Andrew Homan & Chris Miller - Redefining Semiconductor Progress - [Invest Like the Best, EP.395]
01:26:22 Share

Two fund facts about our newest sponsorship partner, rap. First, they are the fastest growing fin tech company in history, reaching a level of revenue in five years that I can't quote exactly, but is eyebrow raising. Second, they're backed by more of my favorite past guests, at least sixteen of them when I counted.

Then probably any other company that i'm aware of, a list that includes ravi gup, dead aoa, josh, Christa, thrive key or boy, a foundation conventual at john Michael ovis brag garson. The list goes on and on. These facts demand the question why, having been personally obsessed with the great businesses through history, one clear lesson is that the best of them are run by disciplines Operators.

These Operators managed costs with incredible detail, and they are constantly thinking about how they can reinvest every dollar and every hour back into their business. This is rams mission to help companies manage their spend in a way that reduces expenses and freeze up time for teams to work on more viable projects. First, on expenses, the average american business has a profit margin of seven point seven percent.

This means saving one percent on cost is the equivalent of making thirteen percent more revenue. The average ramp customer is able to save five percent on their expenses each year. Of course, every entrepreneurs looking for ways to grow revenue by fifty percent, they should just seriously seek to save five percent on their expenses.

Second on time. Unnecessary complexity is why most finance teams spend eighty percent of their time doing Operational work and only about twenty percent of their time on the strategic work. Rap may spend management very simple by handling your company's expenses, travel bill payments, render relationships and even accounting.

It's notable that some of the best in class businesses today, companies like airbnb, Andrew and shop fy and investors like coa capital invest equity, are all using RAM to manage their spend. They use IT to spend less. They use IT to automate T, D, S, financial processes, and they use IT to reinvest, save dollars and hours into growth at both classes and positive.

Some my businesses we've used ramp for years now for these exact reasons, go to ramp dot com slash best to sign up for free and get a two hundred and fifty dollar a welcome bonus that's R P 点 com slash invest。 This episode is brought you by tiguas, where were changing the game in investment research, step away from outdated, inefficient methods and into the future with our platform probably hosting over one hundred thousand transcripts with over twenty five thousand transcripts added just this year alone, what that tickets apart. It's not just the sheer of volume, it's the unmatched ed at which our library expands consistently outstripping competitors.

Our platform grows eight times faster and adds twice as much monthly content as our competitors, putting us at the forefront of the industry. Our collection is investor LED, ensuring unparalleled quality and giving you access to questions in topics investors careless about. Plus, with seventy five percent of private market transcripts available exclusively on tickets, we offer insights you can't find elsewhere.

Forget the traditional way of doing things with tickets. You have the most comprehensive insights and rapidly growing transport library at your fingertips. See the difference that a vast quality driven transcript library makes.

Unlock your free trial at tickets. S stock com slash Patrick. Hello, and welcome, everyone. I'm Patrick macy, and this is best like the best. This show is an open ended exploration of markets, ideas, stories and strategies that will help you Better invest both your time and your money. Invest like the best is part of the colosse family of podcasts, and you can access all our podcasts, including edit transcripts showing tes and other resources to keep learning .

at join glosses dot com patro SONY is the C E O of positive sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect to the opinion of positive sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of positive sum may maintain positions in the security is discussed in this podcast. To learn more, visit P, U, M, dvc.

I guess today are Andrew homan and Chris Miller. Andrew has spent two decades at maverick capital and as a managing partner at maverick alicant, where he leads the firms technology investments. Chris is a professor at toughs and the author of the new york times, the best in book chip war, which details the geopolitical battle to control the semiconductor industry. Today, we get into a comprehensive discussion on semiconductor ecosystem and a silicon backbone of our digital age under increased her insights on how venture capital is navigating this complex industry and what IT means for the future of computing, we discuss the AI driven revolution and chip demand, the geopolitics of semi manufacturing and the next wave of innovation behind invidia. Please enjoy my conversation with Andrew man.

and Chris is Miller.

So guys, I thought a fun place to begin would be to talk about one specific thing in the ecosystem, just as a jump off point. And intel seems like an interesting business to do that with because until been around for a long time, there's been periods when intel was an unbelievably dominant and pioneering business and were in the midst of a paradigm shift. So maybe you can describe why into the business is an interesting one to study, to learn about what's happening in the field of semiconductors today.

So I guess maybe to kick IT off I was at in event last week with the global some picture reliance, and this topic came up and someone asked, what film genre would you describe in terms of what is happening intel right now? Is IT a horse? what? And as romantic comedy.

And i'm not as good sure what turn out you to describe that right now. But someone from the company got on stage and his common was I would describe IT as a hang gg action film, which is probably a pretty good example of explaining the state of play right now there. And I think there's a lot of moving pieces.

You've got geopolitical tensions laid on top of all them. You've got competitive issues on the design business. They're trying to break in to the founder business out of the next new leg of the stalls for them.

But there are a lot of moving pieces. Guess is how I would kick IT off. But Chris, you've been found the company for a long time.

So I think they illustrate how hard IT is for a very simple company to pivot to a new business because intel's problems, in my view, are actually a function of their success. They were too successful in producing chips. Rep is too successful in data center for a long time, and that made them risk averse and hesitant to invest in new products. They missed the mobile revolution, famous ly. Steve jobs asked to blue chips to I do another, struck with a ee, because IT was always a huge rift to try something new.

Why does this happen? I know that the reason i've read the invitation, let me, like everybody else, I understand jc existing profit dreams. But also, this is so well documented across business history, especially in technology. These are all very smart people running in tell with lots of experience. Why don't they just say we're not going to fall pray to this specific kind of disruption and we're willing to do IT IT takes, including betting, lots of our free cash lower, whatever, to not lose this way again because we know this is the way companies like us lose.

I think it's painful for institutions and for individuals to actually embrace the individual illumina, because IT means you have to sacrifice the products, major career on. You have to promote people from different business streams who are working on new products that threaten the old ones. And so there's just a lot of tension that's involved is much easier to keep. The eb looks like it's profile and long time, longer proof.

Andy growth, one of the initial you I rode a book about this. And when you read the book like why is this not manator reading the entire industry, but certainly the entire management team because as you talked about, you miss this transition of mobile. Seems like they've largely missed the transition.

T. I. And so you've got one of the founders of the company that had highlighted these risks in these paradigm shifts. Now you need to be ahead of them. And for whatever reason, it's been chAllenged for them.

Can you describe that paradise shift that we're going through right now from your perspective?

So I think when you look at technology every decade, there seems to be what I would describe as an architectural change. And so you went from main frame in the eighties to the P. C.

Then you went from P. C. To mobile. Cloud also was happening simultaneously at this next paradise shift that we're just in the very early ends up would be A I in the implications from that. I think i've probably brought her perhaps in any of those other paradise fts that we've touch non so far.

Did you see any differently? You ve study the history this as much or more than anybody. Is that simple? Or is there more unions to appreciate?

I think that right. In terms of the end markets, I think sentinum to that there's been shifted in the business models along the way. So the history the industry is started with companies that weren't entirely integrated. They design and produce their own ships, and you shift to the fabulous sound model with tsp, c bruising ships for everyone. And now you're in a position where you're got a small number of AI companies of cloud company that are not just big buyers of chips. They're also reverting to the pride practice of designing their own ships and getting really deep into the weeds of the technicals of harder chips are produced because they want to have that entire integrated process, but they can control from the software onto the silicon.

Could you pick one company example that's making that transition away from being just hyper scale and going vertical and describe what's happening in a lot of detail, what this actually means in terms of outlaw investment, what they're actually building, what they're doing IT. One of these transitions would be .

interesting to highlights. You mean two directions is happening is happening at the chip. Companies are going up with a hyper skill vel.

They're all down. They bought startups. They've tried to integrate them.

They're all trying to understand how to bring the networking together, and they're having to hire hundreds and the thousands of new employees who have areas have never had to hire in before. So I really the same story. I think ma, at microsoft, amazon, there all sense. Ly, the same playbook .

maybe described the logic as, see.

I think you look at the enormous capital outlet that always companies are making right now, primarily farm video GPU, and they look at where that span is going to go in the future, which I think that number is only going in higher and all their alternatives that they can do, at least for certain work clouds, they can custom bill the chip that is very purpose bt bore one specific or k. Whether that be search, for example, some sort of social media APP that allow them to potentially lower their costs to serve those customers significantly.

It's faster to watch investors try to figure this out, and you can find a brilliant investor for just about any argument for where the value is going to accu in all this. I'm assuming here that A I is going to be a massive, huge thing on par with for succeeding a mobile or cloud or whatever. Obviously, there's some version the world probably where that's not true, I don't know.

Let's focus on the world in which IT is true. As with a very well known investor in the private markets last week, that said, look, we just think it's a hundred present going to be in the application layer. And so that's we're making all of our bets.

You're focus much more on the infrastructure, silicon on player and everything around the chip. Why are you making that bet? What is IT about? You're reasoning through where the profits will crew in the future because of this boom that has you focusing so specifically on silicon and the surrounding technologies.

So when I think about the broader A I ecosystem, we think about IT as a three layer cake, right? So the bottom layer chips in cloud, that's where we focus. The middle layer would be the foundational models, and then the top player would be the applications built on top of us.

So the function models would be OpenAI, anthropic, gami eea. That ship cloud layer would be in video M D T S M C, eea. And then that top player would be ChatGPT office copilot test full soft driving.

And I think the chAllenge right now is that on the top two layers, it's very foggy in terms of who the ultimate winners are going to be and how much value is going to get a crude to those layers broadly eventually. Yet there will be big dollars ago. There are been in terms of who actually captures, though I think it's very OPEC right now. When I think about that chip player, though, to me, almost in any scenario I see that be in the area that's onna capture a significant amount of economic rent through this paradise shift they were going through right now.

If A I, what would be the reason that IT doesn't, let's say that we get to simulate the future and a whole bunch of different ways and in some percent, those futures the infrastructural does not capture. A lot of the profits gets big, but there's pricing pressure and IT gets competed down to the cost of capital and there's just not a great competitive vantage period. Turn money in. Draw that picture or can you draw that picture? How could that happen?

I think there were two dynamics. So one would be to the scaling laws change right now. Everything we see would say, no, they're not onna change in the more compute you throw these problems, the Better the model, the quicker you'll be able to train the model.

I think the question is, do you get into situation where there is much more competition particle that A I accelerate specific area of the market work, Prices come down anything? I think ultimately, we do one Prices come down. That's the history of compute.

But I think you've seen in prior transition, whether be mobile or whether was with PC, that silk way are so capture a ton of value with PC and server into clearly created an incredible business during those periods of time. And then specifically with mobile, you look at companies like ARM and qualcomm, others that monotoned significantly on that trend despite there was more competition. Prices did come down, but the time was so large. These still great companies were building, created over those time periods.

And I think the p mobile a both show that describing is just a silicon actually mistakes the source of those modes because there was silicon plus often are in the interface between the two. Why ships and half the world's pcs? Because of software building top of? And does architecture the same?

True for where's apple in mobbed com or ARM at the base at all. And so you ve actually had a long history of the silicon surface playing big role. And the software is often a big portion of the mode .

in one of the the area that words mentioned on the scaling laws with the release, open a story model, an area of AI that we're thought to be compute light, that the inference is now compute intense as well, is for that model to be Better at reasoning. IT requires much more in a much longer period of inference.

Is really interesting to think about the history just from an investing perspective up until now. And I love you to reflect on IT, where if you ask about semi investors in private markets that off and I just don't do IT, it's capital intensive and brutally competitive. and.

You just not going to make money as a semi conduction investor, feel like a venture investor. And in public markets, there are certainly people that have done tons of semis investing historically, but there aren't a lot of semis funds, semis dedicated funds. There's planning of software dedicated funds, but not a lot of semidetached funds. So why is the technology that has undergraded all of technology progress of the last forty, fifty years? Why is IT been like somewhat of a backwater for public and private investors over save the last twenty years?

I think there's a dynamic where there were a lot of semi focus funds twenty, thirty years ago and many of them did very well. And then you had this rapid rise of software that kind of captured everyone's imagination. And many of the kind of quickly ly the early stage community pivoted to software and did fabulosity well, right, investing behind that trend.

And then you had a situation where a lot of that expertise just retired and less the industry and it's broadly IT is semis are chAllenging. They're technical complex. They are capital intensive though I think like comparing them to software today, looking prospectively over the next ten years. Where to say the last ten years, my guess is that we're going to see more and more capital start to flow into this hardware layer. So much value, I think, created there, and there's just so much need for innovation.

Where do you think the highest perspective returns are within infrastructural layer, even obviously focus on and video, that's for all the market cap is they GPU everyone talking about GPU. There's so much more going on around them. Data built out famous and record pace with crazy monts of power is all this cool stuff happening around the GPU itself. So where else are you most interested to study? Look me with company and so on.

This ties into a point that Crystal ate earlier, less focused on that core GPU itself. There's many companies trying to go directly after that AI accelerator market. I think IT is chAllenging to go after the predator that is at the very top of the food chain right now in video being that player.

And there's all these interesting areas around that gp, whether it's interconnect or memory or power related where you're just onna need innovation because the way of doing business that existed five years ago is just breaking. And the data center that existed five years ago where an intel processors king of that data center and there are some simple networking and there are some storage attached to IT. The example that we get more talking to folks is that essentially you had a tsunami, I N A I tuna I, that specially wiped that data center out.

And jensen, riding his surf pod on top of that sunni. And as a result, you're seeing everything needed to get rebuild because the paranoid of these GPU is are off the charge, are trying to connect as many of them together as possible to create super G P. S.

And so now we're connecting one hundred thousand and get together. The next step is gonna to get to a million them together. And the complexity of doing that is creating, I think, a lot of opportunity consume.

Curious where you've started the history of this in detail. We won't go through IT. All people should read your book that they understand, which is excEllent. But i'm curious if you take all that history up through to today, when you spend a lot of time with companies, with governments, with investors, with through everyone in the ecosystem, what people are not talking about enough that is not yet enough in the mainstream narrative. But you're hearing a lot about in these conversations.

I think everyone is ginning to register how much money will be spent on compute over the coming decade, driven by dis build, driven by AI. The numbers are staggering. They are exciting if you're a data builder or a power provider or your NVIDIA, but that's actually a bad thing.

We would like chea computer to be as cheap as possible. And so I think the key focus right now should be finding ways to minimize the amount of money that needs to be spent on to build out the computing we need for ai. And we're right now at the stage of admiring the scale of the problem, admiring the billions of dollars that we will be spent. And actually, the real goal should be finding ways to economy to bring your technology, to find ways to bring down the cost because that's going to be one of the key gating factors of A I development.

Is can we bring up about the political considerations around both the infrastructure and the first two, stax will leave the apps for later, maybe, but we're talking about the judy infrastructure, the fabs, everything that needs to go to run these things, and then the model layers. U. S.

Companies seem to really dominate this. Maybe except T, S, M. C. In a couple in europe, talk about the most important political considerations, given that if this technology, I, E S, the most important, whenever, like every county, going to want some answer to their foundation model, their data center, their infrastructure, what are the key variables in the political landscape right now as you see them?

I think if you start at the level of word ships produced, if you've got a lot of concentration, taiwan, as you mentioned, and the economist of scale, the industry are such that you're just not going to have multiple locations at the size of taiwan or chips produce maybe a couple of the countries career, but it's not going to be every contributions on chips, totally impossible.

The data center of level, if you talking my data, enters for foundation models. Those are huge, brutally expensive projects. And so there are two, although lot of government that would like to have their own the scale of investment about that unless you're sitting on a lot of oil or unless you're a very large economy, you really can't justify this span.

And so that's why if you look at where big data and products are happening in the us, happening in china, there's a desire to have meet east and there's political issues that come up right there because the U. S. Wants most of the role gp s be used in the united states and is wary of letting too many get deployed in other countries, both because there's concerns about china and competition with china, but also because the us thinks will be really benefit to training most the world's foundation models.

If we had like a country leader board of who is the most advanced in a couple different areas that say foundation models and infrastructure, who do you think we'll be the largest movers up and down over the next select .

five years and the surprise player right now is the U A E. The to put a tone of money into A I. Their leadership is very include analyst issue.

They spend our time studying IT. And if you look at the filter models of they have develop, they're pretty sophistic ted. And I think most people won't expected five years ago that the U.

A. Would be from center in AI. I'd actually the surprising downside players is china. If you look at the huge capital expansion uses that U. S.

Big tech firms are along right now, that's not really matched in maybe the exception of bite dance. And that's partly for U. S. China reasons and controls and technology, but also for china's own domestic economic situation, is not making film excited about putting out billions or tens of billion of dollars and data center span and investment.

They I and can you talk about the motes in this business that people might have a hard time appreciating T S. M. season? I just want to talk about because everyone knows what a political danger hot that this could be.

Most of the worlds d's cutting as chips are built very far from us and very close to arrival. And IT seems of unsaying that we just don't replicate this in the U. S. I know, were trying. But the fact that we haven't begs the question, why is IT so hard? What is IT about the infrastructure businesses that makes them so hard to replicate or copy IT just seems crazy.

It's changed over time, right? Because there was a period of time where intellectually was at the leading egg and was the most technology advanced manufacturer chips in the world. And then T S M, C.

Lab frog. Then that rally happened around ten when intel was struggling with that node, so called that cashed. I was rally and was seven, eight years ago, something in this of code.

And so like, why is T S M C been so successful? What they do, I think in part of it's because they're just manually focused on just one business model, right? All they are doing is the foundry model.

Where is the other players? Intel has a very large design business for designing the chips that will get made. Intel founder.

Samsung plays in multiple areas. Memory founder. They also designed some their own chips. But if you're just singularly focused right on the basket hand in making the highest performance chips right and then learning from your customers, they have all the most sophisticated design customers at this point, making their chips on the leading edge. Tmc IT just gives them the ability to continue to improve and sharply in the manufacturing process to a level that I think at this points, we be very changing for anyone to catch up.

Thing that strikes me is the ecosystem in taiwan is just extraordinary, early, well developed. And we talk about soft requiem, silicon valley, but there's a many of traction assist taiwan's is the chemical suppliers, the materials producers, the people who know how to fix the tool s in the tool breaks.

And they're all right there in a very small country within an hour and a half of each other on the high speed rail on the western shore of taiwan. And so if something goes wrong in your factory, need to prototype of new machine or new process is very easy to get IT right away in a way that if you are in arizona or if you're in japan, uh, you just don't have that efficiency. And so tim, see, we will say, I think they are right that the bullet train that goes up and down taiwan, bringing people back and forth in a single day between their babbs, is one of the key source.

Come to the why is the yield solo in these things? If you're producing cars in a ford factor, something is very few lemons, whether you tell me what the numbers are. But I think for some chips, this as low as like seventy percent of the ones produced, actually, then go get used in production, which means you just like literally throwing away a lot of the product. Why is that still the case? Despite all that sophisticated.

the manufacturing is the most complex that humans have ever done. So in comparison to a car, cars can have fault tolerances measured in millimeters for certain parts. In ships, you're talking about nanometers, so a thousand times more precise.

And if you look at just the tip in your phone, for example, and pop open, there are billions of tiny transistors, each one of which is the size of a cron of virus, and almost all of which have to work for your trip to function, to let you like things, instagram or send a text message. And IT changes every year. Every year, T.

S, M, C rolls out a new manufacturing process that is dramatically Better. The old one, which means that the transistors are smaller or packed more densely together, and the tolerances get even tighter year after year. And so the question of yield, ld, how do you produce ships that work? Means got to be yielding a new process on a regular basis. So you start lashes process, and you start all over again with the next process, with even smaller tolerances for air.

And course, what you think like the five fish most important companies are right now outside of event tisa.

I would put back on that list. I would put S K hino x on that list on the memory front. I would put M D on that list right now. I think when you look further down, you would put some of these other smaller component players that are in the A I infrastructure orbit as well. There's actually play a longer list of those more than two or three, but you definitely have a whole group of companies that are benefiting from what I think is gonna be one of the most significant infrastructure builds that we've seen.

If you think about a community S M or something that seems pretty singular, which companies, if they just broke completely with most mess, but a emails .

definitely on the last more, that's a time to chat about U, V, all these different accounts that people throw around that, I mean, only get a seventy percent year older, eight percent year older, maybe even lower than that in some cases. But we are talking about at launch how credibly complex the E. V.

Processes and what exactly you're doing. And we're going to use this laser to heat this tin two hundred times the temperature of the sun. And then we're going to bounced off these mirrors that are the most technologically advanced and smooth mayors in the entire world.

And then we're going to essentially, the layer of complexity of the analogy is i'm going to hit a golf ball from earth and get a home one on the moon. That's what you're doing, is what you're doing with that scope. I think that and gives you a sense for why sometimes these chips don't come off the line and perfect ah .

what do you think of the top concerns of the U. S. Government about chips act? We can talk about where we will accept external capital to finance companies or projects just by revealed preference.

What is the U. S. Government care the most about in your actions?

How on the ball are they like? We're talking the day after David, some veto this bill in california, which I think technologies is very happy about. How well calibrated are our politicians around this stuff? Do you think they've been effective?

What are they doing? There are three broad goals and has one is to mitigate chip making concentration on taiwan. Number two is to keep U.

S. Firms out in front technologically in every statement that's possible. And three is to prevent adversary china of all from accessing cutting A G I chips. And I think we've been a lot of moves the last couple years on each of these fronts.

The chips act, which subsidizes must manufacturing controls on technology transferred to china government, is relatively including, I think, the people who are working these issues understand them well. But I think government works at government speed. The chip industry works that mores law speed.

And so what really matters in the long run is actually that silicon valley, the u ship. Industry, can keep that expands worth going. If IT does, then the political issues sort themselves out. And if IT doesn't, we have more problems in the political issues because our entire autonomy is levered to more law.

Why can you trying to catch up? If so much money, so many incredibly smart people, so much talent there, obviously the will seems like more and more of a centrally plan, more centrally focused effort. They can just say we're going to vote x amount of effort. We don't direct the private sector the way they might understand why they can get parity earlier. You said they might be on .

the downside less. No one can do the other own right now. You can produce clean ships on japan.

can't. Taiwan certainly can. They rely on imports of chemicals and soft technology. And the fact that china struggling through they're Normal for, are strugling to do their own.

In fact, they're trying something that harder than anyone else has already tried, which is to actually be self sufficient. The trend of the last thirty years has been to become less self sufficient. We've collectively decided that we'd rather have a specialization of taiwan specializing in manufacturing, japan specializing the chemicals.

Us specializes in design because there's cute difficile encies that come from that. And did we expect a bit too much specialization? In the case of iwan, you one can argue that, but I think china's effort to do IT all by themselves is much, much harder because they don't get the benefit of access world markets and access the whole world's expertise.

Because this is such a burgeoning space. I'm curiously where, if I ve got a room for all the semi investors that you respect, however those people are, where you think your personal views would most diverge from that group.

Couple of all, I think there's been a view, you know maybe take the biggest of them all, right, in video. And what is the view on that company? And how has that change in the broader public markets over time? It's interesting a lot of active managers that I ve spoken with actually in video has not been a huge position for them, which is interesting because I think people always felt this is too or the stories already been discovered.

And so that's probably been one of the most, I don't know, a puzzle energy interesting options that I surprisingly, like few people actually own in deep feeling, like very well known story, right? And I think people talk about all the time, right in the newspaper, all the time acta. But I think people have this fear that all my god is the stock is too expensive.

It's around, it's ready up so much. But if you actually look at what's happened there, at least on a fundamental vases over the last two years since chat ch. GPT really was released, the fundamentals have tracked dead in line with the stock Price, right? And so if anything, I think the burner's multiple is probably flat to slide down versus where IT was when chat B.

T. Was released. And so theyve actually delivered on a fundamental basis like over that time period. I think many smart some investors probably do on the video. I don't want to say that necessarily the case. I think broader in terms of the broader market view of that company is always struck me a little bit interesting .

given that dynamic. You talk about the changing cost structures and business models of the digital slash software world in this new era, where he used to be very people engineering intensive, becoming much more compute intensive. I know we talked about we want that cost to come down, but just a game on the field right now, the difference in the business .

mills that are sing this site struck me probably only twenty three. I was tagged along a meeting with the ventures team, meeting with one of the front chair model players. And this was early in their life cycle, and we were getting their financial model for what the business could look like over the next three years, five years, what the cold cure look like.

And this was one of these I opening moments where you looked at the cos space and literally ninety percent of IT was compute, right? So these are dollars that we're spending with asia A W S, or G, C, P or oracle cloud. And literally just for training these models, and that was another one of these ahm people don't appreciate the magnitude of spending is gonna happening here over the next decade.

The stats that people feel around train a competitive model in twenty twenty two was ten million dollars in twenty twenty three, it's one hundred million dollars. In twenty twenty four, it's a billion dollars, in twenty twenty five is ten billion. And then I think the latest the lyell was highlighting was hundred billion dollars in twenty twenty six.

Like stagnant amounts, what types of companies is actually keep up with level of span to be able to be on that frontier model? Edge them lead? That's one area. Just you're seen just the cost of doing business in software, gn particular just drastically change where you have these old models that were very people intensive and you've had all kinds of sales people running around on the product, very R N D heavy programmer, actually a and IT. Is that now changing where just the cost of business is being much more compute and ship driven versus people driven? And think you're seen that shift in high per scale cabet's numbers.

Acta key outline, what you view is the key sectors within sems, whether that's design, whether that's manufacturing, whether that whatever, what do you think over the key three, four, five areas that matter most?

So the design piece is critical, right? And that's where the companies like NVIDIA dominate, right? They're designing the best chips and then tmc is making those chips. And then I think what is interesting is that you're seen as more as lawd slows. And this is people are well aware there's been pressed talking about this for now for probably the last ten years.

Where are the opportunities to create value? Is IT gets harder and harder to get those more law benefits on the front end of the initial manufacture of that chip itself. And that's where you're seen so much interesting innovation on the back inside.

So the events packaging inside of the equation, for example. So alright, we made this chip. Now tmc is on three nanometer.

How can we package that ship with other chips and memory, for example, and get that packaging to be as efficient as possible? Because basically people just having care about that because you are getting these massive cost stone with more law every two years, double other transistors in that chip itself. And so you're finding these areas at historical have been ignored that are now made.

Some cases of like coasts is probably turned that many people heard thrown around, actually limit in the number of NVIDIA gp s are able to get manufactured. And so finding those areas, the historical Prices have been ignored. And actually some of the most interesting.

if you think about the emerging players, everyone heard some of these big names mentioned a lot, S, M, L and video kings and so on. What are the most interesting emerging players in your mind? That will matter a lot more. Like my rankings question, names that will hear a lot more than five years, and we're hearing about today.

Even the first thing you have to say is that you would be surprised by how short the list is relative to any other sector technology. Look at A I, at least the five most important companies. And A I, half them were found in the last decade, do the same some icon doctors and it's a very different ratio. I don't think that speaks to the fact that there are some big g combine who play a very big role and is LED the less company information. And I think you might expect, given that the history, the industry is actually start up merging.

And so I do you think about returns, I know you and I ve talked about this in the past is really just this is like a massive cycle that could be bigger than the ones that have come before IT. And obviously, technology investing about writing those waves. And I how do you think about what that means for perspective returns relative to everyone's opportunity costs, the five hundred and or something like that .

at the big player? It's interesting when we look at opportunities on the private side, this is where at least today, you've seen valuations will also be quite much different than you see often are, particular the last few years. I think the opportunities for returns are actually going to be quite attract of interestingly, looking at the performance of the semiconductor index probably is actually done quite well.

I think something like twenty five percent k over the last decade. And so I think when we think about the opportunity set in front of us right now, you look at the private world of samon, then you look at the public world in theory, that private world should not perform the public companies, given that there should be an a liquidity premium. Exeter and what we see in the market now is actually a pretty attractive set of companies that should be either large dependent players in this space looking out three to five years or will become important parts of many of the large existing comments, right? Because as you see, IT very chAllenging for these public companies to merge with each other. They're gonna need to do these token M A strategies to really beef up their internal like ship efforts .

to about anti trust. This seems like a really important the future of the system when IT comes to returns. If the video wanted to buy all the companies that you backed and they're not able to give anti trust that the impact your returns, maybe what's the data play with anti trust and gonna allow.

We've heard love and just last week with discussion of a potential l cocom. Deal with that would face. And I trust a regulatory issues hard for me to know. There is actually two types, sani trust that matter in the chip. One is real anti trust from western regulators and to his fake anti trust review from chinese regulators.

And a lot of the deals that we've seen be sunken recent years were from the chinese refusing of approval, not because there is actually real cern, but this was a part of the political game that they're playing with the united states. And so I think for every major transaction and even some minor transaction in a chip industry, the question of chinese approval hangs over and is something that companies themselves can do much about. IT depends on the state of U. S. China relations and once the presidential summit, that what matters more of natural substance of the anti trust.

Can you give an example of that? The case study of that playing out the most recent.

I think, was with intel's effort by towers and micon ductor. So tower si conductor is a very small foundry company producing different types of chips and entire produced. They're both in the foundry business but through a small company. And western regulators will wait through and then china refused to.

And why can't we just say, okay, why isn't there somewhere around that?

I think this is a really interesting question. And at some point, we're going to merger that approved by everyone. But china and the company, just since a scrut with the promise, is a pretty big economy and everyone is got assets in china and everyone needs the chinese market. And so unless you are in a niche where you don't care about chinese customers and you have any assets in china, you don't know exposure to the chinese market, you've got to listen when .

they say no there is this mean going around earlier this week that in the late nineties, we religion have an internet bub. We relied on telecom bubble. Why might this not be of a similar character or nature to that?

I've ve got a lot of thoughts on this one, the cabeca to be rages on. And I think that Gavin Baker did a nice job setting the state of play on this about a month ago. So let's go back to that bubble, right? We spend a lot time thinking about IT.

I started in maverick into for right is IT was in the aftermath, right of that explosion. And we pulled the data and looked at the biggest offenders and biggest spenders of cab cks during that time periods. So the global crossing, the level three is the quest.

Ca, and you look at the amount of capital that they were spending was just totally unsustainable versus where we are today and all get into that. But the capex, so that are are spending with the r or of optical gear was on average about two hundred percent of their Operating cash low. So for every dollar that the business generated on Operating bases, two dollars spent.

S, they weren't self funding and they had to borrow in the high yield markets or issue equity or heavy. But IT was not a self funding sustainable situation in some years. We pull the day and there was even more out pop in b five hundred percent of Operating cash in a given year.

And then you look at what was the utilization rate of that fiber in the ground. And I think as of you like is the nineteen ninety nine and early two thousand and only like two percent of IT was really up being used. So was all on this build IT.

And they will come in because the internet was the early days back then. And so the thought was this is organize consumed in very shorter. And I just turned out that IT wasn't. And I think it's very different from today where you look at the spending that is primarily done by the high per scale or so metta, alphabet, amazon and microsoft. What are they relative to history on capex s.

And if you look at their Operating cashle on this case of cash, the business generating right now, their capeci is only about fifty percent of that, which is dead in line with it's been over the last five years. We are very much in a sustainable level where you could double cap backs and these businesses would still be self fun and you wouldn't have to have run into you where they're going to borrow or issue curate or something like that to fund these builds. Furthermore, when you look at the utilization rates of these GPU in their clouds, IT is nearly hard to there's nothing sitting around I L.

Under utilized. The dynamic, I think, is very different. Looking at the telecoms od of time where you have this mass amount ever sure that, that was built IT wasn't really built in a sustainable manner from like a financing perspective.

IT wasn't even being realized. Where is today? We can debate what ultimately will get generated by these large language models in the value that you will get created. But I can tell you the gp s are definitely highly right now, so feels like a very different.

different time. So hard to take your analysis and applies to today. The place that looks like toko is the foundation model companies. They are spending five x there way more. The Operating cashes, negative Operating cash list that does exist.

So does that mean there all toast? If we get to a hundred billion dollar model, there's just not many places in the world and get one hundred billion dollars and the very few places you can or the companies that you ve mentioned that have these amazing Operating casuals. So that just mean that this technology is like an incumbent stream and that these upstarts, even the open eyes of the world, are in trouble.

That's interesting way of cutting. And I think you look at the time companies that could sustain that levels span that list is pretty short, and we talked about them just a few minutes ago. So that's a very interesting way of framing. And I guess the implications in terms of who ultimately can afford these, the list is not very long ties .

back into the midnight east, which we discuss before because we're looking for a very large positive capital.

Will that be is up? Well, the U. S. We have in various cases, pretty close ties to the countries of the most money. Talk about the range of visibility.

I think the U. S. Is pretty comfortable with middle east n governments investing in U.

S. Space companies to build the U. S. Space data centers.

And we've seen, for example, microsoft mobile ler. I announced a joint investment vehicle. I think U.

S. Is a lot more worried about investment in the middle because then the us. Has less leverage.

The midwest has more certainty to will be data centers built all over the world, including the middle ase. But the U. S. Government really doesn't want the most advanced data centers being built in the goal rather than in the us. And so there's a lot of regulation coming on the pipe right now about that particular issue .

we must worried about.

I worry about the time, eyes on which we see the return that Andrews talked about. I think making cure that lies up is going to be critical to sustaining if the scaling laws are true. The capex scales up alongside the size, and we need real returns right away that cap backs.

I think next couple of years, we have a runway in which the herschel will spend regardless of return. And then at some point, we're going to need to see real business models emerging. And of the time arizon for that, I think really matters.

And the other thing going back to the cost compute is the more we can do to find ways to economize and compute, the more sustainable all this becomes. There's a huge center right now economize on computer because in video shown how large the market is to take a bite out of. And the best way to take a bite out of IT is to produce something that is just as performance or even more performance, but a lower Price.

And how about from an investing perspective, were you the most you obviously going to to play out a lot of capital to support companies trying to solve these problems. What IT keeps you up?

IT goes back to some of the points that Christians made around what is the sustainability of all this and is the situation where is our digestion period. I've been doing some mates for long enough to know that this is still a social business despite what people will say. I think it's absolute growth of business.

And so I actually pride. Any sort of downcycled because it's a great time to put capital work. But I think just making sure that there will be a tear of new companies that are form that are able to capture this broader pair time shift they were going through.

And the chAllenge that the world I think faces today, someone is you have one company in incredible company. It's capture the line share of the economics today. And when I think about in video and what they've done, this was company of initially a chip company, right? That then became ship and memory company that then expanded onto network in another sign, this full system.

And what happens if there's one company is just capturing all the economic value of a transition like this and you're going na want to have, I think, a very viBrant start up community that's helping solve some those problems prickly like in and around the GPU. I think going directly after the GPU itself, we talk about that more or two I think is a chAllenging task. But there is no shortage of need for innovation, everything around that GPU.

And so are we going to be able to create ecosystem here, opportunity that exists around all those things that we mentioned that are breaking in earlier, around power, around in a connect, around memory exit. And I think when you look at the big customers that we just talked about, their highly motivated, diverse and healthy silicon supplier base, and we were chatting about, let's look at the last big paradise shift and what happened with mobile. And you had a situation where when apple announced the iphone 1 t is the launch partner， 8t had a market cap of two hundred and fifty billion dollars。

When the iphone came out, apple at a seventy billion or a market cap, you fast ford fourteen years. And apple's market cap now is three and a trillion and eighty, and t is one hundred and fifty billion. And apple is extracted all the economics out of that transition, capturing, I see the hardware with the iphone and also the software with the APP store.

And I think when you think about the biggest fires of a lot of the infrastructure right now, they want to make sure that they don't become what everyone describe a dump pe, right? And they want to be able to be capture value from this A I transition they were going through right now and not the metal agreed to one of their support. And think he creates attention in the ecosystem. When you have a supplier that has a bigger market cap than you, that has more power than you excited, I think there's a real incentive to have a healthy ecosystem generally speaking.

What are the corporate strategy options available to those players that want to accomplish that? Just looking for the apple T N T thing. How do you avoid that happening to you this time?

They are going to help their own chips. And so that customers to work is probably very high. The list of what they're trying to do, and I think they're very forward lining. We are having lunch with one of the hiper scales last week where they have a team.

Their whole role of this team is just to be talking to startups in silicon valley, understanding what technologies are coming down the pike so that they can be put into their data center and to see if they can talk, solve some of those pain points that I already. So I think they're trying to have a very wide appeal in terms of what they're looking at, in terms of what other opportunities are out there, even for the incomes right there. I think a strong desire to see AMD succeed in this market as well.

I think we will see more, more success come from them as they catch up and were releasing new chips that are higher performance. They have figure out some the software chAllenges that y've had historically. I think that's another piece of the equation.

So it's gonna be accommodation of custom ships that these hypotheticals are developing themselves. It'll be looking at other in comments like the established in companies. And I think the third buckets going to be the start ecosystem, providing value to help them diverse.

I know he said maybe less interesting to go directly after the GPU and Better to attack the area surrounding, but let's say the GPU is the desta at invidia and there's that one medal era for luck, sky ockrent to shoot the laser being through to destroy IT. What is your best guesses to where that vulnerability might be for the media?

I think the chAllenge will be that in videos, ship is very flexible and limber. That's why people like IT. But if the models change, the G, P, U, can still save the models of tomorrow.

If there is a scenario, this transformer technology, for example, ends up being the structure that everyone is going to be building models on going forward, that will potentially be that opportunity for companies, for storms able just isolated that one very specific use case and do a much more cheaper. That will be an area, I think that could in approving to be a chAllenge for them like that s in video, respond to that. They can, obviously custom chips that kind of go out. These are specific markets to today. They chose not to.

okay. I wanted talk about the ways to stay informed in this very fast changing world. And i'd love for each you to call the mine three specific people. If you only got to ever talk to three people with the goal of always knowing what's going on, just imagine you could access. We want what three people would you pick?

Think IT would definitely be someone building infrastructure at one of the large hybrid illah leading edge of all of this, you can say even one in the frontier model folks. I think that's definitely at the top of the list. I think it's going to be someone T S, M, C.

That's on the leading edge of the manufacturing front. I think it's definitely someone in the video that's on the front and be someone video, right? You can say who that would be.

Those are obviously three good answer. I think the other asked this that I look at that from is understand the business motives. And that, to me, is something that you understand, not by talking at someone who's building the data and of hyper scale, but actually who's running one of a hyper scales because they're looking at this saying in the video, a is not bigger than we are.

What's approach to the fact that our suppliers now beating us and scale? And what of the just take to deal with? I think that one.

I think the second is the capital allocation. Where's money flowing from into? And I think you learn a lot talking to some smart investors in this base. And I think the third would be actually politics matters more here than IT has in half a century. And see, you really can't understand what's coming in terms of who is was access to which market terms of which technology .

can you celebrate without sid? Questions would ask six people in this room. What would be the couple most pressing questions that you don't already know the answers to that you wish you could have an answered to.

I think on hyper scale built out what are the key problems that they want to be solving next year. The year after that will provide you a road map into a or they be buying, but also be how they position in themselves in the industry. And a hugely important question. That's what all of the money is allocated in the hyper scale er's decisions as half the market right there. You talking ai.

but I think of the government piece that Chris barros solely important here to around just understanding what is the folks quickly of the U. S. Government and where they are going to be putting potentially restrictions in place or be more important, are they trying to enable right, with the the starter eco system in the U. S. Operated in a much more higher level? L, I think that's another key question in topic to explore.

What about beyond the hyper scare half market? What about the other half? What's going on there?

I the market that hasn't really taken off yet, but will is on the edge of networks. We're going to see a whole lot of innovation coming as we need to deploy some models in the data centers, some on the edge. And we're not sure what ratio yet, I think, were the very, very early innings of innovation in some is on the edge to be super power efficient.

I just thought on a processing power for edge devices, that will be different for cars. I think that, that will be for phones or for whereabouts. Others, unlike that is, were inning three or four right now for dis center, were still in inning one on the edge.

What does that mean? So does that just mean tesla and apple are going to develop Better chips? Doesn't mean the end user devices are pretty concentrated and is the biggest companies in the world.

They are already designed chips because just them doing a Better job. What they are already doing? Or is there some new paradise shift there too?

Is so in certain cases, like apple, I see they will most likely be making their own chips for their I think when you look out a little bit more broadly, there's different thing to me. The opportunity for new companies to take advantage of how this ribe, where in video has clearly shown that IT dominates the data center and think cans down and think most people would not debate that whether I think on the edge, the story is very much on written at this point, and a much of the game is really even begun.

And if I look at in video and where their focuses, I think jenson's point all are the vast major the resources to focus on data center piece of business because as all the values is getting created right now, whether I feel they perhaps even decrease focus on some of edge products. And you can see this even on their financial statements where their data business has grown at a rate that I don't think have everything in any company grow at that scale over the last eight quarters. Where is on their business? For example, that business has largely been stable over the past, a quality, whether you seen the losing in the days and important that is, is, is not in their DNA as much to be offering blow power chips, which is not really critical for the edge when making these things are going to be battery Operated and performance for wise.

The key metric, the joke in the industry, is that video makes a killer gaming in laptop. But if i'm getting up my free from F, K, back to four ago, and I plugged that laptop in the plane, the whole plane, to infect because the power usage is so high, is there an opportunity on the edge to create, I think, some transformative companies that can take advantage, I think what will be intelligence just moving from that day is center into your phone, into your car, into your PC, into. The broader industrial footprint that largely didn't see a lot of innovation and probably four decades.

And so that's something in area that's really exciting and worth expLoring. I think the other point just in terms of who's buying these things, that this whole rise of the start of GPU cloud, right, companies like corry, for example, that came really out of nowhere to build these dominant franchise. Right now.

There's one in the AMD ecosystem called tender is trying to do a similar strategy. And I think there's a real need for these start up clouds because they're just they can move more quickly. And it's interview that we view amazon, a microsoft, these incredible businesses, which they are, but they're also at a skill they just don't move that quickly.

So there is the benefit of partner in with the court then invidia like an engineered job really bringing them up to scale where they can get chips out six months before. I think if you look at coral, they had their eight one hundred ds out in march of twenty twenty three. And you couldn't spend up an instance of each one hundred in azure or A W S until probably almost like six months later. And that's a lifetime, right? Given how quickly things are moving right now, the broader A I ecosystem. And so if you're able to be a nimble start up that can move quickly in a line with a chip company, this case in video tense AMD, there's an opportunity there to create a lot of value both for the chip company and for that start up cloud itself because you just get this product out into the market and these GPU clusters are very finicky, right? And so if you just specialized in making those work, there's a big market.

the edge. One more point there is that I think since we're in the early stages, we underestimate demand of change and products that will emerge. Easy to assume it's going to be air pods and iphones that are defining the edge in ten years, but if you looked in two thousand, what would be the primary use? More building? You might not guess the smart phone.

And so I think we should be pretty open minded about where will glasses, beer, whatever else. It's interesting to watch facebook for two years ago. The metaphor seemed like a idea that wasn't realized very soon.

Now you won't all the metaverse, but actually, there is interesting glasses applications coming out on the consumer side. And I think that the thing is the industrial side, which is less exciting because it's nothing were going to buy for own use. But as Andrew said, there's been so much a less innovation on the industrial side there potentially use cases for AI on the industrial edge, we're talking over launched about tractors having GPU and them. I think that's probably just the first of many use cases for industry, which could be a very large mark.

meaning a joan dear, that has a couple of invidia G, P S. On IT. Anything that could benefit from the deployment of a cutting edge A I model will actually have these cutting edge ships, cheap chips, like the most things on devices.

And any sort of robotics in the future is gonna a holiday computing power. And maybe some of that will be calling back to the cloud if lencs matters, like in a self driving car, but also in a throne you don't want to call back to the cloud to ask the question on what you want to have compute on the device.

What about power? Any commentary on power? There's all these crazy headlines about build more nukes and all the things that we're going to do to power the data center is in any layer deeper inside song what's going on the world of power.

So I maybe to be a little bit IT of a controversial view. My thought on the so paris critical and this probably not enough of IT brother, but I think there is a dynamic where capitalism usually finds a way to work. And I remember what if you go back, took solar, for example, in the two thousand and six, two thousand and seven time frame there a fear capacity comes online.

And then there is actually oversupply in that case, in that with power in particular, when I try to folks that I would view to be experts in this space, I ouldn't say experts empower by any means, but I think there is a nna ic when you see these forecasts for data centers are getting built. Is that a dynamic where a data center says that it's going to need pick number and make city mega? And will the real use is actually be a less than what they're saying going to name plate capacity will be just because there's almost like a double orden dynamic.

They're telling the planner in that city that state how much power they want. They are going to give the absolute kind of higher number that they can that in the case, they might need that much, they can have access to that much electricity. I think the other dynamic, and this is an area were talking about the edge, the broader energy infrastructure in the U.

S. I think people don't actually have that a Crystal clear sense of how much capacity actually is out there and what the utilization rate actually is. And there's this concept of you're building to handle that peak load.

But are there creative ways where in atlanta, for example, in the heat of summer when everyone has their c on in August, can you somehow in alive folks to turn very conditioning til warmer temperature and bring that peak level down? And then all sudden, maybe there is actually more access capacity than people. So the production of me, because the mindset brolly and the investment community is that power. And I think power is very critical. But I think there is a dynamic where maybe some of these kind of shortfalls find solutions.

I think the other dams in data center chips, power has not really been a key focus the past decade. It's mattered. But IT hasn't been the case that the primary driver has been bringing on power consumption. And as a result, we have haven't brought a power consumption.

Where is in mobile? Most people twenty years ago be shocked by the computer per watt that is in your iphone because apple and others have fixed IT on bringing that down, bringing about the pupa what up, because by ter lam is limited. And I think now and video in a lot of calls with customers saying, hey, we'd like more performance, but also you get to take power less seriously.

I'm current how things get done in this world. This is like a side door into asking about the way that you're structuring mavering silicon, specifically the players you have around the table, the investments of your time and energy that you're making that will ultimately be called back to benefit the companies that you invest in. IT strikes me as a world that we could count on two hands and pretty quickly get to players that really have the power and whose influence and support you might need to be successful as a break out company.

Talk through that dynamic of how things are getting done, who matters, how you're cultivating relationships with them seems to have much more hands on style of investing than maybe even mather is done in the past or it's like mostly a public markets investment firm to do anything you can Better IT from going up IT. Seems like you're going away the other direction with a really hands on approach. So how and why are you doing at that?

Yes, it's funny. We spend more time in this and pulled on this thread. I think we've got more, more excited about the opportunity here.

And when you talk to many of the founder that we've work with, there's three big pain points for S. M. I start. Pain point.

One is getting the design tools, so that would be from a syn ops caddies pain point two is getting the right I P building blocks right, because a lot of these ships you don't need to reinvent well on everything. There's off the shop I P, you can buy for various pieces of that. And then third chAllen bucket is obvious ly, the foundry peace and the initial tape out and then brought a manufacturing.

Is there a way to work with the key players in each of those bucket? So E, D, A, I P and foundry to just ease the burden for these startups. And the concept that we use is that many startups are on the airplane there in the middle seat, the last of the plane.

Can we get them upgraded? The first class, all the large semi players benefit when they're interacting with those foundry players, the IP players, the eda players, exeter. And so it's been exciting to have those conversations because of many of those incumbent.

Almost three buckets benefit from having these startups in their ecosystem rather than one of their competitors ecosystems. Not all these companies will be successful, someone that will become big gosper in which company will be the next entrepreneur company, will be the next stair labs. There will be others like that in the future. And I think those different incomes are very excited about working closely with them.

But it's been fun. I've work with tim keller is a well accepted genius of the field, who's also entrepreneur. Is that generally the description of the founders and the space forces, the twenty two year old drop out Y C types, or the sort classics, so I can value a disrupter talk us through, like the nature of the people building these things.

I would say the vast majority tend to be much more experience. So these are individuals that have run divisions at the large public semi companies, right, that that have had the experience of building note chips scale, which is, I didn't deff much different, the software where you have a much a Younger average of your startup ound and maybe even twenty years difference or something along those lies are some exceptions to that.

I know with, actually, Gavin does IT fascinating work there. And I were laughing within the first time. I was like, do they teach you any this at harvard? And not really. But IT is, I think, for the most, getting more of the exception than the rule. This is going to be a much more experience, space, manage that you're gna be investment .

behind what do you compete with for deals. One of the things i've learned talking you is the interesting distribution of where capital is available for investing in the especially in private companies. And obviously, this can change really quickly and supply rises to meet demand and so on.

But IT does strike me that there aren't a lot of people doing in a dedicated way what you're trying to do in private markets for the semis ecosystem right now. So who in a given deal that you are trying to invest in that you you think is the best deal of the year, who are you up against most? sofi.

Think about the lay of the land of the semiconductor prison or investing landscape. You have some very good early stage folks like lippo IT, well in for examples to find its hill. These are just experts in their craft started that early stage. Company information period. And then I think when you look at the very latest stages and even the pri po round, you have some what I would deem to be the public market players felt has got a very strong team and a management leading that they will come into this preist o around. But in that DDL section, let's call IT series b and beyond, where the company is really start to pick up.

I think there's a real earth and shortage of capital in that market and think that the everything that was so excited about because as capital pick up for these companies, and that right is that, that tape outs happening in the IP costa starting to ramp because you actually starting manufacture chips, the E, D, A. And emulation costs start to pick up because you're test in the tip to make sure that it's working. Know that in total fifty million hours.

And there's just not that many players that are willing to commit that level of capital to these companies and make once you're able to satisfy that need, the rework can be enormous. We've seen that with some these recent ipos. I think just comparing IT took software, for example, where I think there's a very long list of capital providers that are going after that market. I think it's Frankly just much shorter when IT comes. And I conceive .

if you could fly anywhere tomorrow and have a meeting of your choice based on what's going on right now, where would you go?

I like the part of the industry that is harder to understand right now is in china actually just went back early this year, first time since covet. But because of the politics, the regulation industries that used to be very closely net, our being split apart. And so I think standing, for example, how we, which is the A I leader in china producing chinese leadership, use I D moderate scale, that is to be the company that defines where china .

is on the comparison is small .

right now out of individual and they're trying to ramp up and the governments trying to slow the rap. And one of the key questions, I think, over the next couple years is will they succeed or will they face real supply constraints for the next couple of years?

You tell the paw I story, how we got to hear.

And what happens is really is a fascinating company, I think had had not been for the politics. We probably on song, real electronic powerhouse and in some ways, still telecoms and that working equipment founded in sanja, right across the board from hongkong, and grew by selling first to the chinese market, but very early on pivoting to international sales.

And so they were learned from our customers abroad, improve their technology, some my piece after along the way, but was IT more or less than the average. Hard to say. IT is only really the last ten or fifteen years of the relationship between the quality and chinese stake got closer.

And now given the politics, who always business model is to sell as a almost politics producer into a chinese market that is constrained both by the U. S. And the chinese government of the company itself said to shift a lot, and now it's the chinese A I monopoly player may .

be described what the U. S. Government did visible while I just to complete the story.

So the us. Government doesn't want china to produce cutting edged chips. And so how we works with sm c, the leading chinese chipmaker taxi manufacturer, its ship smoke is like the tmc of china, but smc is allowed to buy the most events tools, including the newer lathom phy systems from sml.

And so right now, sick is trying to produce pretty close to cutting GPU using second tear machinery. And it's learning and just how hard that is to produce namesake ed transistors using tools that weren't really designed for that purpose. And so I think there is a scrappy company. They are innovating on the fly, but they're also hitting real scale liberations.

at least for now. We think about the components of returns you have mentioned. The entry Prices can be reasonable sometimes in the space rather to mothers stuff you've seen.

What about just like the size they can get to the capital intensity along the way that will delete U. S. An investor if you're an early investor and the past to exit? Those seem like the remaining three components. I love you just like pine on each of the three.

I think maybe will start with the last one in terms of the path to exit. If there's ten companies, my guess is one or two, one of going public of IPO around. I think the majority will end up getting acquired by the large establish semi companies that will be using these startups to help basically bold onto their flagship efforts where they can buy a engineering team, where they can buy a product that can really solve product gap that they have.

In these are companies that are generated and collectively, something like two hundred billion dollars, a free cash show year. So there's ample drive power to pay for assets like this, even to the tune one to two to three billion dollars. And we've seen several examples that happened over the recent turn on that is how I think about the exit path.

A couple of work in terms of the capital needs and the dilution on is required. I think it's interesting because that there is this mindset was semi, that they are very capital sive, which I think is broadly you I mean, each companies making fab, making their chip, but there are capable needs in terms of some other start of cost that we talked about having a deep engineering team, acta, that are not inexpensive. But I think comparing to software, yeah, software is not very capital intensive.

When you start, you can only have three or four people and you can do IT on the cloud and doesn't cost that much to spend up the first product. But I think to scale soft, actually capital and city is just that. Capital intensity hits you later in semis. IT hits you early.

But then I think actually the margin profile of a Young semi company can be actually be far more attractive than often are a company that hundred million dollars of sales and is Operating at thirty five, forty percent Operating margins relative to a software company that could be still burning significant cash, which may be the right decision for that. A company of the ltv of customers that when in will justify that level of span. But I don't think that you should kill yourself.

That software is not cap, but intense, I think just went in that lifecycle of the company that IT hit. you. Tell me that how for that?

What about the word bubble? When does the word bubble into your minds, if at all?

I think what people talk about bubble, that means you probably aren't in a bubble.

Usually one role fun.

It's usually only like afterwards club. So I think that's probably one just broad observation. Whenever people talk about about any people who are talking about this being a bubble like twelve months ago, right? And I think it's prevent not be the case to date.

And so I like just one over arching view. As I mentioned earlier, i've been investing in something to long and up to. This is a cycle, an industry, and at some point, there will be a digestion period.

But I will say that this is a growth cyc. This is actually very attractive and market to be putting dollars into. And I think I think a if you look at where the industrial structure is today and where the Operating margins are relative to some these other sectors and tacket very attractive.

And so if you have the stomach to deal with, if there is a bubble, there's a digestion period, what have you? I think there is actually can create opportunity because I think we know the end point is going to be up to the right because that can be a strain line up there. Probably not. But I think taking advances of any downcycled hopefully can do that from .

a position of strength review to study or incurred fuel to study any of the historical episodes and semis to Better understand the future. What would you have them study the P.

C. Revolution and intels role in IT. Give you example what it's like to be early in the stage of a grow cycle. The chip industry was always take on d RAM, the memory chips that were also also puter consumer devices.

But l was early the PC revolution and is able to ride that wave two and half decades, three decades of which the recycles in the mist of IT. But because more people around PC, you didn't notice those cycles anywhere near the way you notice the DRAM cycle because they were just less protected. And so I think that's the optimistic view of where we are right now is that is a cycle as understand heading up to the right.

You can suffer some up and down. And I think that's probably revenues because we're seeing an entirely new type of new category of chips emerged. The prominence that we hadn't used before. And that's what happened in the eighties with microprocessors that were used for the first PC really mesh market before then. Then PC made the mass market.

How do we get to this state of the world? Were so much of the manufacturing done overseas?

I think there are two drivers of the shift overseas. I think. One was that U. S. Companies naturally, in some cases, rightly focused on what they were good at, which was designed.

And two is that manufacturing was cheaper overseas today, is cheaper in taiwan, not primarily because labor costs. There is actually very people at work and highly automated is cheaper because of the ecosystem dynamics, very hard to reverse. And I think for that reason is, is also why companies are less excited about investing a dollar manufacturing versus design.

Because if you're invested a dollar manufacturer and you've got higher costs, less develop ecosystem, you probably rather developed IT, put IT in design government and to change that with the chip's act within forty billion dollars at work in the us. And that's a lot of money. I think important to keep that.

My money in context, T S. M C. Spent almost at much in capex last year. And so even really substantial government money is enough to moves the needle. But IT moves the needle on the margin. And it's onna take, I think, a long time to really shift for manufacturing happens.

What could most disrupt the trajectory of this whole story? What could be an example like a new architecture, not the transformer, but something different that rekick off a different kind of rate of improvement? Or usually talked about maybe the scaling laws level out and that could change things. What other things could happen that would drastically change the trajectory of the semi industry memory?

One of the key limiting factors in performance improvements, they need to move data back and forth from processor of memory. And so there's a variety of new memory architectures that are being experimented with, emphasize, experimented rather than prototyped. But a lot of researchers are part of this is the memory wall.

The key pain point is your memory is too far from your processor, and this slows everything. And so i'm not going to bed on that next year, the year after. But down the road, we might need new member architectures.

Maybe just explain in a little bit more detail what is literally going on there, what is happening? And then if I were to get revolutionized, what IT would unlock.

you literally have a memory chip and the processor ship. So the case of export of your G P U in your habit of memory, and you need to move dat betwen them, which you mean to move electron's between them and the distance, coupled with the speed at which electron's move, which is slower than the speed of light, for example, you could move tonics and increasingly do create friction in your training. And so if you were the design ships differently to have processor memory more declined, close together, or if you switched to optical interconnect, those are always you can increase the speed at .

which that happens and the not kind effect that would be just a more efficient system.

more system, but change across, across and changing producing them. I think .

that's fair. You have this dynamic and industry where the GPU performance has been increasing at a much faster rate than the networking in memory performance. And so you're not able to get as much realization out of that GPU as you would like because of some of these other bottle.

Next, what happens if the networking piece and that memory piece can keep up? Does that end up just slowing down, just the level of innovation that we're seen in the marketing, the ability to train a bigger and bigger model? I think my expectation would be there's a lot of interactivity happening at networking in memory level to solve some of those pain points and catch up with the performance improvement that you're seeing with the GPU. But to me, I think that's a key issue to understand because if you are in a situation where you got this incredibly powerful processor, but you can get data in and out of IT quickly, for you can get enough data and out of IT, that being a bottle ck, that slows down innovation.

I just imagine that your whole portfolio s already built the general types of companies that are going to invest in, on average, what do you think the source of sustainable power and differentiation will come from business power, not a little power will come from for those companies? Is IT mostly scale? Is intellectual property? Is IT cornering the best researchers as you think about not just who's gonna grow, but who's gonna grow a defensible way? Where do expect most of that defensively will come from?

Think it's gotto come from the intellectual property peace train, the value that these companies are bring the table on, the problems that they're solving because the skill piece comes from the founder players. These starves are designing the chips from the most part, and there are going outsource. That peace of the scale gets solved by that third party foundry, whereas think the value really is going to be created from the pain point that company is solving in. Are they able to do IT in a way that no one else can?

Is that world like A I model research where frustrating how few people seem to really matter in its like dozens of people that seem to have any real impact on miles?

I think it's broader than that to be told. candidate. I think you've got in silicon in particular, right? You've got thousands of engineers, tens of thousands, Frankly, even hundreds of thousands, right, that are very technical.

Are they all going to be start at founders? Definitely not. I think IT is actually proud a wider town pool.

And maybe what you're seeing with some of these I kind of front of your mom place seems like there's these hundred eight engineers, right? If you're able to have one of them, that solves all your problems. I think there are folks like that in the same industry for sure. You highly jim keller, his reputation and they clearly proceeds him and there's others. But I think in the brother speaking, and I don't know, you know quite that same painting.

what do you two most commonly or disagree on?

I don't know if we do. I mean, we talk a lot about government stuff, right, in terms of when headlines come out and there's a lot of news flow coming out of D C. And just trying to understand what IT means and what the intentions are of the various restrictions that are coming out of where chips act related.

What are the end goals that they're trying to achieve with got the fab side in the R N D side. Chris has got incredibly tight relationships there. Member, I told I was in d last minute.

Then you had me meaning with biden. Chip is on. I was like in the west wing of the White house, like four hours later.

So you d definitely have those people really trust to respect what you have to say. But I think it's more of a collaboration is turning. Like what does this really mean? I think you are probably more effective than i'm reading the T, V.

is a lot of this, yeah the other and is a first statement. There are more different ecosystems. I need to be brought together to understand the full picture. You can understand the manufacturing, the IP, the end users in the data center, the supplies is longer than a soft company. For example, you've got a stand where all the players are to see where technology's going.

That's a very interestingly, right, because I think you look, it's often that tends to be much more salad and in the same mechanical systems. And we touch on these different players earlier in the conversation were like so many different folks just inner lap. And it's very interwoven where everything from semi cap is making the tools to the foundry player to the design folks.

And even with design, there's these different buckets like mobile and computer and there's analog. Any of the E D A fox are signed, the tools that are designing the chips and then you have the I P players like ARM exter. But like everyone needs to work together to push the industry forward and so be able to see that bigger pitcher and have the connection vary with those different players, I think is a really critical piece of successfully pursuing this market. And I just think it's different than what you see in in software.

For example, I was when I first started learning about this industry, I figured with someone I could call to get a view of the whole thing, and I quickly realized that there are brilliant people who know their slice and no bit about the slice below and above. But there are so many steps and supply in that actually it's really hard, almost impossible, to have a view of the whole picture.

What are the most sensitive parts of spot chain manufacturing?

Stepper sensitive. But so too are the tools that make manufacturing possible. The sm level, if you will. But then you can do the manufacturing of the chemicals.

And the chemicals are made in many cases by a couple of japanese firms, which have unique knowledge to produce chemical acts with ninety nine point nine nine nine nine nine percent precision. And that's only the manufacturing inside. Then you look at design where the IP and the E. D, A. Tools are it's hard to say, one of is more critical than the other because you need all them.

Just feel like iraq or something in dune, like precious compute resource that just fuels everything else. And there's a revelation going on in IT that kind of we've never seen before. How big do you think this gets? How big is this industry in five.

ten years? And how big is the second is probably hundred billion dollars software, Polly, a little bit higher than that eight hundred billion dollars. When I think about where we go, maybe this is a old David. I think the market will see the software market by of the which just given this low, a shift into compute from people into compute. Is that being the primary resource?

It's gna drive the eye. I see that, that makes me optimistic with the future rate progress is the application of compute to solving chip design and ship manufacturing problems. This is sort of fly wheel of we make Better chips, we produce more, and we use IT, first and foremost s to make Better ships and produce more compute.

So whether is that the chip design, where the leading chip design companies have essentially been doing A I for chip design for decades now, you can't lay out a chip with ten billion transistors at computing to do IT for you. Or whether is the ethnography process, how do you know what shape to print is now extraordinary ouns computation? I go into the actual figuring out how to get the shape you want on your transistor out of the shape of the transistor in your piece of silicon is already huge use of computing in the designer production process. And I think there's immense excitement about ways to find even more applications for eye to make the chip design and production process more efficient.

Anything we've missed that you think is critical to whats going on in this world that you spend time on that we have talked well.

thing that sends or to me is that one might think the more advanced ships we need, the fewer, less advanced PS we need. But it's actually exact opposite. The more advantages we need, the more less advantaged PS we need to be in the systems around.

Take a new car for example, you've got a couple of pretty high processors for the self driving. But if you need to self driving, you need different sensors around that. You need more chips managing the data transfer between different sensors. And so a new car will have more lower tech ships than an old car will, in addition, having a couple of higher and ships. I think that's an dynamic for the entire industry because that means even if you're not in video, if you're an analog devices, which produces a lot of sensors, for example, really interesting transformation for you too, because the senses converting the data that will be necessary for A I training inference. And you see that and basically every corner .

of the industry investment opportunities, that strikes me as something that is a far less sexy story and therefore perhaps more interesting Price, but nonetheless similar long term prospect in terms of market size.

Ah absolutely. I think that kind of tizer in that broader A J I bucket. And we thought a lot of these companies there more in a log base trade and so not the leading age digital chips where you try to make a very clear with them what just happened in the data center, that tsunami I is coming for you next.

And if you don't embrace change here and wide ed in your appeared to be able to put intelligence into your historically kind of dumb chips. You're going to be on the wrong side of change. And some onna come disruptive, just like you saw in video, disrupt the incoming in the data center market. And I think anyone is on the edge. SHE realized that that wave is coming for them next.

If I to force you both to leap frog one or two degrees away from just core semis, focus what is going on in the world right now. The most interesting.

the plays sending out of time is in biotech genetics space, which I think not this decade, but next decade will be a very interesting intersection of improves my auto tragical logic and biotech. I I think if you spend time at places like I leaving universities and ask to the P. C. Researchers working on, there's a whole lot of interesting work being done there, which I expect to lead the information technology down the road.

So i've been living in reading, sam here in particular, over the last nine months. So this is a refreshing question, I think, just trying to zoom out. And this ties into the semi piece of the equation. But like how do you help create a more viBrant, healthy status ecosystem for this set of and for this industry? And how do you collaborate with folks in dc, which I think, to their credit, the chip's officers, doing a great job being proactive, talking investors, talking to company, is not trying to be in that dc echo chAmber.

What can be done to the bottles, neck all these pain points that makes you so chAllenging for a chip company to start up? How do you motivate that very small engineer? Who's that in video right now, our M, D, or broadcast say, hey, i'm going to start this new company because I think the rewards of that are going to be both financially and personally massive.

And so help creating that culture is something that I try to think about because something is real important for the U. S. Broadly to be competitive on the designs I were dominated today. But the fear is that does that slip away because you don't have a very active start a pipon pushing the accompts to innovate and keep moving this ball down the plane field.

And so that's something that I think about a lot and try to figure out how can you collaborate with a different players in this ecosystem, whether be the inconvenience, whether be the U. S. Government or other governments, to create a really viBrant, healthy ecosystem.

Having read the book and talk you a lot about what you're building at max silan, i'm really excited to see the companies that started to come through here like you guys have put so much time and thought and attention to not just who to back, but how to help back them. I'm sure the signal coming off of what your investment is gonna a fascinating way. You'll be my answer to my earlier question about who to call the figure, what's going on, my traditional closing question for everyone, what is the kinds thing that anyones ever done for each of you?

The caveat here will be this is excluding family. So this ties back to when we had lunch this summer, I was leaving for that trip, and my first trip, taiwan. And so if I go back to high school and college, the mentorship that I got from my teachers and my coaches over that period time, and such a huge impact on who I am today, that is kind of anyones ever done from reason, my bring of that taiwan trips.

So that end up turning into two taiwan trips over the course of eight days. And I was at an airport hotel working out of the jam, trying to maintain some sembLance of fitness, and got an email from my rowing coach from high school, the prize, and coach me on the water for over twenty years. Just hey, and check in and how you doing? How is your family doing? And I just totally bright in my day, and just individuals still cares about me.

Long after I left this program and I left you A I high like really brought us milk in my face. And I think just trying to do that at maverick, can we mentor that next generation of leaders at the firm and spend time with them and invest in I try to do that like people do with me. I think the job, I think that is a really valuable lesson. And I just encourage everyone take the time to do that .

and think it's important to mentor as well, adviser, when I was doing my PHD in history time at us, who often when you A P advice or your advice want you to study what they study, that's the standard trend. And Johnson studied everyone to study and gave me a sort of an intellectual blank check and trust that I could find something interesting and accurate to say whatever feel I wanted to.

And at the time, that was a niche area of russian history when I was study them. But IT, I think, helped me realized that if you other set of tools, you could have confidence to go to new area, understand what was happening there, makes sense of IT, and produce something of value at the end of IT. And that initial willingness to take a bad and write that black into actual check, something that I valued today, hofus ly produce.

valued from the word's infrastructure, is changing. Thank you for teaching us about what's going on.

Thanks for time. Thank you. thanks.

If you enjoy this episode to check out, join colosse dot com. There you will find every episode of this podcast complete with transcripts showing tes and resources to keep learning. You can also sign up for our newsletter glosses weekly, where we condense episodes to the big ideas, quotations and more, as well as share the best content we find on the internet every week.