind of super concept that combines both of them of resilience and optionality that you look for in investing. And neither of those are terms that most investors are familiar with. Can you define what you mean by both of them and maybe give a few examples of companies?

Sure. On the resilient half of the portfolio, we kind of say, you know it when you see it, which I think is kind of an unsatisfying answer. But generally what we're looking for are companies that are further along kind of in their S-curve and their growth trajectory.

And so this would be companies we own in the head of the portfolio, be someone like a Microsoft or a TSMC, where we're not looking for value stocks or kind of like cheap companies. We're looking for companies that are healthy growers that we think can durably grow for the next 20 or 30 years. And our turnover in this half of our portfolio is around 10%. So this is the hopefully set it and forget it part of the portfolio. And so then

They know a few characteristics we tend to see in that part of the portfolio are mission criticality and switching costs, which I know you guys cover well in some of the deep dives you've done. I mean, just scale like TSMC, we can talk about it in more detail. It's just like a classic portfolio.

scale company where we talk about power laws in our investing framework and pockets of industries where one company can take 90 to 95% of the profits in a given industry. And TSMC is a great example of that. And then you exactly the way you guys laid it out so well on your episode on TSMC is you really get the flywheel going, right? Where the more scale you have, the more you can reinvest, the more you can impact your customers. And it just becomes this

beautiful compounding machine. The last bucket we probably would see in the resilient part of our portfolio is just network effects-based competitive advantages where you see companies that really hit their inflection point. And again, especially in digital markets, you'll tend to see a handful of companies or potentially one or two companies take most of the economics in a given market like digital advertising or smartphones. There are just so many markets where there's one or two players that have 80% to 95% of the profits. And so we tend to see those in the

resilient part of the portfolio. And one thing I think was really counterintuitive about how you guys think is you're actually not looking for moats. The companies you just described, and I think a lot of investors think about like, oh, wow, well, they've got really deep moats. And so of course, you want those as your long-term compounding holds. But you guys have a little different perspective on this, right? Yeah, we have a little bit of a different take. And so I don't want to offend anyone that uses the term moat because I think it's a great term and it's a great part of anyone's kind of

investing toolkit. But I guess one of the things that we're careful about is looking for companies where part of their moat is inserting themselves into the value chain or into their customers' share of wallet, basically, where they put themselves in a position to extract as much economics as possible. And so I think that can be viewed as especially in kind of a more of like an industrial age view of how competitive advantage

has evolved, is that something that we try to avoid? We're not trying to look for a company where they feel like they have customer lock-in and then all of a sudden they can raise price 3% to 5% for the next 10 years. Part of our framework and the reason we named the firm NZS Capital is we're looking for non-zero sumness. So looking for a win-win outcome for all constituencies across the value chain. That includes the company's

employees and their customers and also kind of society and the environment at large. And so we're just careful looking for companies where all of a sudden, if I have this moat, then I can screw my customers over the next 10 years, right? I think that's what we're careful about. I noticed, I don't think you guys hold Apple, right? But you do some of the other large tech companies. And this feels like a perfect example to me of like, oh, Apple, like incredible moats. They're getting pretty good at value capture over there. Yeah. Yeah.

Yes, that's a very nice, polite way of putting it. You obviously see it with Spotify and the EU and Epic here in the US and being in the news flow constantly. I just think when you get to a point where you're taking so much economics for your business, which already has the largest market cap in the world, that your key partners on your platform are taking you to court or...

taking you up to various regulatory bodies or writing white papers on how you're screwing your customers. That's just what we're trying to avoid. And who knows, it might work out perfectly for Apple over the next 10 years. It makes you less resilient. Yes, exactly. Is that the idea that if there's consumer surplus money left on the table for consumers where they're not getting every dollar extracted that they could by the company, that that company is more resilient over time, even if they're not making every profit dollar and growing as quickly as they could today?

I agree with that. If you really have a management team that's thinking really long term, I don't know why you wouldn't give up a little bit of extra economics for your key partners, whether that's suppliers or developers on your platform or your customers to really solidify your trajectory over the next 10 to 20 years versus...

I don't want to pick on Apple too much, but what's the gross profit impact if they cut their app store take rate from 30% to 15% across the board? What is that, like 5% of gross profits? It's meaningless to them and it would create so much value. There's obviously knock-on effects of that. But anyway, I think that's what we're looking for as companies that are paying it forward. Again, back to the TSMC example, because you guys covered it so well. TSMC has lower gross margins than most of their customers.

And so at any point, they could probably take their margins from 50% to 60% and say, hey, I basically have a monopoly in this market. But the way Morris Chang architected the culture there is on long-term value creation and really creating a platform for their customers to create massive, amazing businesses. And so I think that's the way we think about it.

Yeah. And just for listeners to put some numbers behind this concept, which I think is just great from this NZS white paper, 15% growth over 10 years would deliver more than a 300% return. Not bad. But 15% growth over 15 years would almost double the 10-year return.

If we could populate our top 20 positions with these types of resilient companies, we'd only trim and add around periods of volatility. All the real absolute dollar value of compounding shows up in the out years, so you just want to make sure that you're still compounding in the out years.

That's right. And it's sort of where we take issue with Porter. This is, of course, Michael Porter, competitive strategy, Porter's five forces. Thank you. Yes. If we take a cynical view of Porter, a lot of people have interpreted it of, hey, build a moat around your business and then stick it to your customers on price, right? I'm not saying that's what he said. I'm just saying that's the way it's interpreted. And that's a terrible way to build a business because eventually someone will undercut that and offer actually a better value proposition.

to the customers. And because all of this value are in the out years, it was really not a value maximizing way to run the business either. So we think this concept of creating more value than you take is really important. And Porter agrees. He actually revised his thinking and in 2019, wrote a paper in Institutional Investor called Where ESG Fails, where he talked about this concept of shared value. And that's not...

him trying to say that purely because it's good for the world to care about all your constituencies, not just your shareholders, but also your customers and partners. He's literally making an economic argument for shareholders that that's the long-term value maximizing thing to do, right?

I think that's right. And, you know, he was a consultant for Intel back when ARM processors were starting to come out and actually dominate the mobile space. And they came out with this sort of dumped down processor. The Atom? Yeah, exactly. But even before that, they came out with a cheaper version of it. But in reality, that's not what they should have done. They should have actually embraced a totally different business model like ArmDead, where they were just selling IP and enabling a whole ecosystem instead of trying to take all the profits for themselves. Right.

Okay, so that's the resilience side of NCS thinking and the portfolio. Then you also marry that with something very different. Tell us a little bit about optionality and how you think about that.

Yeah, really, we're thinking about the future. We're just thinking about how broad and safe is the prediction we're making. So we can make these very broad, safe predictions. Like we think electronics are going to push deeper into the world, right? I think in nine out of 10 copies of the multiverse, that's happening. But there are other predictions. Like we think,

EVs are going to dominate the world and Tesla is going to be the power law winner inside of EVs. That may only happen in two out of 10 copies of the multiverse. It's certainly not 10 out of 10. So these predictions are much narrower and the range of outcomes is much broader. And so

That doesn't mean we can't invest there because it's incredibly asymmetric if we end up being in that copy of the multiverse. But if we're not and it's a zero, it also doesn't torpedo the portfolio. So I feel like actually you guys could give the masterclass here since you're such great venture capital investors. But that's really what we're trying to expose ourselves to in these earlier stage public equity companies. And

And so how do you actually then apply both of these very different principles inside the same portfolio? Are you picking some stocks because you're maximizing for resilience and you say that, look, this is a great portfolio?

compounding, slow growth, but durable company. And then there's other companies that you're investing in because you say, oh my gosh, if this thing's right, it's going to be really right. Like venture capital, asymmetric upside, right? Or is it blended in some of the same companies?

You're right. It tends to be these two portfolios in one. So we concentrate resilience. That's about 15 names in just over half the portfolio. And then we distribute optionality. So that's about 40 names. Also, just under half the portfolio, max position size, one and a half. And in the middle, it's between one and a half, two and a half, one and a half, three. We don't own anything. And that's percent of the portfolio? Thank you very much. Percent of the portfolio. I say this stuff so much. Sometimes I don't complete the sentences.

And then sometimes we find these very resilient companies that are actually layering on optionality to the business. So they have both. They have this resilient base, but then they have optionality on top of that. And you call those companies rootmos? Yeah. We're such geeks. It's so bad. Oh, we love it. It's the resilience with out-of-the-money optionality. It's just a shortcut on the team that we use. And those are the companies that you bump up to...

of the portfolio, right? Yeah. What are some examples of those? Well, I mean, sort of there's a couple, right? The classic example would be Amazon in 97 when they went public, you know, I think around a billion dollar valuation. Nobody could have foreseen AWS, right? That wasn't anybody's DCF. Oh, yeah, they're going to also create infrastructure that everybody in the world is going to use to create businesses. Yeah.

They sound so silly, right? But another one that we had in the portfolio years back was eBay. I don't know if you guys remember the marketplace business was struggling. They brought in a new CEO, John Donahoe. They had PayPal. And really, you weren't paying for any of PayPal. If the marketplace business would recover, that more than cover the cost of entry. Of course, marketplaces did recover. PayPal ended up being great. John Donahoe is an amazing leader. And that was a classic remote stock.

So in that situation, you've got a fairly resilient business, or the hope is that the marketplace is a resilient business. And then PayPal is the out of money option that you sort of have that's being valued at zero, but clearly is a very valuable business. That's exactly right.

What I think what's cool about this, yeah, right. This makes a lot of sense in investing. This also makes a lot of sense in how you should run your company. Yeah. And that the best CEOs think this way as in, how do I create resiliency in my core business? But then what are the options that I'm investing in for the future on top of it? To my mind, there's no better example than Amazon of just, this is the whole operating philosophy of the company, right? Yeah.

So we love this concept because we think it's true in the universe. And so therefore, the narrow slice of investing that we're using it for, we're pretty sure is also true. But it works really for everything. I mean, it works for parenting. Parents don't know what we're doing. Like I have four kids, I have no idea what I'm doing. So I'm just trying new things all the time. Well, that didn't work. Okay, well, you know, so there's a little optionality involved too. And then sometimes your optionality becomes resilience. And that's what you're hoping for. But you just try a lot of new things. So I don't know, for us, it's a life philosophy.

Fantasy football is another area where you can apply resilience and optionality very well, actually. You can't get out of your head once you start kind of practicing this, which is so funny. All right. I'm going to take us in a totally different direction. I want to talk about the difference between normal distributions and power law distributions. And listeners of the show who are in venture capital or in startups and have tried to raise venture capital or successfully raise venture capital will know that

They get this answer from VCs all the time that our portfolio construction is really a power law. We fully expect a third of the portfolio to go to zero, a third to return capital, and really only one or two companies at the head of the curve are going to be this hopefully 10, 50, 100x that gets us a great return regardless of what else happens in the portfolio.

You guys, interestingly, are sort of applying that thinking at much later stage companies, hopefully ones that aren't going to zero, you know, the way that a frail $10 million valuation startup could. How does that work? And what was the insight that made you realize, hey, the world is not normally distributed? Actually, in certain scenarios, it's very power law distributed.

Right. Well, once you accept the fact that all of life is governed by complex adaptive systems and the markets are also governed by complex adaptive systems, which means emergent behavior, you can't predict the future, you focus on adaptability, then, of course, those complex adaptive systems tend to be governed by power laws. So it's sort of a natural follow-on.

But the insight here is all risk models are based on these Gaussian, these normal distributions, right? But in fact, the world doesn't work that way. And so there's a really fascinating economist named Ole Peters, who's done a lot of work here and said, wait a second, your risk models are sort of like airbags that go off at stop signs, but not when you get in a crash, right? This has always been my beef, you know, when I was in business school with economics as applied to like

and investing in the real world is you study this stuff and you're like, wait a minute, I actually work in the industry and this is not how it works. Right, exactly. And being venture capital investors, you see this all the time. With public companies, it's also true. There are a few big power law winners. We see them in the market today. They're driving the entire market, right? This is our reality. There's a great study that you guys referenced in the paper.

paper that I'm wondering if you could just talk a little bit more about it. The toy example of a coin flip, a coin flipping contest, where let's say you win $50 or whatever, when you lose, you lose 40. That sounds like investing to me, like there's an expected return of 10%. What actually happens when you run that contest?

Wait, real quick before you answer. Is it that? Is it win $50, lose 40? Or is it 50%? Oh, it's percent. Okay. Yep. So if it's $100, then yeah, you're right, David. It's 50% down 40%. Got it. Okay. Yeah. The concept, it gets to the heart of modern portfolio theory and expected utility theory and the flaws of that. So back to the multiverse. The way this works is if you had 100 people flipping coins and they did it for enough time, you would actually see a nice, steady, positive return.

And that looks like on average, the experience of the participant is winning. But that's not really true. What you get is a lot of people going bankrupt and a few massive winners, sort of the Buffett's and the Soros's of the investing world, right? So the average is not average.

And in one portfolio theory, you're taking an ensemble of all these, but that doesn't really make sense because I don't really care about your outcome, David, or Ben, your outcome. I care about my outcome. And I only get to live in this one universe. I don't get to live in yours. I don't get to live in the multiverse. So my outcome on average is that of loss. It's that of bankruptcy. So when the time average does not equal the ensemble average,

That is called a non-ergodic system. And you guys can put in the show notes, Ole Peters works on this. It's super geeky, but really fascinating. That's so cool. Well, this is so counterintuitive. You would think if you presented that game to me, I would be like, oh, for sure I want to play that game. The odds are stacked in my favor.

but most people who play that game will lose. And then a few will win really, really, really big. That just blew my mind reading that. Yeah. The distribution set is not normally distributed. It's power law distributed. And so that changes everything. And that's why all these risk models are like airbags that go off at stop signs, right? It's because it

It turns out the world doesn't work that way. So, you know, we hear on a regular basis, oh, this was a three standard deviation event, you know, which if you understand the math of three standard deviation events, you expect, oh, wow, I'm so lucky to have seen one of these in my lifetime. But we see them a lot, according to the media. And so it's sort of ridiculous.

99.73% of all events should fall within three standard deviations is the way the math works. Meanwhile, I mean, I've been in this business for 13 years and already been through two recessions that are way outside of three standard deviations. Britain has as well. And so it is just kind of a funny kind of common sense thing that when you're practicing this stuff, the normal distribution doesn't

Doesn't really make that much sense. Britt mentioned I took this course at the Santa Fe Institute. It was actually around this time last year. And like, I mean, just power laws are just like so cool. It's amazing. I mean, one of the examples we use in the white paper is if you plot like earthquakes by frequency and intensity that just like in nature naturally forms a power law, which actually makes a lot of sense. You're going to get, you know, one or two or three heavy magnitude earthquakes a year and then a lot of small ones.

But also if you plot the frequency of every word in the book, Moby Dick, that actually also forms a power loss. The is mentioned 15,000 times. And then the next word is and, and that's like 7,000 times. And then there's like this super long tail of words that are only used a handful of times.

Again, it's one of these things, once you see it, and it's amazing for starting companies, obviously, like we mentioned, especially in kind of digital markets, because, you know, the world is going towards more markets where a winner can take all and, you know, becomes much more of a power law dynamic. And so that's why we just always have our kind of antennas up for these power law dynamics. That's honestly a big part of when we're looking for optionality. That's what we think about is like, is this a company that is relatively earlier stage in public markets that has the opportunity to power law, you know, a large market?

Which is why, of course, you're making 40 diversified optionality bets here. Because it's funny to think about this, but the statement of... Let's go back to the stop sign example. Yeah, the vast majority of the time, the fact that these airbags don't work is totally not an issue. But it's a massive issue the moment that you need them the most. Very similarly, all of the gigantic outsized economic value is created from the 3, 4, 5, 6, 7 sigma events in the world.

And so it's kind of ludicrous to be like, well, the vast majority of the time, this investment philosophy is very sound. And you're like, yeah, but we're not really trying to index on how many days out of the year it's sound. We're trying to index on how much value can get created at the end of the portfolio 50 years from now. And that's going to be driven by the outliers. So we have to be prepared and fully optimize around the outliers, not close our eyes to the few days that they might exist.

It's such a good point. I'm really glad you brought it up because what we're really playing for in the optionality half the portfolio is asymmetry. And really, it's not about batting average. It's okay if we're only right 30% of the time, which is not intuitive at all for public markets investors. I think everyone wants to be right 55% of the time, 60%, whatever. It doesn't have to be that high to have

good long-term returns. But we're playing for a slugging percentage where even if only one out of three work, but those are multi-baggers and can really create a lot of value over a long period of time, then that's the beauty in that half of the portfolio. And we've seen it. We've been doing this for years now. And it is amazing how you see these companies emerge as value creators and generate a lot of value for the portfolio out of relatively small starting position sizes. It gets back to this whole idea of you don't know what's going to happen.

If you set everything up with the idea that you don't know, I think in a lot of ways, most venture capitalists...

grok this idea and set up their portfolios in this way. But I think lots of people, myself included in the past, didn't fully understand this. You say in the paper, I think you use nicer language than this, but I'll use my own language. This is my quote, that conviction, this idea of conviction that so many people in venture talk about and entrepreneurs like I've conviction, I'm convicted, which convicted means you're convicted of a crime. But anyway, I have conviction that in this company, I'm going to lead this investment

Conviction is kind of stupid. Conviction is saying, I think my view of the future is going to be right. And really what you want is optionality. And you need people to have conviction because otherwise there would be no entrepreneurs, right? That example of the coin flipping contest, that is exactly the dynamics of becoming an entrepreneur. The expected value is positive. And yet the vast majority of people who start down that path go back.

go bankrupt. And then a few people win really, really, really big. But when you're constructing a portfolio, what you actually want is a lot of those bets. You guys have 30, 40 optionality names in your portfolio. As a venture fund, you want 30, 40 quote-unquote names in your portfolio. Venture portfolios with five or 10 are

are very non-resilient. Yeah, that's right. I mean, we use conviction as a synonym for overconfidence. I think that's what really is the right way to think about it. Conviction for us means, hey, I've done a ton of work, so I've got a lot of sunk costs, which means I've got bias. And I think that my view of the future is better than yours. That is literally what you're saying when you say that.

Yeah, right. So who knows, right? What we're trying to do with the tail of the portfolio, and I think what you guys are trying to do in venture capital investing is maximize the probability that we get lucky. I just ripped that off for Malbison. He's very good at calling it what it is. And we're just trying to maximize the probability that we get lucky.

I also think, I mean, just changing your mind is the hardest thing to do as an investor when you're wrong. And I think if you kind of stand in a row and say, this is my highest conviction idea, it just makes it that much harder to Britain's point on just like introducing bias into the equation. And so, and I also think if you kind of invert it, I think it's fine to have an optionality position that you'd actually don't have that high of conviction on like to Britain's example on Tesla. Like I don't,

think that we have super high conviction that Tesla is going to power a lot of the EV market, but is there some probability where they do that and it's worth multiples of what it is today? Sure. That's the way that we kind of think about conviction. And again, I don't want to piss anyone off that uses the word conviction similar to moats. It's fine. Everyone has their own process and we're not trying to push what we do on anyone else.

This is just what has worked for us over time. And so one thing we're very careful about is just introducing bias into our process. And luckily, everyone on our team knows each other well and can call each other out. But I do want to be a little careful. Well, I have two points to make. This cultural one, I think, is the second. But let me start first with

At the end of the day, this is a Buffett concept, this idea that it's better to be approximately right than exactly wrong. That's another way to describe this optionality phenomenon here where it doesn't sound nearly as strong to stand up in front of an investment partnership and say, I have very little conviction in this, but it could totally work. And if it does, it'll be really big. That's about the best I can tell you right now. That does not get everyone around the table excited, but in a very Buffett sense, it's

if it works, it's going to be so freaking successful that this is a great price to own it at. And like, do I know if this is the right price to own it at? Not at all. How many versions of the multiverse do we have railroads? Like all of them, right? And

And they're important. And how can we recreate that? We can't. It's impossible. And so I think you're right. Buffett and Munger say this so easily, so naturally, and we're just saying it in a much more complicated way. That's the second point I'm curious about. What are some guardrails that you have in the internal culture to reward non-conviction? To reward like, yeah, I don't know, but it could work. And if it does, it could be really big. And here's why it could be really big.

Well, the funny thing is the way we view team. So we think investing is inherently a team sport. It's a terrible solo sport for the most part. And the way we view team is our role is calling out bias in each other. Now, these are uncomfortable conversations because nobody likes to get their bias called out. But we all know that bias is really easy to identify in other people and really difficult to identify in yourself.

And so by opening yourself up to having being called out, then your probabilities go up as an investor. So it feels unnatural to us at this point to say, I have super huge conviction that this micro cap stock is going to rule the world one day, right? It's like, that would feel really odd. Everybody would be like, are you okay? You feeling all right? Yeah.

I think the way we've also set it up with our framework is we just inherently expect failure, I think, more than other public markets investors might. Like if we put something in the optionality tail of the portfolio, and by the way, that half the portfolio turns over a lot more quickly than the resilient part of the portfolio, which makes sense. Like we are going to be wrong a lot.

And luckily, there's more positions. And so you're not going to torpedo the portfolio as long as the most important thing to do is just admit you're wrong and move on. And so I think building that into the culture where, you know, it's okay to be wrong and move on and fail quickly versus like string ourselves along on a three year journey on a tough position. And so that's one cultural way that we've architected the way that the team works together. That has really helped. It basically gives yourself a license to take some risk that maybe you otherwise wouldn't take if you were sitting on a different team or within a different organization. Yeah.

And do you try and document, here's the reasons why I'm making this optionality bet so you know you can decide to rotate it out of the portfolio if those reasons are no longer true? Yes. Everything's written down. And actually, Brad...

which is an investor on the team, is amazing at pulling the stuff back up and saying, "You said blah, blah, blah." He must be really popular on the team. Oh yeah, he's great. We love Brad. He's just very good at remembering and then pulling the source data and saying, "Hey, look, you've drifted." To your guys' points too, it's not because their operating margin exactly 22% this year or as a revenue exactly at this run rate we thought it would be. It's like, are we approximately right on the thesis? It's not like we

I feel like we can predict the future, but you can certainly have checkpoints along the way. We call those usually with any stock, I think there's usually three or four things that really move the stock as we call those key leverage points on any position. And so you can generally check in on those and make sure that we're on track. Well, one of my big questions are things that didn't quite make sense to me.

in reading your paper. Can you talk about what you do with your optionality part of the portfolio as things evolve in it and how you start an optionality position and

there are two copies of the multiverse where this works and then X amount of time passes and you start to have more of a view of which copies, you know, two out of 10. And now, you know, maybe it's like two out of five or two out of three or like, you know, as it evolves, what do you do? Exactly. There's like kind of two scenarios that you can really see this happening. Like,

A good example is we own Peloton before the pandemic. And the stock obviously went parabolic. They were a huge beneficiary of work from home, but it's also just a really dynamic company that's early in its life cycle building a brand and a platform. And so with a company like that, I think it's still...

early days to call that business resilient for many reasons, both just like the context of the company of we're going through a digestion after the 2020 kind of a record year for them and off the charts here, I should say. So for a position like that, we'll just trim it. And, you know, we have this cap of how big in the portfolio we allow optionality positions to get. It's generally pretty clear, like how much of this is something that's a really durable inflection in the business. And sometimes I guess it is both like, I think Peloton is definitely a different company in this version of the universe versus the non COVID version of the metaverse. Right.

but I think we can't cross it over. Well, this is so different that you trim it. You know, the canonical VC wisdom is...

ride your winners as long as possible. The things that are working are likely to continue to work, so don't sell. But that's not the approach you guys take. Yeah, it's a really good question. We talk about it a lot because you're certainly, in some cases, leaving money on the table, I think, if we're not letting our compounders really express themselves over time. And so there are stocks we will let them, we'll own them earlier in their life cycle and let them cross over in the resilient head of the portfolio. And so we'll do that

a few times a year. And I think it does kind of force us to average up if the company, you know, and actually buy more stock potentially at multiples higher than our initial purchase. Oh, it's so hard to do. It's so hard to do, but I think it's actually, I've thought about this a lot. This is actually kind of part of our process where we're kind of forced to do it, which is really helpful because otherwise it's harder to just look at the stock and buy more. But we're saying we're making this explicit decision that we're going to take this from 150 basis point position to 250 basis points.

And so we're going to add capital because this business has structurally changed and actually belongs in the resilient part of the portfolio. And so there are companies where we've done that, where they're honestly just more mature or they're becoming like more of a platform. You can actually see like the network effects starting to hit. And honestly,

Some of that is a valuation conversation also, that there are plenty of platform-like companies we might want to own in the resilient part of the portfolio, but in the current market environment, they're trading at valuations that we would not consider resilient. And so that's another reason we would own them as optional positions. But it's a really good question. And I think what we try to do is not make sure that an optional position ends up in the head of the portfolio, because that's something we've just learned.

The hard way that if you have a stock that can have a 50% to 70% drawdown, and the starting point is a 5% position, not only does it crush your performance, but then you're also probably hamstrung where you've got a stock that's still a relatively big position, and you don't really want to add to it. And then you just kind of have to take your licking. And so that's something that we've learned through experience. Yeah. Well, I'm wondering if even just thinking about this past two-year COVID cycle, I'm

You've kind of seen this happen. The stocks that were huge multiverse winners in the beginning, the Pelotons, the Zooms, you know, and the like. I'm thinking Zoom. You know, Zoom went from, I don't know, what, $70, $80 a share to $600 a share. And then back down to, you know, I think it's at like $280 right now. So you've kind of seen this happen, right? The optionality played out. That was correct.

but then returns pulled back. We're always looking at like what's happening to the range of outcomes. Is it widening? Is it getting broader? Is the prediction becoming safer or is it remaining narrow? And so with a company like Zoom, it looks a lot to us like a feature. So now the question is, can it become a product and eventually maybe a platform? Can it develop an ecosystem around it? Yeah.

We don't know. But to take your example, let's say in the middle of pandemic, it was a sort of a cool feature. It was better than everything else on the market still is. And then this big ecosystem came around it and it became a full blown platform. Well, then the range of outcomes would narrow and the prediction would get safer, right? And so then that would warrant that becoming a bigger portion of the portfolio. Valuation is a key piece. And this is the piece that we get every day as public investors and valuations, expensive valuations, force predictions. I

I have to believe a lot more at 10 times sales than I do at 10 times earnings. So we're seeing, okay, what is the prediction of the company? And what is prediction the market is forcing us into? And are we comfortable with that? So we are in an unprecedented investment climate where everything on a, you know, whatever basis you want to revenue multiples, earnings multiples, unprecedented highs, you

any asset you could invest in, be it stocks or farms or crypto, is forcing you to make predictions. And what I've heard this whole podcast so far is you actively avoid trying to make predictions. So how do you respond in an environment where there's very little resilience in your ability to invest without making a prediction and have a margin of safety there? You just inserted yourself into the weekly NZS meeting, investment meeting, Ben, I think with that question. This is most of our dialogue.

Oh, just dial us in anytime. Yeah, part of this is interest rates, right? We've never had negative interest rates and then stimulus. And so those effects on all assets, which are unprecedented. We think about this a lot. We don't know the answer exactly. But this could get us into our top of semiconductors because there's a few building blocks of the information age that

and we are in an epic shift. We're still early days from the industrial age to the information age, and semiconductors are the new oxygen in this environment.

And so we look at some of these companies, we think the valuations are actually quite reasonable, and we choose to sort of bring the portfolio more towards resilience. And these are one of the ways we do it. We do it multiple ways. Yeah. Like you use the railroad example. There are 10 out of 10 copies of the multiverse in the future going forward where railroads are important. Right. There are probably also 10 out of 10 copies of the multiverse where semiconductors are important. Exactly. Yeah.

Well, that's an amazing way to transition to semiconductors. I mean, I was looking for the right hook. And, you know, Brenton, I think you bring that up. I was prepared to make some joke like, wait, you guys know something about semiconductors? I think TSMC is like a top three position for you guys. I think your other top positions, Amazon, Microsoft, they use a lot of semiconductors. And I think Salesforce is probably up there too. And TI is one of your top positions, right?

Yes, that's right. Maybe even before getting into some of the nerdier semiconductor topics, let's stick with an investment one. What semiconductor companies do you own right now in the name of resilience and which in the name of optionality?

Let's start with the two different versions of semiconductors, right? So there's a lot of semiconductor makers that are on the digital space, on the leading edge, right? They're making three nanometers and on, and these are the high compute functions. And there's other semiconductor makers that aren't really dependent on that leading edge. They're more dependent on having the breadth of a catalog. That would be like Texas Instruments that has 100,000 parts.

or a microchip. And so in our top positions, we're more heavily weighted towards the catalog names, these names that

The lifetime of a part is 30 or 40 years. And the margins are high. The growth is pretty good. There's clear NZS in the business. They're definitely creating more value than they take. And they're very hard to replicate, not because what they're doing is so technically hard. It's hard, but it's because the breadth of what they have would take you decades to recreate. I've always kind of hoped that Buffett would buy a catalog semiconductor business. I just feel like those are like just classic companies.

Buffett businesses where they're probably not going to look honestly that different in 20 years than they do now. They'll have higher margins and be bigger and they'll be selling into cool electronics that we don't even know about, but they're also still selling into water meters and coffee makers and just everything in your household or in a factory or anywhere you look just has these cheap but high margin chips in them. I guess to answer your question a little bit just on

the semiconductor impact on the portfolio. We have about a third of the portfolio in semis, and that goes across the whole value chain. Like to Brittain's point, we invest in kind of like the catalog, analog microcontroller companies. We'll invest in a digital company like NVIDIA that's actually in the optionality tail of the portfolio right now, just because...

for valuation and context reasons. We're big investors in semiconductor capital equipment, and those actually are head of the portfolio. We do those as resilient. TSMC is a resilient position. And then kind of the broader ecosystem, like Cade InstaDot and systems, which you guys brought up in the TSMC episode, that's kind of the key, one of the two key kind of cat software platforms for designing a chip is also a position. And so there are kind of more less household name type positions we own as optional positions, like a company like Cree, which is early in silicon carbide, which is an alternative technology

technology to silicon that's used in electric vehicles, including the Tesla Model 3. And so that's a classic example where there is some version of the metaverse where it's a massive platform and silicon carbide, the market goes from being a $1 billion market to a $30 billion market. But I don't know if there's a 50% chance of happening or 20 or that kind of thing. It's like a little bit of a walk around the portfolio in terms of semis. That silicon carbide thing is the first time I'm sort of hearing of it. Is that changing the substrate of the wafer?

That's exactly right. Yeah. Instead of using a silicon, like a bulk silicon wafer, he used a silicon carbide wafer, which is actually the wafer itself is much more expensive and it's very hard. It's one of these classic semiconductor processes where there's some black magic that goes into it. And honestly, most of the people that know how to do this are all in like the research triangle in North Carolina.

Cree is the one company that has two-thirds of the market for the substrate itself. And then they will sell the chips. They're the ASML of silicon carbon. So potentially, then that's a classic optionality, right? We honestly don't know. But there's a chance where either this stuff isn't that hard to do and they have a two-year lead on their competitors, or it's incredibly hard to do and they do become one of these companies where they're doing something that no one else in the world can do. And so that is kind of classic optionality for us. But it makes electric vehicles...

charging and also like renewable energy and really high voltage applications much more efficient. And so it's like really one of these companies where their core competency has just like all of a sudden the market really needs what they can offer. And so the market is growing extremely quickly. You're seeing a lot of activity around from other companies trying to get into the market as well.

All right, listeners, our sponsor is one of our favorite companies, Vanta, and we have something very new from them to share. Of course, you know, Vanta enables companies to generate more revenue by getting their compliance certifications. That's SOC 2, ISO 27001. But the thing that we want to share now is Vanta has grown to become the best security compliance platform as you hit hyper growth and scale into a larger enterprise.

It's kind of wild. When we first started working with Vanta and met Christina, my gosh, they had like a couple hundred customers, maybe. Now they've got 5,000, some of the largest companies out there. It's awesome. Yeah. And they offer a tremendous amount of customization now for more complex security needs.

So if you're a larger company and in the past you showed Vanta to your compliance department, you might have heard something like, oh, well, we've already got a compliance process in place and we can't integrate this new thing. But now, even if you already have a SOC 2, Vanta makes maintaining your compliance even more efficient and robust.

they launched vendor risk management. This allows your company to quickly understand the security posture of the vendors that you're choosing in a standardized way that cuts down on security review times. This is great. And then on the customization front, they now also enable custom frameworks built around your controls and policies. Of course, that's in addition to the fact that with Vanta, you don't just become compliant

Once you stay compliant with real-time data pulled from all of your systems, now all of your partner's systems, and you get a trust report page to prove it to your customers. If you click the link in the show notes here or go to vanta.com slash acquired, you can get a free trial. And if you decide you love it, you will also get $1,000 off when you become a paying customer. Make sure you go to vanta.com slash acquired.

Let's do a little sidebar, if you guys are game, that I just thought of that. I think this could be a really cool case study if you're willing to talk about it, about the NZS process. So you mentioned, John, that you own Cadence as a resilient position in the portfolio, having just done our TSMC episode and got a deep dive on the whole semi-industry infrastructure. You mentioned they also have a competitor, Synopsys, and the two of them, it's like a duopoly in the EDA space.

How did you decide to own Cadence? And I'm assuming not Synopsys. Or do you hold both? We've talked a lot about both of them over the years. We've owned Cadence for nine or 10 years back to our days at our previous employer. I think on Cadence, it's a few things. I mean, it's actually kind of a cool story of just kind of like how we even kind

kind of got into the idea of investing in EDA as part of our kind of process for finding new ideas and just kind of being up to speed on what's going on in the industries we follow is going to like industry trade shows instead of investor conferences. Like we don't generally go to a lot of like big investor conferences. And so in early last decade, I used to go to all these chip conferences and like every presentation, it was like someone from TSMC and someone from Arm and then someone from either Cadence or Synopsys. And at the time, Cadence and Synopsys were viewed as these sleepy, crappy companies. And

We love TSMC and we love arms. Let's do some work on Cadence. And then the more work we did, I mean, both companies are amazing. They deserve a lot of credit. I think what steered us towards Cadence, one is the management team. So Lipu Tan at the time was the CEO. He's actually moving into the executive chair role this year. He just was like one of the iconic leaders in the SEBI industry over the last 11 years. He was actually a VC previous to being the CEO of Cadence. And he was just on the board and had to come in and basically turn around the company.

but he was just so focused on the culture of the company. I mean, he told us what we wanted to hear, you know, which helped.

But in terms of turning around the culture and really taking a company that was in a very difficult position in the financial crisis and really like re-architecting the product positioning of the company. And he's so customer centric. That was how we kind of first got involved with Cadence. We do think they're taking market share, especially in digital markets, like where they would be selling to an Intel or an Apple or an NVIDIA. We think they're the share gainer, but both companies, it's like an extremely high quality duopoly. And so I think you've been fine either way.

And do you end up doing one-on-one meetings with the CEO at the level of capital that you're deploying?

Yeah, so this was back to our previous firm where there was more than $100 billion of AUM to deploy. And technology was a decent chunk of that. And so Cadence is a great example. We were actually their biggest shareholder for multiple years. And so at that point, we had really strong dialogue with them. And I honestly would just bump into the CEO in airports and at conferences because some of these aren't that big of a universe. Everyone's kind of going to the same things. And he's very tall. He's very tall, too. So it's easy to spot. Yeah, yeah. You're not going to miss him.

And he's honestly just like such a good person too. Like we would talk about life and kids and a lot more than just our investment in their company. And so it's actually something we think about a lot as NZS is a younger company with less AUM behind us. And we have a lot of relationships from being in the industry for a long time, but it's an open question of how often do you really need to talk to

Do I need to talk to the CEO of a company four times a year? Probably not. Or six times a year, 10 times a year, like the way we're investing, especially with a resilient company, realistically, maybe a check-in every year or two. Or if there's something obviously that's really critical to the thesis, we can check in. But that was kind of the way we grew up investing is a lot of management interfacing. I'm always curious with public market investors, like how do you think about that? And it sounds like you do find it very useful to have conversations with management, but

versus all the information's out there. I would imagine on the one hand, it's like, well, of course I want to know. I could glean so much more information and subtle signals from talking to somebody in person. On the other hand, I kind of think, well, I really care about what you do, not what you say. And I can just see what you do in your filings. How do y'all think about that?

This has changed a lot over the past decade because, of course, seeing a management team talk is easier than it's ever been, right? It's publicly available. There are some times when it's helpful. There are some times when it's harmful. You know, it probably nets out to be net helpful. But I'm just thinking of one interaction that we had with Rich Templeton, the CEO of Texas Instruments in early February 2009, right? It's a terrible time. Everybody's unhappy. It's really rough. And Rich walks in the room, big smile on his face.

How's it going, boys? A recession is a terrible thing to waste. And you're like, what's going on? I love it. And so he clearly had a different mentality of, hey, this is where we make all of our returns over the next decade. We're going to go buy equipment for pennies on the dollar. We're going to sort of systematically lower our CapEx to sales ratio. And we're going to go get customers and sign them up because we're running our fabs full out still and other people aren't. And

He's just one of these amazing leaders. You know, when he took the company over, they have 40% of the business geared towards wireless. And Nokia was a massive customer, their largest customer, and he bled that down to zero. So clearly this embodiment of a company that's built around adaptability instead of these point predictions. And so people like that are helpful to interface with. But honestly, we probably could get everything we need at this point without meeting with them as well.

David and I were explaining our research process to some friends the other day. And one of the things that I think that is chronically under-viewed on YouTube is presentations by executives at industry conferences. And that's a thing that we've relied on really heavily. Everyone goes and watches Elon Musk give his talk at the Recode conference. About 358 people watch the YouTube video of Gwen Shotwell presenting at an aerospace industry event.

There's a lot of really interesting information about the company, probably more so than the big shiny public facing stuff. That's exactly right. It amazes me. I remember the last time I watched the computer history interview is between Jensen and Morris Chang. Oh, so good. It was under 3000 views, something like that. I was like, how does this not have 3 million views?

You're like, am I watching the wrong feed? Or how could there not be more views of this? It's like the next no said, there should be a million views. It's so funny. We send stuff like that around all the time. I couldn't agree more. Brandon and I were sending some stuff back and forth earlier this week. It's mostly free too, and from industry trade organizations. And it's a huge resource. And you also, you do get a little bit of a different flavor if you see a management team at an investor conference, or if they're just even on a roadshow and coming through town, but you're the fifth investor they've seen

That day, it's like you kind of are getting the company line versus hearing what they're really pitching to their broader stakeholders at an industry conference is such a good resource. Yeah. When we're doing an episode just us, two, three, three plus hour deep dive on a company, we almost never talk to people actually at the company. Maybe we should, but we get all the insights we need from...

obscure YouTube videos, books, presentations, you know, white papers. To Britton's point, there's so much material out on these companies and stuff by management teams these days. And we do it in our, you know, basements. I think you guys are onto something with this acquired thing. We'll see.

Okay, so I'm going to take us into the more technical side of semis now. Britton, you sent an email when we were batting around topics and you said, well, what if we start by talking about the UFO crash that happened in Roswell in 1947, where we got the first semiconductor technology and then began to reverse engineer it at Bell Labs. Winky face.

It was the first tech transfer. That's exactly right. I mean, we know exactly when semiconductors came to planet Earth. It was July 7th, 1947. And then you needed a backstory. You know, when they took the UFO over to Area 51, they were like, how do we get this in the world without people knowing?

And, you know, of course, enter William Shockley. Fresh from the war, doing research on radar and submarine warfare, and he's already got top-secret clearance. We're like, okay, where could this come out of? Bell Labs, Shockley? Oh, yeah, that's it. That's the backstory. We'll give it to Bell Labs. That's the whole backstory in semiconductors. I think we're done. There you go.

Was Shockley... I actually don't know the history. I know he was super involved in World War II, right? He was, yeah. There's this great book, by the way, called...

the idea factory. It's the history of Bell Labs. So if anyone's interested in the history of the semi-conductory, you should definitely check it out. Not only the semis, but information theory from Claude Shannon, which came around the same time. Also from Bell Labs, right? Also from Bell Labs. Yeah. And so, yeah, in the 40s, he took a leap of absence from Bell Labs and did work actually with the Secretary of War on radar and submarine warfare.

Okay, and so just to keep pushing on this, the reason that this is a plausible story that we got these from UFOs is because...

The magic behind how a semiconductor works is so mind-blowing and unfathomable that you could just sort of experiment your way to finding this, right? That's sort of what you're going for here. Yeah, I think that's right. It really was this, this is overuse, but quantum leap. They had vacuum tubes. That's what the switches were made out of, right? It was one of the few places where there was still pure science being done at Bell Labs. And they said, we need this switch that doesn't break because we sent a lot of people out in the middle of nowhere to replace these vacuum tubes.

This could go on very long, but there was some key insights around doping germanium at the time. You're on the Acquired podcast, so it's okay, Brent. Indulge yourself. Right. There are these few key insights, and it wasn't just Shockley. It was two other guys, Bertain and Bardeen as well. And so the three of them together came up with these insights. And just right after the war, some of it was during, but most of it was just after. And they figured out, oh, you can dope this substrate germanium with

with different sort of n-type and p-type is what they're called. And when you run a current through it, it changes. So it actually does the switching in solid state. And this idea of solid state switching, which of course came about because of the transistor, and then later on was made to integrate a circuit by Kilbia TI, is what sort of enabled the foundation for all modern electronic devices. Over the last decade, I've read the Wikipedia pages for...

for transistor, for... I remember the first time trying to look up, like, how does a flash drive work? Like, I've got this cool USB drive and I put it in my computer and I read the whole Wikipedia page and afterwards I was sort of just blinking. Like, yeah, I still don't understand. This actually was not helpful. And it is one of these things where most of the time, especially having, like, a computer science education, I feel like I can connect

every building block to the next layer of abstraction building block on top of it, where eventually at some point after a few years of studying computers, you're like, wow, cool. I pretty much get how we go from physics to operating a operating system on a monitor. I understand all the building blocks in between, but somehow there really is something right around this layer where I never quite can jump from the physics to how it actually works and then how it manifests

in information and bits on a computer. I think I just need to go read a few more books, but it is one of these things where when you sent the alien joke, I was like, you know, you're right that I've just taken it at face value that this works, but I don't really understand how it works. Yeah. I told my daughter that this morning and she was like, wait a second, dad, that's really the way it happened. Right. And I was like, yeah, let's, let's back up. Ah, parenting. Yeah.

Okay, well, getting tactical here. So on our episode, I think we did a little bit of a high gloss shine on the story and the current state of the market, especially with TSMC and Samsung. We basically equivocated them and said they're basically doing the same stuff. TSMC is one to two years ahead. Obviously, Samsung has the whole consumer electronics division as well. Okay, that's TSMC and Samsung. And that was probably too simplistic. So one thing I was hoping from you guys today is helping us better understand who's good at what between those two

companies. Yeah, well, Samsung is an amazing company, probably not super well understood, maybe like TSMC, and people know them for consumer electronics and phones, obviously, but they have 50% of the market share in DRAM and about a third of the market share in NAND. So of course, as we do compute, we need more memory, we need a lot more DRAM, which is the fast memory on your phone or device or whatever, that queues stuff up. It's like a funnel, right? If you think of the funnel, you've got solid state or stuff sitting there and NAND flash, and

moves into DRAM, then actually goes under the chip with SRAM, which is a really fast funnel, and then it goes into logic to get processed.

So Samsung is very good at making memory. And like I said, they have over half of the market share in DRAM, which is incredible. There's really only three major companies in the world that make DRAM. Two are in Korea and one's Micron in the US. And then in Flash, they're big. They also have a decent foundry business. It's about 17% of the total foundry pie. So not as big. But DRAM and NAND are easier to make than Logic. And are these branded Samsung products or are they manufacturing them as a contract manufacturer? Yeah.

The DRAM and NAND is all branded Samsung, but of course, everybody uses Samsung. So it'd be next to impossible for Apple to get all the memory they needed without having a massive relationship with Samsung. So they're frenemies. So the four or eight gigabytes of memory in your iPhone...

That's coming from Samsung. That's coming from Samsung. Yeah, exactly. And so these are easier to make. They have fewer steps, but still they're very hard. So DRAM takes around 400 steps and over a month in the fab working 24 seven to make. And NAND has a little bit fewer steps than that, but actually NAND is getting more difficult because they're stacking it into 3D. So as you get more layers, it's actually getting more complex, but logic is still the hardest stuff to make. These system on chips that TSMC makes. And of course,

It's just imagine, you know, making one thing over and over versus making like a menu of what if you had to be a restaurant that made every kind of food on the planet, right? It'd be really hard to be good at all this food. So that's what TSMC is doing. So that's some of the difference. Samsung is also doing logic, but they have a different business model, right? They compete with their customers. So it's harder for their customers to trust them. Whereas TSMC doesn't have that conflict. Yeah.

What do you think makes for a more resilient company? Playing at multiple spots in the value chain such that you compete with your customers and have optionality, or being super pure play so that you have no strategy conflicts?

I think it depends. If you're talking about something in semiconductors, and Intel is the classic example of this where they're more vertically integrated. I mean, the hard thing about doing that is you have to fight battles on multiple fronts. Intel has to fight TSMC on process technology, which in itself is one of the hardest things any technology companies had to do over the last 20 years. And that's why Intel has been surpassed by TSMC, right? But they also have to fight AMD on...

in their core kind of like chip design market where AMD enabled by TSMC is innovating faster than they have in the last 20 years and, and like really delighting customers and taking share from Intel kind of real time or, you know, Nvidia where they were trying to just basically marginalize the CPU and make the CPU less relevant. So Intel is less relevant. And so I think,

That's the hard thing about being vertically integrated in semis versus being more of just like a horizontal pure play is the needs of Moore's law are just so difficult. It's hard enough to just do one of these things well and doing multiple of them well makes it harder. So I generally, I think Britain's point on just like the business model difference between Samsung and TSMC is so significant.

spot on because I mean, TSMC is like the neutral party that will never ever compete with their customers. And if you think about the amount of trust that the company has to put in TSMC because they're betting their entire company on TSMC's ability to make this chip for them and to have capacity for them when they need it. Just like the amount of trust and this has kind of been Morris Chang's

Hallmark, since he founded TSMC, that that's just something that Samsung can't quite offer because they just have a different business model and they're not willing to... Not that they're not willing, they just don't have the capacity to build kind of a massive foundry. They can't change, right? Like...

Like, they're not going to shut down two thirds of the company. Yes, exactly. How do y'all think about, especially since TSMC is such a large position in the portfolio? How do you think about the geopolitical risk? Because we did this whole big long episode and the conclusion I think we came to was this company is amazing. There's like no fault we can find in this.

Except that China might want to take over the land that they sit on. Except this enormous company ending risk. Even on TSMC's last earnings call, someone asked them what they thought about Taiwan's sovereignty. And I'm just like, what a world we live in that that's an open question that an analyst can ask on an earnings call. What do you think about China invading your country?

So, Britton and I are not geopolitical experts at all. I think we spend a lot of time thinking about just the importance of TSMC to the world. And I do think TSMC is top five most important technology platforms to the world. I think TSMC is more important than Apple. If Apple disappeared off the face of the earth, I actually think it would be painful for everyone that loves TSMC.

message and FaceTime, but it really would not be as big of a deal versus if for some reason China moved to seize Taiwan or however it went into play and all of a sudden TSMC's fabs were shut down, then the Western world would be set back at least five years, if not 10, just in terms of technology progress. And by the way, technology progress is driving most of GDP right now. And so you do read about what's happening with the auto sector and shortages

It's like you've seen nothing. If you think shortages kind of from the way that auto guys manage their inventory and kind of just like the classic post-recession semiconductor shortages you always get, you go from there to what would happen if TSMC, if Taiwan sovereignty was in question and TSMC stopped making wafers for some period of time. I just think the impact on the global economy would be extremely painful. And that brings you to the logical conclusion of

you know, hopefully the U S and the West would move to protect TSMC at all costs, or at least get the people out of there. We were, I mean, I guess I shouldn't joke about it, but it wouldn't be totally dissimilar to what's happening in Afghanistan, where I think you would just airlift as many TSMC folks out of there as possible in a short period of time, but then you have no fabs there to produce. I mean, TSMC fabs a quarter of the digital chips made in the world right now. And so it would be a lag of multiple years between, you know, when you get those people out and when you can actually start making wafers again. So it's a very complex topic. Yeah.

I think another way to think about this is just an ecosystem perspective. So like TSMC as a company is very valuable, but it's not super valuable without ASML and LAM and AMAT and KLA, right? So when you think about the ecosystem of semiconductors, and ASML, of course, is not valuable at all without TSMC and Samsung, there are this handful of companies, call it 15-ish, maybe more, that...

If you think of them as one super company, which is kind of what they are, it's like a super organism, right? Kind of like my bees are a super organism. It's like an ecosystem, you mean? Maybe like a complex adaptive system? It's like a complex... Yeah. So predicting the future is really hard. So anyway, yeah, if you think of it as a super organism, this is probably the most important super organism on the planet. If it's not, it's certainly one of the top, most important. So could you recreate that elsewhere in the world? You absolutely could. Yeah.

if you had access to the rest of these pieces, which in the West we do, it would just take, to John's point, a long time. And that's why people are sort of saying, maybe we should take some risk out of this place. And ASML calls this semiconductor sovereignty. We're building this fab, of course, in Arizona, this five nanometer TSMC fab, but it wouldn't be a stretch to think that they will be built again in Europe. This happened before, but of course,

When technology reaches a fairly stable state, you want to optimize around efficiency. So you get these horizontal type models, right? I remember when Apple bought PA Semiconnector, I was on record going, oh, this is the stupidest thing ever. Qualcomm makes these really good Semiconnectors. Broadcom makes them. TI makes a great application processor. And this is the origin of you deciding that you don't know the future and so you should. Yeah.

Yeah, what I said was the dumbest thing ever. That's for sure. But when technology changes, I just didn't have a concept for the smartphone. So technology was about to change quite a bit. And in that change, you really want to be vertically integrated because you're not pushing efficiency, you're pushing sort of product technical ability.

We see this with Tesla as well, vertically integrated. So can you imagine Ford having an AI day? That's just kind of funny, right? Well, they could have something then they would call it an AI day. I can see that happening. Yeah. Oh, man. MKBHD just did this awesome thousand mile road trip with a Tesla, a Ford Mach-E Mustang and a gas car. And it was just amazing. They're like, look, the Mustang, it's a good car. It's really good. But...

Yeah.

It directed us to the charging networks. It told us which stall to go to, you know, all this stuff. So I guess like the full circle question or answer to your question on like how we even incorporate the geopolitical risk on TSMC is, I mean, I guess it'll never be our like a 10% position for that reason is there is always this risk. But I also, I think it's important. I mean, my like kind of cheeky answer, that's probably not fair as if

if there is enough conflict between China and Taiwan, that TSMC, that they're like kind of, you know, business sovereignty is under concern or people are worried about them being nationalized by China or something. The whole U.S. market is going down. It's not just TSMC. I mean, I think they would be kind of like the epicenter, but it's not like this is going to happen in isolation. And so at that point, who knows if there could be, you know, more of a more World War III type global conflict coming from that. And so I kind of say like at that point, the performance of our TSMC common stock is probably not

my biggest concern that day, you know? Okay, so let's say TSMC is a resilience position. Let's say there is a black swan event, which the fact that we're all forecasting it and so is everyone else means it probably isn't that black swanny if their sovereignty gets challenged. So the point of optionality positions is to benefit from these black swan events. Do you guys have any ideas on if the whole US market went down in this situation, what could you hold that would hedge it?

The obvious hedge would be like defense stocks, but we probably wouldn't own them because of, I just don't think those are very high NDS businesses. But honestly, that would be the one pocket of the market that probably would fare. Okay. But Brent, I didn't mean to interrupt you. No, no, no. I was just going to say, if you've got to rebuild all these fabs somewhere, you're going to need a lot of equipment. And so you're leaving 20 years of equipment in the ground somewhere and you've got to recreate that. That would probably be pretty defensive against that outcome.

There's also, I guess, owning China tech companies. I mean, I guess it's questionable. If all this is really going down, I don't think China's... My ADRs are probably going to stop working. They haven't stopped working already? Do you have any China tech position? You know, Tencent or others? We don't. We've owned China tech for a long time. But when we put this portfolio together at the end of 2019, it just looked...

sketchy to us, honestly. I don't know, for lack of a better word. And so we own zero China tech. We just thought we don't have to be there. There's other places that are very interesting. And it's a question of ownership. We're not really sure who owns these companies. And through the ADR structure and the VIE structure, we know we don't own them. So we just set out. Yeah, makes sense.

Well, I want to come back to some more technical questions. Another thing that I think we kind of glazed over in our TSMC episode is the current state of Moore's law from a literal perspective, but then probably more interesting, the current state of the spirit of Moore's law. And I was wondering, maybe, John, let's go to you. Could you give us a little bit of a download on like,

Does Moore's Law still work, at least spiritually? Yeah, I'm glad the way you framed it that way, because it is kind of like a religious debate and people much smarter than me in the semi-industry are on both sides of, is the true Gordon Moore, Moore's Law still holding up? I think for the spirit of Moore's Law, we still have visibility probably for the next...

10 to 15 years. And to be honest, that's like the industry never has more than 10 to 15 years of visibility. I think obviously the death of Moore's law has been pronounced for a very long time. But I think that's one thing to keep in mind is there are a lot of things out there that's going to keep us driving down Moore's law. And so one thing that you guys covered well in the TSMC episode is the implementation of EUV systems and

from ASML and they actually are allowing us to shrink the transistor, kind of the fundamental building block, you know, two-dimensionally. So it actually put more transistors into a chip. And so ASML is kind of on record saying they think that the EUV will last about 15 years.

In terms of they'll allow us to keep doubling the number of transistors on a chip every 18 months for 15 years, or just it will be an effective way of getting any performance? Yeah, I think it will be an effective way to drive performance and like drive shrink, basically, is the way the industry kind of frames it is you'll be shrinking the transistor to pack more performance into a chip. But I think the broader point you hear from the industry a lot is this concept of more than more, which is kind of a cheeky pun. But yeah, there you go. Yeah.

Those semi guys, they're real hoot. Exactly. So geeky. But so the broader point around where Moore's Law is going now is it's not just about the transistor. It's really about the package. And so you're seeing a lot more innovation, not just in like, can we put more transistors onto one gigantic chip to drive more performance? You can actually split up

you know, chips into multiple chips called chiplets, which AMD is doing. And this is a big part of Intel's future strategy actually also. And so having like, you know, one gigantic like GPU that you would buy from Nvidia, you can have four smaller chips and you can kind of like stitch them together to drive more performance. And so that's,

It's another way that we're going to get a lot of benefit from Moore's law. And then actually one thing I skipped over on the transistor side is we are moving to a new transistor architecture, either a two nanometer or three nanometer, depending on which company you're talking about, to a gate all around architecture. Previously, we were on FinFET, which has been around since, I guess, for the last seven or eight years. I don't know the exact numbers, maybe 10. There's line of sight into from here to

more gate architectures, different materials. And then when we get to like the mid 2030s, we'll kind of see how it goes. But it's just, it is amazing. I was actually at a virtual chip design conference this week, earlier in the week. And there's so much focus, not just, I guess, on this like packaging idea, but even like if you have to extract that one more layer from that, it's about system level performance. And if you look at what NVIDIA or AMD and kind of like the leading digital companies are talking about, it's can you make more processors and actually like stitch them together into a cluster with some sort of proprietary interconnects.

They're really thinking more like computing companies than just like chip companies now. And so that's a big change. But all of these transistor level innovation, package level innovation, and then system level innovation, I think we've got a pretty good line of sight into the spirit of Moore's law continuing at least through kind of the mid 2030s.

And so it seems like with the sort of creation of the system on a chip that at least let's just talk about the iPhone because it's the one I understand the best. Apple became the aggregator rather than the old days of I'm going to go build my computer and I'm going to go buy a motherboard and I'm going to buy a GPU and I'm going to slot it in the PCI slot and blah, blah, blah. Yeah.

Apple basically says, well, we've designed this logic board and we've designed most of the chips, the important chips, and we've situated them together. TSMC manufactures it. They do all the packaging. So you have like the bare metal to bare metal packaging of these things. So we don't need to run it through these buses that have low bandwidth to get information from one piece to another. Let's keep playing that out a little bit based on everything you know. Where in the value chain do you think the point of aggregation shifts to over time where

Who gets to own where do we put all this stuff together? And I get to capture a lot of extra margin because I'm the one putting it all together. I think that's the interesting thing about the semiconductor ecosystem is actually there's a lot of people capturing margin and they're capturing really high margin. So this is the sign of a healthy ecosystem, right? There's not one company that's making all the money throughout the whole chain. We've seen margins come up and

you know, here's a good trivia question. Who has higher operating margins, Texas Instruments or Microsoft? Right? Because I ask it, you know the answer is TI. Wow. But, you know, it's not really appreciated how good these businesses are. That's shocking. Right? Yeah. Especially because TI, the core business, is not the leading edge

digital processors that TSMC is doing. It's the commodity stuff, right? I think this was Rich's key insight. He said, okay, we're doing all this leading edge stuff. We're fabbing at TSMC. And what if we just trickle that business down to zero? They tried to sell it. No one wanted to buy it.

and boring is beautiful. And so you look at TI's end markets, two-thirds of which are industrial and auto, and I guess it's a tech company. They make chips, but man, it seems a lot like an industrial company too, right? Yeah, I think you can make the same point that you made on TI on NVIDIA, like the fabulous business model. Really, I mean, it's like a software company. It's like NVIDIA has got close to 70% gross margins and low 40s operating margins. Really, with

very little cyclicality because TSMC offloaded all that cyclicality, right? So it is just like one of the best business models besides, I would say, enterprise software. It's one of the best business models in the world. And so TSMC is the enabler of that. But I think Britton's point is spot on. You're going to see kind of in this future world, one of the things that's kind of cool is it takes the whole ecosystem to really drive kind of the future of Moore's law. Like it used to be just about ASML needed better litho tools and like

Intel would use them and shrink the transistors, and we'd just brute force ourselves down Moore's law. And now all this advanced packaging needs litho advancement from ASML, but it also needs improvements from the other equipment guys, like LAM Research or Applied Materials or Tokyo Electron, because you need deposition and etch steps to build these advanced packaging and multi-die packages for the advanced packaging applications.

You also have all this off-the-shelf IP they're getting from companies like Arm or from Cadence or Synopsys. Part of what Apple does, to your point, Ben, is they're really just an aggregator. They buy IP blocks off the shelf. A lot of what designing a chip is about is just buying a lot of individual IP blocks and aggregating them. The great thing about it is since Moore's Law is really freaking hard, everyone in that ecosystem does really well. I think TSMC, sitting in the middle of it, will obviously do very well because they're driving a lot of the innovation also and obviously are the key partner for a lot of this. Real quick on...

TSMC. I think we glossed over this on the episode because we weren't deep enough to understand it. My sense is that the open innovation platform that they've created is really important and is kind of what orchestrates all of what you're talking about here, that it really takes the village of the whole industry to push things forward now. Is that true? What is that and how central is TSMC's open innovation platform to all of this? That's exactly what I was going to say. I think it's really true. And there is no GitHub to the

semiconductor IP ecosystem, right? The closest you get is kind of the TSMC Open Alliance. And there's pockets of it elsewhere. The EDA guys have a ton of IP as well. And of course, as you make these chips, you need to emulate them to see if they actually work, hopefully before you put them in the fab, because that's really expensive. So

All of these things really play together. And so when you think about how Intel was doing this for a long time, it was a closed system. It was Intel's way, Intel's process flow. And TSMC said, oh, wait, let's form an alliance with everyone because this is really going to take everyone to keep driving this forward. And this open architecture or this open approach has really won over.

One thing that's always hard for me to understand is how hard it is to do each layer of the stack. And by that, I mean, wow, it seems like the ASML guys create a pretty unbelievable machine and they have a lot of services associated with that machine. Like they're even in the TSMC factory helping to assemble and operate these things. And then I sort of

scratching my head thinking, well, could ASML just kind of like become TSMC? Could they just operate their own equipment? And then I dove down the other side of the slope and I was like, well, who makes the stuff that's important to the ASML machines? And I was like, what is this Trump company doing?

And then, of course, you go on the Trump website, which let's just let the name lie for a moment here. And they make this unbelievable laser. And they've got this crazy video on their website that shows off their laser. And I'm pretty sure what they're showing me is actually the magic of the ASML EUV machine. And I'm like, well, shoot, why can't the Trump company just do what ASML does and then also do what TSMC does if they are the only ones in the world who can make this

Unbelievable laser. Can you guys shed any light on is that ever going to happen? Could it ever like vertically integrate? First of all, the videos of the simulations of the laser in an EV system on the Trump website are like so freaking cool. They're amazing. Yeah.

I hadn't seen them until recently, and I've heard so many times that like 50,000 pulses a second drops of molten tin, the whole spiel from ASML, which Ben, by the way, you did a very good job on the TSMC episode. I was doing my best John Bathgate impression. I could tell you were excited to do it. I think like multiple times you're like, can I do the ASML thing now? But anyway, a few things to think about. One is there's so much innovation. So the laser itself actually is

10 tons, but an EV system is 180 tons. So there's a lot of other equipment in there that's not just the laser. And actually, I was trying to find this from a dollar value. I didn't track it down, but I'm sure that that number is out there. ASML's partnership with Zeiss, the lens company, is also really special. And I guess

Trump and Zeiss could try to partner to... I don't think they have any ambitions to do this, but they could try to partner together to kind of circumvent ASML. But the lenses that Zeiss is coming up with are literally the most uniform lenses designed in the history of the world. And some of the metrics that they throw out on that are incredible. And I think...

One of the things that's unique about ASML is they shipped the first EUV tool in 2010 to TSMC. And EUV didn't even really start high volume. And this was after a decade of R&D already. But then they didn't start high volume manufacturing on EUV until 2019. So they had almost a decade of learnings in TSMC's fabs on how to get these things to actually work, how to get the throughput to levels where the economics actually make sense.

And so there's actually this really cool conference called SPIE every February where all of ASML's customers come together and basically give feedback on UV and kind of give the updates on where they're at. And so ASML lived through 10 of those with all the feedback, not just from TSMC, but from all their ecosystem partners, right? And so I just feel like the learning cycle that ASML has been through, there's just so much more innovation in addition to the laser and the...

The lens is, but it is, I mean, the lens is a really critical component. I mean, ASML actually bought a laser company in 2012 called Symer. And I think they actually had like an internal laser bake-off between Symer and Trump. And I think Trump won for EUV, which is also kind of a funny, you know, trivia question. Wow. Fascinating. Okay. So the answer is they all add a ton of value on top of the previous layer of the stack.

And like a lot of things in the tech ecosystem, there's fractals on fractals, right? It's like you say, okay, well, what's the most important part of this ASML tool, right? And there's a fractal down and then like, you know, you keep going. But it's really this ecosystem approach that makes sense. No one can do all of this. It's just way too hard. It really does take a village.

Our sponsor for this episode is a brand new one for us. Statsig. So many of you reached out to them after hearing their CEO, Vijay, on ACQ2 that we are partnering with them as a sponsor of Acquired. Yeah, for those of you who haven't listened, Vijay's story is amazing.

Before founding Statsig, Vijay spent 10 years at Facebook, where he led the development of their mobile app ad product, which, as you all know, went on to become a huge part of their business. He also had a front row seat to all of the incredible product engineering tools that let Facebook continuously experiment and roll out product features to billions of users around the world. Yep.

So now Statsig is the modern version of that promise and available to all companies building great products. Statsig is a feature management and experimentation platform that helps product teams ship faster, automate A/B testing, and see the impact every feature is having on the core business metrics. The tool gives visualizations backed by a powerful stats engine, unlocking real-time product observability.

So what does that actually mean? It lets you tie a new feature that you just shipped to a core metric in your business and then instantly know if it made a difference or not in how your customers use your product. It's super cool. Statsig lets you make actual data-driven decisions about product changes immediately.

test them with different user groups around the world, and get statistically accurate reporting on the impact. Customers include Notion, Brex, OpenAI, Flipkart, Figma, Microsoft, and Cruise Automation. There are like so many more that we could name. I mean, I'm looking at the list, Plex and Vercel, friends of the show at Rec Room, Vanta. They like literally have hundreds of customers now. Also, Statsig is a great platform for

for rolling out and testing AI product features. So for anyone who's used Notion's awesome generative AI features and watched how fast that product has evolved, all of that was managed

with Statsig. Yep. If you're experimenting with new AI features for your product and you want to know if it's really making a difference for your KPIs, Statsig is awesome for that. They can now ingest data from data warehouses. So it works with your company's data wherever it's stored. So you can quickly get started no matter how your feature flagging is set up today. You don't even have to migrate from any current solution you might have.

We're pumped to be working with them. You can click the link in the show notes or go on over to statsig.com to get started. And when you do, just tell them that you heard about them from Ben and David here on Acquired. Great. Well, John, I know you have a ASML story that you want to share, so I'd love to hear it.

Yeah, I think this is a really cool story of how ASML came to be. ASML and so strategic to the world is, you know, they had one competitor, which is Nikon in the lithography market. And then Nikon gave up on the market kind of coming into the last decade. And so I think TSMC and Samsung and Intel kind of looked around and realized, like, we're betting...

the future of Moore's law and this one company, ASML, which at the time was like kind of a sleepy Dutch company that would have like a $20 billion market cap. Like no one really knew who ASML was in 2012. And so it was so unique is that Intel and TSMC and Samsung partnered and actually bought 25% of ASML to inject capital into ASML to develop EUV systems. And that started really like the iteration path of developing EUV to get EUV to where it needed to be for high volume manufacturing by the end of the decade. And so it's just such a cool story of like,

the ecosystem coming together and everyone in the semi industry knew how important ASML was at the time, but the world didn't really understand that. And so getting ASML where they needed to be on a UV obviously is now enabling Moore's law for the next at least 10 years. And is it right that they've all largely divested at this point?

They have, yes. Yeah, I mean, they honestly should have just held on to it. I mean, it would be not immaterial. So especially like Intel's enterprise value. There's so many stories like that. Like you guys covered the ARM origins, like all these like major ecosystem players have all like had, you know, stakes from other companies at various times, which is just kind of a funny way that the semi-industry has worked. I couldn't find in our research, how much of TSMC does the Taiwanese government currently own? Because they started by owning 50% of it.

That's right. I looked for that too recently and I couldn't find it. I want to say it's still in the 20s, but I could just be making that number up. I thought I actually was going to say 20 off the top of my head, but maybe don't quote us on it because I don't know if that's true or not. Wild. Could you imagine if the US government owned 20% of Intel? I know, right? Or Apple? They owned a lot of Ford at one point. That's true. Yeah.

Yeah. And sometimes people ask, well, why can't just another company buy TSMC? Well, it's a national champion. The government owns a bunch of it. It's just impossible. Right. Some things don't have a price at which they are for sale. Exactly. All right, Dan, you had one other trivia question for us. Yeah. We were talking about, I kind of made the point that fabulous chip companies are kind of the best business models in the world. I think one of the things that's so cool about TSMC

If you just look at all the value they created, like NVIDIA is a half a trillion dollar company and then add up their next few biggest customers like Qualcomm and Broadcom and AMD, that's like another half trillion. And then I was trying to think of like their aggregate value creation. And so I was trying to think about, I mean, Apple is obviously the biggest customer. And so I think in an acquired episode about the top 10 acquisitions of all time, you guys assigned a value to PA Semi and how much of kind of like Apple's differentiation value.

is driven by semiconductors. And so if you guys know that off the top of your head, that'd be my guess for kind of like Apple's contribution to the TSMC value creation for the world, if that abstraction all makes sense.

Oh, man. That was the most hand-wavy part of that whole analysis. I kind of feel like we ascribed half of Apple to Next and then 10% to PSMI or something kind of arbitrary. We were literally carving up the Apple. Well, here's the framework. Here's the reason it's hard. It's because the notion of necessary but not sufficient is really hard to frame into a percentage.

Apple would be worth zero if they didn't acquire Next. But does that mean that Next is responsible for 100% of the value of Apple? Absolutely not. So what percentage do you assign it? It's tricky.

Yeah, that'll make sense. I actually don't know that the number you guys use, I thought it was 25% off the top of my head, which maybe that's a good rough number that, you know, Apple's contribution is, you know, another half trillion dollars to the TSMC value creation story, but you get the broader point, right? It's just, it's like trillions of dollars of market cap that TSMC has created for their partners, which I think is just so cool. Yeah.

Yeah, it's amazing. I mean, it truly meets the Bill Gates line of the definition of a platform that they've created way more value for their customers and their ecosystem than they've captured for themselves. Yeah, which is so cool because it's obviously not like a traditional internet or e-commerce platform the way like most of our SaaS platform, the way we think about platforms, right? It's just like a manufacturing platform, which is just so unique. There's one point that

you're getting at here that is part of the white paper and what we discussed earlier, which is around sort of leaving money on the table for your customers and leaving money on the table for your partners. And it reminds me a lot of when we did the Altos episode with Honam. And it's really this idea that if you as the management team, or if you as an investor who deeply understands the company,

knows that in a way that other people outside the company can't underwrite, then you can do a much more intelligent job

valuing the company than anybody could with a brute force metric such as industry average earnings multiple. Because if you actually understand, well, our earnings could be this if we wanted it to, or our growth rate could be this if we wanted it to, but we're making strategic trade-offs to not do that, then you actually have

unique ability to underwrite the company's value and thus actually more of a margin of safety or more of a willingness to pay up than anybody else. And so it's interesting being deeply studied about these companies where you do know that they're sort of leaving something on the table for other participants that you can be more comfortable making an investment than other people can.

I think that's a really insightful point, Ben. And the thing it gets to for me is duration of the asset, duration of the growth. So when you leave money on the table, what you're doing is you're creating goodwill for your customers and you're buying the company duration, which is oftentimes the way to maximize total value, right? So when I think back about Ho talking about Roblox, he was effectively saying,

we just really understood how big this ecosystem could become. And we kept seeing the value accrue and then it moved beyond our original investment case. And therefore we became more comfortable investing more money over time. And what people get wrong oftentimes is this duration. Because duration, if you can go 15% back to your earlier example, it's extremely nonlinear if you can keep that flat, right? All the value comes in the tail. And so we just aren't very good at thinking like that. Our brains don't work in that nonlinear fashion.

But when you create more value than you take, and if that's your driving factor, and you want to take a lot of value, it's a hard task because you have to all the times think, oh, wow, we want to take a lot, but we need to create even more. How do we do that? And then, of course, that buys duration, which is a feedback loop, sort of the happy feedback loop, if you want to think about it that way. I think Morris Chang got this very early on, and that's what created TSMC into such a great company.

It's so good. We touched on this a little earlier, but just to double underline, one of the things about you all and your ethos that was kind of an aha moment for me is flat growth versus hyper growth. Flat growth extended over time will beat short-term hyper growth. You mean the derivative of being flat, right? That a company grows at the same rate every year. If you grow 20% a year for like 50 years, like TSMC, you will destroy...

every Groupon out there. So what you need for that is a negative feedback loop, right? So the negative feedback loop for TSMC is, I'm going to come in, I'm going to take what used to be the special sauce of your business, and you're going to trust me to do that. That's extremely hard to do, right? No one wants to do that. But then the more it happens, eventually there's a game theory to it. Everybody has to do that eventually because it works so much better, right? So you're never going to get 100% growth. It's impossible. Right.

But you might get 20 for 30 years, which I think the number that you guys said in your podcast was 17.7 for 30 years or something like that, which is just incredible to me.

And when you say negative feedback loop, you basically mean a governor on the growth, like a natural force in that particular business that makes it so you can't have ludicrous Uber-style hyper growth. And it ends up being long-term good for the company to have that growth governor or that negative feedback loop. That's exactly what I mean. So when we think about ASML, right, they can ship everything they can make, but they just can't make anymore. It's impossible, right? So there is a governor on the growth, right?

Wow. Well, that's a great place, I think, to leave, especially this Emmy's discussion and most of this episode, that it's so counterintuitive, but the way that you've sort of framed up why it is long-term good for an investor to want slow, methodical, governed growth is just very different than a lot of the things we talk about on this show. One of the things we think about sometimes is we're looking for companies that can double in five years and double again the five years after that.

And all that means is we're looking for companies that can grow 15% over a decade, all else equal.

Easier said than done. Easier said than done. Much easier said than done. I will just say one thing about our days. We get a lot of questions of, well, how do you do this at a small company versus a big company? And part of it goes back to what we were talking about earlier. It's easier to do research in these companies now than it's ever been. But the second piece of it is, yeah, there's more details we have to deal with at times. But how much extra time would you have in your regular job if you only had two meetings a week and you never had to worry about office politics?

you know, my guess for most folks is it's about 20 hours, right? And so then how would you use that? Well, we just use it for unstructured research time. And the way we think about it is we can wander around not knowing what we're doing and waste 90% of that time. And 10% might be really useful and 1% might be absolutely watershed. And that's really all we're looking for. But you can't ever just get to the 1%. You have to wander around to find it. And so that's how we really structure our days.

And Brendan, that's why you run for like 24 plus hours straight. And that's why I run and that's why I keep bees. It's all the same thing. It's where your best investment ideas come from. That's right. I do think this idea of like linear time versus nonlinear time is really...

interesting and something Brittain and I talk about a lot because like the linear time it's like it's probably very similar to you guys getting ready for your next episode it's like you're going down the rabbit hole on one topic and we do spend time doing that obviously but then we just have so much extra time to like just be out there trying to connect dots and so you never know when you're gonna have that aha moment or that that insight but having you know as much opportunity for that as possible is kind of how we've intentionally tried to structure our time.

I love it. Well, we could spend another hour on how you structure your time. I was actually wondering, maybe we can get one or both of you, if you'd be up for it, to join our next LP call and I'll chat about it. I'm sure folks would love to pepper you with questions and hear about how you spend your time. That'd be super fun. We always learn from those questions, so we'd love to do it. Awesome. Well, Brenton, John, where can folks find you on the internet?

You can go to nzscapital.com. We do try to write a lot. We put it all on the internet immediately. It is everything that we use internally. Nothing's held back because we know when we put it out there, we're going to get more value back. So this is our way of trying to create more value than we take.

Our partner, Brad, also writes a newsletter every week. And so if you want to see how he spends his time, you can sign up for the newsletter on ncscapital.com. It's called Sit All Week. It's just Brad's process of sitting all week and what he thinks about. And Brad is, he's like a microprocessor. He's literally the smartest person I've ever met. And the way his brain works is incredible. And so if you'd like to sign up for that, you can do that there as well.

Great. And on Twitter, I think you both have Twitter handles. That's right. The whole team is on Twitter. NCS Capital has a Twitter account. I'm bjohns3. Brad is at bradsling. And John is at jbathgate. Great. We'll link to all those in the show notes. Thanks for having us, guys. Thank you. All right, listeners. Hope you enjoyed our conversation with Brinton and John. If you found it to be

frankly, as eye-opening as I did. Feel free to share it with a friend. If you learned something new about complexity investing or their barbell strategy or resilience and optionality, I'm sure you can think of someone that you'd want to share that with. So feel free to do so. If you are not already a member of the Acquired Slack, come join us, acquired.fm slash slack. Some of the best discussion you'll find on the internet about

Lots of things that you care about. If you're not an LP, you should become one. It's a way to get closer to what David and I do here at Acquired. We have these awesome LP calls.

Once every month or two. We have been on a tear recently with great LP exclusive content. We just dropped about a month ago now, a great interview with Ronil, the CEO of Audius, which is the largest crypto application, Web3 application out there with over 6 million users. And then just before this episode, we dropped another sort of Web3 episode, this time on Web3 marketplaces centered around brain trust. With...

Adam from Braintrust. Wow, that one was a blast.

Yeah. So if you're Web3 curious, or maybe you're Web3 skeptical, these are fun episodes to listen to because they're super non-DeFi, non-crypto use cases. Yeah. They're like real world applications. Yeah. Which, of course, we wanted to dive in and tell those stories. You know what I'm curious about after this episode? And I feel like the universe works in funny ways, right? I think everything's sort of aligned that a bunch of stuff happened all at once.

It's the Santa Fe Institute and complex like everything we talked about on this episode. It just seems like such an amazing place. I've always wanted to go to Santa Fe period. I've never been there. But to go like taking class there like we should do something because we before this we had our Michael Mobison episode. He's the chairman of the board there where kindergarten is actually investing in a company that the CEO is also a board member there. Just like there's too many stars aligning. We got to go. The universe is telling you something. I think it is. I think it is.

Well, with that, we will see you next time. We'll see you next time.