0:00
Brenton John, so glad to talk to you guys today. Thanks, Shane. Thanks for having us on. We're really excited. Likewise. Thanks for having us.
Search this video
0:06
So, we're going to talk semiconductors today.
0:09
And this is like such a a big concept.
0:12
It's hard to wrap your mind around.
0:14
I I think the best starting point is sort of like going back to the beginning and the origins of the semiconductor and then almost working our way up to now.
0:26
Yeah, it's a fascinating topic.
0:26
It it really isn't that well understood by lots of folks these days.
0:32
But of course, you know, the semiconductor came out of pure science from Bell Labs back, you know, when AT&T did pure science when when we had companies in the US that did pure science.
0:45
And really morphed and changed the world over time, but morphed certainly changed lots of companies that created Intel, created Texas Instruments, created really the modern electronics movement.
0:56
So, there's been a lot of moving pieces.
0:59
A lot of the foundations have been in the US, but certainly it's been global over time. That's awesome.
1:05
So, where where do we start?
1:07
Like what's the entry point into this vector here?
1:11
John, you want to give it a go and I'll I'll tag on? Sure.
1:15
I guess the way we think about it and what we talk about a lot is that semiconductors are really the building blocks for the for the digital economy.
1:22
And what that means is you can kind of look around the world and think there are pockets of our lives and pockets of the economy that are you have pretty clearly already been impacted by technology and my semis.
1:32
Like we're all living on Zoom and you know, working on the cloud these days, but also spending, you know, 8 hours a day on our phones.
1:39
But then you you go around the rest of the economy and think about, you know, the next car you buy.
1:44
You know, that decision will likely be somewhat impacted by semiconductors.
1:45
Is it, you know, an electric vehicle?
1:49
Does it have like some level of autonomous driving or or kind of you know, active safety or like you know, lane departure detection things like that.
1:55
Think about how, you know, drugs will be delivered in the future and even how drugs are discovered.
1:59
Like think about right now living in in COVID times that we actually mapped the the COVID-19 genome I think faster than any other you know, germ that that I'm aware of.
2:09
It was it was mapped in in January before you know, COVID even really hit the US.
2:13
And so, we think about it just from a strategic perspective, you know, semis are going into everything.
2:18
And if you think about all the the really most exciting trends over the next kind of 10 to 15 years, whether it's you know, like I said, you know, automotive or artificial intelligence or kind of the connected factory and internet of things, semis kind of are support all of that.
2:30
And so, we kind of feel like semiconductors are kind of the unsung hero of the technology sector that everyone gets excited about software and we all kind of know that software is eating the world and I think that that's fair, but kind of the underlying support layer of that really is is semis.
2:43
And that's why we're we're excited about it.
2:47
It's like a key strategic resource for everybody, but there's only a couple people that can actually fab these days, isn't there? Yeah, that's right.
2:54
If we think about, you know, back to the question, uh you know, how do we get here, right?
2:59
It's it's a fascinating story.
3:01
We won't go through all of it here.
3:05
But if you think about a company like Texas Instruments, which was one of the first semiconductor makers along with Intel and Fairchild, um they really used to do everything soup to nuts.
3:15
They made their own equipment.
3:17
They had their own factory to make the chips.
3:20
They of course sold and marketed the chips and they even sometimes sold and marketed the end device.
3:24
I mean, we still have Texas Instruments calculators.
3:27
And they did the speak and spell sort of back in the day.
3:30
So, they did they they didn't they made radios.
3:32
You know, they they did everything.
3:33
Over time, of course, more entrants entered the market.
3:37
The market the industry matured quite a bit.
3:42
And then we really went to this horizontal model.
3:44
So, there are equipment makers that only make equipment for to to fab chips.
3:49
There are companies that build the chips.
3:52
There are of course other companies that build the end device like Apple or Tesla.
3:55
And then now it's sort of coming full circle where Apple is making some of their own chips and Tesla is making their own chips.
4:03
But when you say making, do you mean fabing or do you mean making in a different way?
4:09
Yeah, no, I do mean fabing. So, okay.
4:14
When when Texas Instruments was making their own chips, they had a very promising CEO candidate named Morris Chang.
4:22
And Morris turns out was passed over for the CEO spot and decided to take a bow out of the industry for a while.
4:30
He went back to his home in Taiwan.
4:32
Taiwan said, "Hey, get us into semiconductors."
4:33
And he said, "You don't have anything here.
4:38
It's really hard to build chips and you're on an island that has a lot of earthquakes.
4:42
This is, you know, like I don't know what to do."
4:44
And so, what he realized from talking to customers is that actually fabing a chip is the pain point. It's really expensive.
4:52
It's really difficult even back then.
4:55
And there are only a few companies that are all that good at it.
4:57
So, he said, "Okay, well, what if we made a business that we just made other people's chips.
5:03
We didn't compete with anybody.
5:03
We just actually fabed their chips."
5:05
So, to answer your question, Shane, when I say Apple and and and Tesla are fabing their own chips, to be more precise, TSMC is fabing chips for Apple.
5:15
And in the case of Tesla, Samsung in Austin is is fabing that Tesla chip.
5:20
So, Morris really created this new business, which we call the foundry model.
5:26
So, walk me through the value chain there.
5:28
So, if like Apple is designing the chip somebody on a different architecture.
5:33
So, you have an architecture, you have Apple engineers designing the chip and you have TSM or sort of Samsung fabing the chip at the end of the day and then Apple integrating it into their their hardware product through Foxconn.
5:47
Where's the value add in that?
5:47
Where's the where's the strategic sort of nature of that going?
5:51
Why aren't these companies vertically integrated?
5:53
Why are there so few people that can fab?
5:56
I mean, I have so many questions. Those are all good ones.
6:00
Yeah, I I can start with that.
6:02
So, I I guess to start out like thinking about just the chip design process overall.
6:06
So, if you're a semiconductor engineer at Apple and they have, you know, thousands of them now, which in itself is incredible.
6:13
You design and lay out the chip on on EDA software, electronic design automation software from really only two companies provide this now.
6:20
It's Synopsys and Cadence Design Systems.
6:21
And so, that in itself is kind of CAD software ecosystem they've created is already a like a really robust part of the overall kind of value stack.
6:30
And so, if you're an EDA company like Cadence or Synopsys, you actually collaborate really closely with Apple, but also with TSMC, who's the end manufacturer to make sure that you can actually, you know, build the design that Apple has laid out because at the end of the day, you're really kind of bending the laws of physics.
6:44
If you're thinking about Apple you know, designing a 5 nanometer chip that's going to be in the next iPhone or even in the new the new MacBook, the precision with which they're doing that is is phenomenal.
6:54
And so, so Apple will design the chip on on EDA software and then TSMC Taiwan Semi is really, you know, the main manufacturer in the world for for leading edge semiconductors.
7:05
Samsung is kind of the the second source, but in terms of, you know, really like bleeding edge digital chips you would buy from an Apple or an Nvidia or someone like that, that's all done at TSMC.
7:15
And so, and then the other part of the value chain is the equipment part of the value chain.
7:20
So, TSMC you know, buys equipment from, you know, really a small handful of strategic suppliers, which would be ASML, which is in the in the Netherlands, you know, Lam Research and KLA-Tencor and Applied Materials in the US and then Tokyo Electron in Japan.
7:33
And so, one one of the things that's amazing about just the overall kind of like value stack in semiconductors is there's a lot of margin throughout the whole value chain.
7:41
And that really to your question, Shane, is because I mean, the industry has evolved over 40 years to this more horizontal model because it really is just really difficult to do the whole thing.
7:51
I mean, like Intel used to do their own kind of internal EDA software and they ended up just going and outsourcing it more to Cadence and Synopsys because it really is just like so freaking hard every, you know, 2 years if you're following Moore's Law to to do every part of the value stack yourself.
8:05
And so, you kind of see that with with the software, with um the equipment guys, with TSMC doing the manufacturing and with the the chip designer itself.
8:15
And then the important part that that I kind of hit on is is there is this ecosystem dynamic where they all collaborate very very closely.
8:19
And so, there's kind of just this iterative process that's really, you know, unfolded over 40 years that has created, you know, a lot of really well-positioned companies in various parts of the value stack.
8:31
But being so complicated, it's hard to compete.
8:33
Like how would you compete with a TSM today?
8:34
Like how would you build a a foundry as a manufacturer?
8:37
Like how much capital would that take?
8:40
Are we moving to a world where that's going to be more competitive or less competitive?
8:46
Or are the the I guess the the rich getting richer in this case where the more chips TSM makes, the more they're going to make in the future, the bigger their advantage gets over other.
8:55
And then I want to talk about the geopolitical sort of like globalization risks of all of this.
9:02
Yeah, one of the questions we ask ourselves quite a bit is, you know, if this company got hit by a nuclear bomb, would it matter?
9:10
You know, if the nuclear bomb was somehow able to trace every tendril of the company throughout the world and and just blow it up, would it matter?
9:18
Most companies you see, not really.
9:18
You know, CVS gets blows up blown up, I'll go to Walgreens.
9:22
If Starbucks gets blown up, I'll go to Pete's.
9:24
If Facebook gets blown up, I'll communicate some other way, right?
9:27
Um If a couple of companies in the semiconductor value chain that John was talking about get blown up, the whole world goes backwards.
9:38
10 years, 20 years, I don't know, a long way.
9:41
There's really only a few companies that can do this.
9:43
And to your point, Shane, it's not a matter of throwing money at it.
9:47
Technically, it is extremely difficult.
9:49
These are some of the most complex machines humans have ever built.
9:54
They're virtually impossible to reverse engineer.
9:59
And in the case of something like photolithography, which allows Moore's law to to go ahead, there's only one company that builds the next generation machine that is keeping Moore's law on track.
10:08
It's ASML and the Netherlands.
10:11
So, there are a handful of companies that have become um, because of this long history of consolidation combined with uh, sort of the the non-linear pace of complexity um, extremely important in the world.
10:28
And and to answer um, you know, Shane's question on how you compete with TSMC, there's also I mean, there is like a real world example of this where AMD actually used to own fabs the way that Intel still does.
10:36
And they actually spun out their fabs um, and and their fab um, their fabs into it and into an external foundry.
10:42
Um, into a company called Global Foundries.
10:44
And this was probably, you know, 2009 or 2010.
10:48
And they were owned by um, by Middle Eastern sovereign wealth money.
10:52
And so, they basically had unlimited resources to go out and compete with TSMC and try to replicate their model.
10:57
Um, and then about two years ago, they were a little bit behind TSMC and they were trying to migrate down to the next process technology node, which at that point was 10 nanometer.
11:04
And they threw up their hands and they they literally just said, "This is too complicated and the capital investment in the technology, the lift is is just too heavy.
11:11
So, we're we're just going to kind of like hang back here and and make a lot of money actually doing, you know, kind of older technology.
11:16
But it just it really is like a case study real time of of building the ecosystem and migrating that um, down that technology road map, you know, every every two years is just such a treadmill.
11:25
Um, and so to your to your question, I think at least in from TSMC's perspective, you know, the rich do get richer and then, you know, Samsung can kind of linger as a second source also.
11:35
Is it isn't China building a competitor to TSM and Samsung?
11:39
Yeah, they they've been working on this for a long time. Um, over 20 years.
11:44
It's called SMIC um, is the fab.
11:47
It's Chinese based, of course, with the recent uh, you know, limits on um, or tariffs on what we are allowed to ship China um, both from the US side and also from our allies even in the Netherlands.
11:59
Uh, that has set them back quite a bit.
12:02
They've literally spent tens of billions of dollars on this and they remain um, you know, around a decade behind.
12:10
What can't they just do what they always do and sort of reverse engineer or steal the technology?
12:16
Yeah, I mean, it's a good question.
12:16
I mean, for for now what they're what they're actually trying to do, like their their kind of public road map for SMIC is they they're trying to build a um, a factory with no US equipment because I think they have to have to at least, you know, face the reality that there's a scenario where where China gets does not get their hands on US semiconductor manufacturing equipment for the next, you know, decade. And so, what do you do?
12:36
And so, their their road map for that for now is they um, they're trying to build a 28 nanometer manufacturing line, which is actually about at by in in the next three years.
12:44
And so, by the time they actually have that up and running, it will be 10-year-old technology.
12:48
And this is, you know, in an in an industry where, you know, technology is scaling exponentially.
12:54
And so, what they have to do um, really is I mean, they have to re-engineer all the process technology that TSMC has built, which, you know, has taken decades to build with very tight collaboration with their customers um, like with an Apple or an Nvidia or an AMD on, you know, packaging technology, how you stitch everything together, what's the easiest way to get it from, you know, software actually onto the chip.
13:13
How do you actually build chips that, you know, actually have high yields so that customers um, don't have defects and can actually ship them.
13:17
Um, and then China has to um, try to replicate all of the the the manufacturing technology in terms of equipment, which is also um, I I think probably a 20-year project in terms of um, replicating the various kind of parts of the manufacturing value chain.
13:33
And Brit Britain mentioned ASML, which literally has a monopoly on photolithography, which is probably one of the most kind of complex processes that humankind has ever engineered.
13:40
I mean, it literally is more difficult than putting a man on the moon or building a 747 or you kind of pick what, Can you walk me through that? What makes it difficult?
13:48
Like, walk me through what that what what is lithography?
13:50
Like, Yeah, well, there's a couple of things to understand um, and I and I have to get a little bit technical, but I'll I'll try not to get too technical.
14:00
Lithography is when you shine a beam of light at the silicon wafer and and therefore, you harden parts of the silicon wafer so that you can go back and etch or or kind of scratched out the lines between it.
14:10
And so, when you think of a transistor, a transistor is a source, a gate and a drain.
14:14
Uh, and you want to put these transistors as close as possible on the silicon wafer. Right?
14:21
So, you have to etch these lines in between them.
14:22
And then that's what we think of as nanometer, right?
14:24
Like seven nanometer, five Exactly.
14:27
Yeah, so the most leading edge chips these days are are five nanometer chips.
14:31
And a nanometer in silicon is about two to three atoms of silicon.
14:33
So, we're talking extremely small.
14:35
Um, you know, five nanometers less than 15 atoms of silicon divide the lines.
14:40
So, uh, there's a problem.
14:44
A wavelength of traditional light is 193 nanometers, but we're making chips at five nanometers. How do we do that?
14:51
Well, we've had all sorts of tricks that we've done over the past, you know, decade and a half or so.
14:57
Um, we shoot it through a lens that shrinks the wavelength of light.
14:59
We shoot it through water.
15:00
Um, it's called immersion lithography.
15:02
It also shrinks the wavelength of light.
15:04
But uh, at at around 10 nanometers, it became very clear we had to change the light source to something that was a smaller wavelength.
15:13
And the the answer that everyone came to um, primarily enabled by ASML is um, something called extreme ultraviolet light or EUV.
15:22
EUV is 10 and a half nanometers.
15:28
Um, so that Did I get that right?
15:30
John, 10 and a half, right? I think so. Thank you. 10 and a half, right?
15:34
light starts at 170 and 193.
15:36
193 and then the extreme starts at the extreme ultraviolet starts at 10 and a half. Okay. Exactly.
15:41
Yeah, so to to create extreme ultraviolet light.
15:45
Now, this is the part that blows your mind.
15:49
You need to drop molten tin and hit it with a laser.
15:55
You need to perfectly pulse those. Right?
15:57
So, the molten tin is coming down.
15:58
You hit it at the exact point of time with the laser that explodes it into a plasma, which creates extreme ultraviolet light.
16:07
Um, now that sounds hard, right? It's in a vacuum. So, it's not too hard.
16:09
But then you think, okay, well, how how often do I need to do that to create a bright enough source to then reflect off of a bunch of different mirrors and finally go down on the chip.
16:20
Um, well, it turns out the number is about 50,000 times a second that you need to get that right.
16:23
And it has to be exact, precise every time.
16:25
But extreme ultraviolet light is the the wavelength is so small that it's absorbed by all known mirrors.
16:33
So, it doesn't reflect, but you have to bounce that off about 20 mirrors in the machine, plus or minus.
16:37
I may be getting this a little bit wrong, but I'm I'm overall right.
16:41
Um, and then and then still have enough impact on the when it gets to the silicon to be able to um, to do its job, to harden that that silicon.
16:49
So, ASML had to create mirrors uh, to be able to bounce it.
16:54
That was a brand new science experiment. They created those.
16:56
They had to create the laser that was powerful enough to do that.
16:58
Um, there's only one company in the world that makes that laser.
17:01
It's a private company in Germany.
17:02
Uh, so this is mind-blowingly hard.
17:05
I mean, John said something the other day that was like, you know, it's the equivalent of shooting an arrow from the earth and hitting an apple on the moon. It really is that hard.
17:14
It it's just ridiculously like somebody they should win two or three Nobel Prizes hard.
17:20
And it's another example of actually their competitor just dropping out.
17:21
I mean, there for a long time there were two companies in the lithography market.
17:27
Um, it was ASML and Nikon.
17:27
And then again, that at the beginning of last decade, Nikon just threw their hands up and said, "We just don't have the the capital and expertise and um, kind of like willpower to pursue EUV."
17:35
Like they didn't even attempt it.
17:37
And so, that's why ASML is kind of this natural monopoly.
17:41
They're actually now the the most valuable company or tech company in Europe.
17:44
Um, so I think it's a little bit discovered, but um, what they do is is just kind of like unprecedented.
17:48
And another like amazing fact about EUV systems is that it takes four 747s packed full of equipment just to transport one EUV system.
17:55
And they cost right now, they cost about $200 million and that that will actually go up to about $300 million over the next like four or five years.
18:02
Um, and so, to just the the engineering lift to to Britain's point on what they created is is is just like a crowning achievement of of humankind in our kind of geeky opinions.
18:13
What what's the runway on the extreme ultraviolet?
18:14
Like, um, regular light gave us up until 10 nanometers, I think.
18:19
What's the runway on the extreme? Can we go three? Can we do two?
18:23
Or do we need a new source?
18:26
Like, does somebody like ASML um, just get displaced by a new better technology?
18:31
Or do they uh, innovate on this?
18:35
Yeah, so they have they actually have given their road map um, through kind of like the middle of this decade at least.
18:40
They have they have a new tool called high numerical aperture um, lithography, which which Britain can give the equation um, that that has numerical aperture for for the the line width.
18:50
Um, but basically, that that's a tool that will be $300 million.
18:52
And so, that gets you um, towards the end of this decade.
18:55
And so, um, I think when we get down to like, it'll be, you know, two or one and a half nanometer or something like that exiting the decade.
19:01
I I do think that'll be um, time for the industry to start thinking about.
19:04
I mean, there there are different device structures or different materials and things like that.
19:08
Um, but the the ecosystem is so tightly collaborating on all this that I think even if it's not exactly, you know, the the generation after high NA, high numerical aperture um, EUV, like ASML will play some sort of a role in in advancing Moore's law just because they're I mean, first of all, they're putting, you know, billions of dollars of R&D into it along with, you know, Intel and TSMC and the whole ecosystem.
19:28
Um, but but also, there are just only so many companies that, you know, have the the scale and the kind of um, you know, the expertise and the the internal knowledge from how these things are built to to kind of be the the the company that will drive us on Moore's law from here.
19:42
Talk to me about Moore's law.
19:42
So, uh, just give listeners an idea of what it is.
19:46
And then talk about why it's proven so difficult to maintain.
19:49
And then, what's the trajectory here?
19:51
Like are we budding up against it?
19:53
Is there a 5-year visibility where we're not going to hit it?
19:58
At some point, like we reach 1 nanometer, do we go smaller than that? Is that even possible?
20:03
Like physically possible or like Yeah, walk me through your thinking on this.
20:08
Well, it's yeah, it's it's a great question.
20:11
There's a lot of we don't know answers to when we when we get to the logical conclusion of that question.
20:17
It Moore's Law was created by Gordon Moore who worked at Intel.
20:21
He basically back in the '60s I believe he wrote an article and he said, "Hey, we think we can double the transistor density about every 12 months."
20:28
He then revised that a few years later to say probably like every 12 to 24 months.
20:33
Um and so every 2 years what it means is for the same price you get double the numbers of transistors on a piece of silicon.
20:40
Uh that's enabled semiconductors to push into every last vestige of our lives.
20:45
Uh you know, 60 years ago or 50 years ago it wouldn't have been possible.
20:49
They would have been way too expensive.
20:50
But because we get double the amount every 2 years for free um and the the manufacturers get to keep equivalent profit margins at those rates, uh then it's been semiconductors have become democratized.
21:03
They've become into everything.
21:05
Of course that's what a lot of the stuff we talk about all the time is really all about.
21:08
AI is is about high-functioning um leading edge semiconductors, Internet of Things is about cheap uh lower end semiconductors, 5G, you know, we could kind of go down the list, machine learning, um uh uh uh autonomous driving, um you know, these are all enabled by semiconductors.
21:25
As John said at the beginning, um software gets a lot of attention, but really semiconductors do the work.
21:31
Um And it's harder than software.
21:35
And it's way harder Yeah, it's it's it's extremely difficult.
21:37
Um so what happens over time with Moore's Law? We don't know.
21:41
Uh we we like John said, we see a runway to the end of the decade, we meaning collectively humankind.
21:48
I'm not an engineer here, I just play one on TV.
21:49
Um so uh but But it could be that we change the the substrate, so maybe not it's not silicon, maybe it's something else um where where we can uh you know, jam more um lines in a smaller space cuz the atoms are actually smaller.
22:06
Um so so maybe it's changing the substrate, maybe it's changing the way we do architecture like John said.
22:13
And then there's a like a host of other things that go into it.
22:17
So TSMC um recently gave an update and their chief scientist had about five things that they think will allow them to keep on Moore's Law through the end of the decade.
22:26
Um also combining chips into what are called chiplets, um putting stacking chips on top of each other and on the bare metal.
22:32
Uh and then and then other things.
22:35
So the we we spend a lot of time on on photo photolithography um but but there are a lot of other key things.
22:42
Like for example, we are now building transistors higher.
22:46
We're stacking them into what are called 3D transistors and there's one company that essentially that enables that, Lam Research.
22:50
Um the the design rules for doing these are extremely hard, getting harder as John said.
22:58
That's Cadence and Synopsys.
22:59
There's only two companies in the world that allow you If you're an engineer that's designing chips, that's your Excel.
23:04
That's what you go in and you live in every day.
23:05
Um So we don't know uh how Moore's Law will continue.
23:10
There are a bunch of um different options that we may go to in a decade, but this is not really a new problem either.
23:19
you we've never really known how it continues.
23:21
It just does through innovation, right?
23:22
Like in hindsight it's easy to say, "Oh, this is why it continued."
23:26
But at the time looking forward, you get maybe 1 to 2 years and then you're Yeah, I mean with EUV, when John and I started looking at the space, I don't know, let's call it 2005, people were talking about EUV and it was like, "Oh yeah, well, you know, if we can figure out how to get a laser bright enough and if we can get that, you know, 50,000 times a second to to fire precisely and if we can create a mirror to bounce that stuff around."
23:49
You know, there was a lot of ifs and it was like, "Wow, uh this is not a done deal by any means."
23:57
We did it as a species, but but it was really difficult.
23:59
Yeah, I guess what I'd add on the Moore's Law piece is there's kind of like two different ways to think about Moore's Law.
24:04
There's kind of like the Gordon Moore definition of doubling in transi- transistor density every 2 years and that that actually I think has kind of empirically ended I think.
24:11
Like the industry cannot actually scale um that way anymore just because we really are bumping up against the the laws of physics that Britain was talking about with like the number of atoms you're even looking at.
24:20
But then there is um there's really this idea of of can we just double computing capacity every 2 years?
24:26
And I think our computing performance every 2 years.
24:28
And I still I think um with kind of the broader efforts of the whole ecosystem, Britain mentioned you know, TSMC has their kind of five different um you know, efforts around the edges of of computing um with with those and that that's actually called like more than Moore, it's kind of a cheeky pun in the in the semi industry.
24:44
Um but kind of like what can we do outside of Moore's Law to keep driving performance?
24:47
And um and and those I think will will continue to drive us down um at least, you know, the theoretical Moore's Law of like are we getting more economic value from every process node transition?
24:56
And then we're also adopting new architectures like GPUs or specialized chips um for the data center or for for the edge.
25:02
Um you're seeing companies like Google and Amazon do their own silicon with that in mind.
25:07
That it's it's not just all about kind of like Intel brute forcing ourself down Moore's Law.
25:13
Um it it's really about everyone in the whole ecosystem being more serious about designing kind of specialized chips for various applications.
25:18
Talk to me a little bit about those uh competing architectures.
25:22
What are the competing architectures? Who are the key players?
25:25
And what are the the sort of different strategies that each architecture enables and and sort of gives you an advantage in?
25:33
Sure, I I can start and Britain can can jump in.
25:34
So x86 is kind of the the workhorse architecture of the industry for a long time.
25:39
That's kind of the core Intel um you know, architecture that has built kind of the PC industry and a lot of uh a lot of the cloud, like a lot of server chips are just um built on the the Intel x86 architecture.
25:48
And really the only um relevant companies there anymore are are AMD and Intel.
25:54
Um And then the other architecture that really came into um became really visible with the rise of the smartphone is the ARM architecture.
26:02
And ARM had a different business where they um kind of invented and refined the the architecture, but rather than building their own chips, they actually licensed the technology.
26:09
And so they um grant licenses to companies like a Qualcomm or an Apple um to to build processors that would power a smartphone.
26:17
And so what ARM did, you know, amazingly well is is low power.
26:21
Like in the you know, the middle of the 2000s, everything was an Intel x86 processor and everything had to be like plugged into the wall and have two fans and you know, we didn't really think about this world of low power and that's really what ARM facilitated.
26:31
Um and then more recently what we've seen is the rise of the GPU and that's that's really kind of been Nvidia's um stronghold.
26:40
AMD also has has a GPU product and so those are processors that really started for for gaming and kind of rendering images.
26:45
And then what we discovered kind of in 2014 and 2015 is GPUs are just amazingly well suited for especially honestly as Moore's Law and kind of Moore's Law's cousin is um Dennard scaling kind of slowed down that as compute power was slowing you can actually put a few GPUs next to a CPU and speed up a workload with incredible success.
27:03
And that that kind of actually is what started this whole explosion in in artificial intelligence we've seen over the last 5 years.
27:08
It's kind of the rise of the GPU and that's mostly really been enabled by Nvidia.
27:12
They've taken their data center business from a couple hundred million dollars to a 4 billion dollar business over over 5 years doing that.
27:18
Um And the last one that we think is really interesting is actually RISC-V which which is an open source competitor to um really really to ARM I would say.
27:28
But that is much earlier in its life cycle, but it's getting a lot of traction especially Nvidia is actually acquiring ARM right now which has has its own, you know, suite of uh of issues um given ARM kind of needs to kind of feel like Switzerland um since they work well with Qualcomm or Apple or some of Nvidia's competitors.
27:42
That's kind of an interesting um proposition for it to be owned by Nvidia.
27:46
Uh but RISC-V is an open source um architecture that that's kind of just now starting to um get traction that we're watching closely.
27:55
Britain, do you want to add to that?
27:57
Yeah, well, if you think about um if you think about a CPU as like a big V8 engine, right?
28:01
It's like takes a lot of gas, but man it's fun.
28:03
It get like it gets to some places fast, you know, it's a workhorse.
28:07
And that's what Intel created and that's really what computing power is is done on the back of today.
28:14
Um GPUs are what's called a they're very good at ridiculously parallel problems.
28:19
So things like rendering images on a screen, which is a parallel problem.
28:23
I'm going to do line 1 through 1000 you know, 1080, right? Over and over.
28:28
I'm going to refresh that ever faster.
28:29
Um so if we think about sort of a you know, CPUs as a big V8, GPUs would be like lots of little electric engines, you know, or just lots of little bitty engines doing this doing uh parallel processing and then putting it back together at the end.
28:43
Turns out that's really hard to do.
28:46
Um and there were there were a bunch of companies trying to do it back in the day and when Nvidia when Jensen started Nvidia, there was around 100 competitors. Now there's only two.
28:55
Uh it's just it's just AMD and Nvidia.
28:59
That is also very hard to manufacture.
29:02
And those are both done at TSMC.
29:05
GPUs have enabled artificial intelligence.
29:08
It doesn't work without massive GPU clusters.
29:11
Are are Nvidia and AMD neck and neck in that or is one the clear winner?
29:17
Cuz it it strikes me that Nvidia could invest a lot more money into that technology than AMD, but Yeah, that's right.
29:25
Um Nvidia actually was really uh and primarily Jensen who's one of our our semiconductor founder heroes um was incredibly strategic about this uh probably starting around 15 years ago now where they they introduced a software program called CUDA, which was open. Anyone could use it.
29:46
and it allowed people to write code to GPU architectures.
29:52
Uh so, it's it's much easier to write code to, you know, one CPU that's going to process it in what's called serially, and then dump it out, dump the your work out at the end.
30:00
But, it's much more difficult to write code to take that same process, break it apart 500 times, and then have something put back together at the bottom. That's what CUDA does.
30:10
CUDA enabled uh GPUs to go into the data center.
30:13
So, for a long time, when you were when you owned Nvidia as an investor, you thought, "Okay, are they going to get the next, you know, PS4?
30:21
How big is that cycle going to be?"
30:23
It was really about gaming.
30:25
Um but, when they went into the data center, the game really changed. Keep going. Okay.
30:30
Um so, it's changed to such an extent that they are now worth multiples of Intel.
30:38
So, you think about uh you think about GPUs for a long period of time, Intel was really they sort of the largest maker.
30:45
I mean, they they they wrote the architecture for the PC, right?
30:48
And and basically, if you want to crappy graphics, you could just stay with Intel processor.
30:52
But, if you were a gamer, then you wanted that Nvidia or AMD GPU in in your PC.
30:57
Um and and it and Intel was sort of begrudgingly let, you know, let all the work interface with that.
31:04
Um but, with this shift, the GPU actually becomes more important in many cases than the CPU.
31:11
That is a massive massive change.
31:16
I'll I'll I'll I'll I'll interrupt here real quick, Brit.
31:18
But, one of the things that we're seeing in in AI that that is just like incredible is these these big AI models that get written about every few months.
31:26
Um they kind of follow their own Moore's law, where um AI model has um like different parameters or different weights, and that's kind of like roughly kind of like the synapses in the brain and that you're trying to replicate.
31:34
And the number of parameters per model in AI is doubling every three or four months right now.
31:40
And so, if you think about that over five years, like we literally it's like an 800x improvement in um computing power that you've needed to pull off to keep up with just the way these models are growing.
31:49
Um I think the most recent one literally has like 65 billion parameters.
31:53
Um that's the the GPT-3 model that came out of uh out of OpenAI.
31:57
And so, um GPUs have really been like what's facilitated that.
32:00
And so, the um the computing throughput that's needed to to um you know, maintain pace with these models is is just unbelievable.
32:10
And that's that's why Nvidia has just been so well positioned as they've been right there um with uh the computing power and the software stack to um facilitate that.
32:18
Yeah, and and it's of course, even though Moore's law is slowing down, and as John told me the other day, he's like, "Hey, look, even the company that created Moore's law can't keep pace with Moore's law, you know, Intel."
32:26
Um it it doesn't feel that way so much with AI because of of all these other things we're seeing.
32:33
So, you want to go back to something you said, which is the GPU is more important than the CPU. Can you expand on that?
32:41
it it sure, it's just taking more workloads than the CPU used to take.
32:43
I mean, look, this is early days still.
32:48
The the massive majority of the work being done in data centers is done by the CPU.
32:53
Um but, uh increasingly, it's looking that the GPU will play a more important role in the future.
32:59
I don't know, John, would you have have anything to add there?
33:01
what I'd add is just that if you think about like how you actually train, like the classic example everyone always uses is if like you want to train a model to like recognize what's a cat and what's not a cat, you just like force-feed one of these models, you know, millions of photos of cats and not cats, and you know, the um labeled data.
33:14
And um to to do that, you basically need a cluster of GPUs with only a handful of CPUs.
33:21
And that the CPU really is like kind of the uh is just operating in the background, you know, while the GPUs are really what's um where the data is kind of being force-fed, basically.
33:29
And so, that that's why I mean, at least for certain applications.
33:33
There there are plenty of applications will just that will just run on CPUs because like if you're hosting your email or something, it's just it's not that complex of a computing problem.
33:39
But, anything that really is needs intelligence added to it, I think over time, you're seeing more um more of the compute being put into GPU or or even other architectures and kind of being pulled out of the CPU.
33:50
And we we think that's kind of um you know, Jensen's ultimate goal at Nvidia, to be honest, is to um marginalize the role of the Intel CPU, which is also one of his big competitors, um and really push um you know, the GPU into the forefront.
34:02
Intel seems like a dead man walking here.
34:05
Didn't they just give up fabbing?
34:06
They seem like they're all over the place. They're a mess.
34:08
They're falling behind, like so far behind. Am I wrong?
34:12
Or is this like It's been It's been tough.
34:15
I mean, look, In- Intel is still an amazing company with lots of great engineers.
34:18
It's been tough for them.
34:20
Uh it does seem that they've lost their way in many ways.
34:22
Um they haven't given up fabbing.
34:24
They haven't really decided exactly what they're going to do, but they did talk about uh you know, outsourcing to um someone like TSMC.
34:35
Really, it could only be TSMC.
34:35
Um in the future, which after they said that, I said in a team meeting, that's something like, you know, having your evangelical pastor come into the church and saying, "We don't believe in God anymore."
34:45
I mean, I just can't imagine what that's like at Intel, um whose religion is, "We build better transistors than anyone else in the world."
34:53
So, we don't know exactly what it's going to look like, but but it has been tough times both from a fabbing perspective, as it gets increasingly difficult, and from just the way the architecture in the data center and and everywhere else is moving.
35:08
It really is a good example of just like how hard this stuff is that Britain kind of mentioned that, you know, Intel invented Moore's law.
35:14
They've been the company that really, I would say, for the last uh like the 45 years we've been talking about Moore's law, you know, for 40 of them, Intel was the leader of Moore's law, and they were the one that, you know, advanced us down Moore's law down each sequential node, you know, on this like phenomenally tight schedule.
35:28
And then, you know, around 2015 or so, they really just fell off.
35:34
And instead of taking two years to hit the next process technology node, it took them five.
35:39
Um and they're still honestly trying to get there.
35:40
And so, to to the to the question on kind of like, can how can how can China can't this?
35:44
Well, well, if Intel has been doing this forever and has basically unlimited resources, and honestly, it's an existential crisis for them to not be the leader of Moore's law, if they can't do it, like I kind of would bet against anyone else in the world being able to do it outside of maybe TSMC or Samsung because they have also been doing this, you know, for for 30 plus years.
36:01
Is this a race where you if you fall behind, you're you're basically out of the race, no matter what happens at this point?
36:07
Or can you leapfrog a generation of technology?
36:11
Because it sound it sounds to me as I'm listening to you guys, this is so complicated, it's so complex, you if you fall behind even a generation, it becomes exponentially harder to catch up and then to to push forward.
36:28
I guess we could put it this way, no one's ever done it before.
36:29
Um I I think you're right.
36:31
It It's a non-linear problem, and when you fall behind, it it it goes exponentially.
36:38
Time goes exponentially against you.
36:38
Uh to skip nodes, to sort of say, you know, I'm going to skip five, I'm going to go straight from seven to three, is almost impossible because you run into so many manufacturing problems.
36:49
And a lot of times, they're specific to the type of part you're manufacturing.
36:52
So, uh someone asked me the other day, "Well, you know, China has lots of money.
36:56
Why why can't they just like, you know, if they could technically in in a different administration get access to ASML um and and Lam and uh Applied Materials um and KLA-Tencor.
37:10
If they get get access to all these machines, why can't they just get the you know, the company engineers in there, put it together, and start building chips?
37:19
It's It's It's just so hard that when you get one plus one plus one, you could get like blue.
37:23
And you're like, "Where did blue come from?
37:25
I have no idea where blue In it takes you a year to trace where blue came from, just so you could change it to get four, right?"
37:32
Um so, to answer your question, it seems incredibly difficult.
37:37
You would have to say the odds are heavily against them.
37:41
And that certainly is their ambition.
37:42
They're They're trying to kind of pull in, you know, like I guess what their last process technology migration was five years.
37:47
They're trying to shrink that to two years, but they even kind of had to confess um over the summer that their that two-year timeline had slipped on on their 7-nanometer process.
37:54
And so, um it's it's just so hard to to kind of make that leapfrog step.
37:59
But, I mean, they're also, you know, one of the best technology companies of all time.
38:03
So, we have to give them um a a little bit of a chance of potentially closing it. We'll we'll see.
38:09
And they're incredibly strategic for the US, and the US wants to see them succeed.
38:13
Okay, so let's talk a little bit about the the the geopolitical factors here.
38:16
Well, semiconductors are are key to the information age.
38:18
They're, you know, we have a long runway where we know that they're going to be super important. They're in everything.
38:23
I think, you know, the new iPhone has like a billion transistors or something on their chip.
38:29
It like what is going on with the trade war, and what's the role of semiconductors within that?
38:37
Like, what's the role of semiconductors within the trade war?
38:38
Why is it such a strategic asset?
38:40
And, you know, in some ways, has World War III started, and it's just over like semiconductors, and we don't know it? We worry about that.
38:51
So, I guess there are a few things.
38:51
Um I mean, this really tipped off, I would say, um you know, three years ago, ZTE, which is kind of like the it's kind of like the Motorola of China, you know, they make smartphones and wireless equipment.
39:02
It's a big company, but it's not, you know, a national champion or anything.
39:06
We actually cut them off from semiconductors um because they were um you know, selling phones actually to Iran or equipment to Iran, and that violated sanctions um that the US had imposed.
39:15
And so, um we cut them off.
39:15
We put them on the on the entity list with the Department of Commerce, and cut them off from semis for a few months, and it basically like crippled the company.
39:24
You know, it's like they can't build anything.
39:26
Um and and the the issue with the way you know, the way it's the way the semi industry is composed now is a lot of the really strategic silicon is coming out of the US.
39:34
I mean, I think half of overall semiconductor revenue really is is from US companies.
39:38
And China consumes, you know, over half of semis.
39:41
Um at least they you know, they'll buy them and re-export them, but you know, most companies or most chips are actually bought by a Chinese company at some point along the value chain, but they can't design or build anything there.
39:51
And so, um so anyway, we we saw that ZTE was basically crippled with um their inability to get chips.
39:57
And then Huawei saw that and they actually realized they need to um have a plan for if they get cut off from silicon.
40:05
And sure enough, a year later in I think it was probably May of 2018 or so, um Huawei was put on the entity list and and Huawei is kind of the the Cisco and Apple of China.
40:14
You know, it's like they're the leading smartphone vendor.
40:17
They're playing a huge role in um you know, in 5G both in China and globally.
40:21
And if they can't get chips from, you know, key players like like a Qualcomm, you know, for their um for their smartphones or you know, from someone like Broadcom for their um wireless base stations and kind of enterprise equipment, they they literally like can't build this stuff.
40:37
And so, um I think the administration in the US like pretty quickly realized that chips are kind of this like ultimate bargaining chip because if we want to some of the most important companies in China and especially like
40:48
these companies that are really facilitating, you know, 5G and AI and some of the um you know, some of the most important technology trends over the next, you know, 10 15 years, that this is like a chess piece that we can use basically. And so, um that's kind of
40:58
And so, um that's kind of how, you know, the the shot across the bow um was was fired is with ZTE and then escalating it to to Huawei.
41:05
And we've really tried to clamp down on Huawei now.
41:08
We basically have said you know, Huawei you know, actually at one point was about 20% of TSMC sales and they're almost as big as Apple to TSMC.
41:16
I mean, that's just like the scale of Huawei, which is amazing to me.
41:17
Um but the US basically has told TSMC they can't even build chips for Huawei using US equipment.
41:24
And by the way, you can't build anything without US equipment.
41:26
And so, we've just totally crippled kind of what was, you know, three or four years ago, a national champion with over $100 billion in revenue.
41:33
Um and we have other levers that we can pull.
41:35
I mean, if China can't build their own chips and we're not going to give them equipment or let TSMC build it for them, then then how do they roll out 5G?
41:41
How do they keep up with AI?
41:43
Um especially in China where, you know, surveillance and using facial recognition and a lot of um you know, vision vision processing technologies are used.
41:52
Um they have to have access to leading edge technology.
41:55
And so, that that's why it's become this crazy, you know, pawn in this whole geopolitical um game that we've been playing over the last few years.
42:02
Yeah, we have no comment on the current administration, but but somebody there certainly understands how to hit leverage points.
42:09
And they hit every one of them with extreme precision.
42:11
If you were going to try to an adversary, this is the way you would do it.
42:17
You would go after these specific companies.
42:19
Um so, it's amazing uh how effective their their tactics have been.
42:26
And then the other, you know, to your kind of joking comment Shannon about World War III is, you know, most of the critical chips for everyone besides Intel that's building kind of advanced semiconductors in the US is manufacturing them at TSMC.
42:38
Um whether you're Nvidia for um you know, for AI or high performance computing or doing, you know, heavy simulation um or I mean, for chips that go into um like strike fighters and fighter jets and into defense applications or going to space and things like that.
42:53
Um that's all being built in Taiwan like right off the coast of China.
42:57
And so, that's why it's become such a geopolitical football also is is TSMC has just become so strategic in all this because we really dropped the ball on keeping manufacturing of chips in the US.
43:07
Like besides Intel and one Samsung fab, no one really builds leading edge chips leading edge chips in the US.
43:13
And so, um you know, TSMC is just so strategic in all this.
43:17
And so, that's um you know, another kind of lever in this whole geopolitical game.
43:20
In your opinion, what doesn't a country like Canada uh with its vast land and water resources and all of that take take this on um or Switzerland than sort of a neutral party, but domestic capability of manufacturing semiconductors.
43:37
Like what goes into the the policy considerations about bringing this back?
43:43
I saw something about TSM creating a factory in Arizona, I think, but it was only for 20,000.
43:48
Like it just seemed like a a political puff piece more so than reality.
43:52
How do we get that technology back into friendly uh or closer, I would say Taiwan's friendly, but uh how do we get it back closer to um sources that we consider more reliable than ones that are possibly being circled by chips.
44:12
Yeah, um it's really hard and it it takes a long time.
44:16
And this evolved over 40 years.
44:19
And it's not like you just, you know, break ground in Arizona and and start it over.
44:23
Uh there's a whole ecosystem around TSMC.
44:26
Um uh there's an alliance of people that are are work together to do this.
44:32
So, it just it can happen, but it it's we're talking What are the steps?
44:37
Like what are the steps to make that happen?
44:39
To to repatriate that technology and and then also to create organic domestic capability?
44:46
Well, it it begins with with a um you know, favorable program to bring to bring those factories back to the US, uh which it looks like we have with the CHIPS Act that has bipartisan um support to it.
44:57
And then and then over time, you begin to sort of uh you you begin to bring those capabilities back.
45:05
However, TSMC has such a a head start here.
45:08
It's such a knowledge base and it is a national um champion of a company.
45:12
They wouldn't like to bring it all to the US.
45:14
And and can we do it without their help? Um maybe, but maybe not. We don't know.
45:19
But they're a public company.
45:21
Couldn't couldn't you nationalize them?
45:22
Like couldn't the United States buy them and then repatriate that technology and then relist them?
45:28
Like I mean or is that crazy? Nationalize TSMC?
45:32
Yeah, like I it's a publicly traded company.
45:35
They could take it over, couldn't they?
45:36
That would be World War III for sure, right?
45:38
So, um so, about 70% of all chips either on the front end or the back end, meaning, you know, on the on the manufacturing side on the wafer or on the packaging side, um go through the island of Taiwan.
45:49
Uh we're pretty sure it's a sovereign nation.
45:53
China's pretty sure it's theirs. Right?
45:56
So, it's a problem for the world. It's a huge problem.
45:59
Uh it it's sort of an artifact of history after Morris got passed over from TI.
46:04
You never want to build fabs there, but they're all there.
46:08
All the collaboration is there.
46:08
To bring it back is extremely difficult.
46:10
And there's also there is I mean, to Britain's point, there's the whole ecosystem of material suppliers.
46:14
There's assembly and test, which is kind of the the back half of the manufacturing process.
46:20
And so, um I mean, there there are like, you know, hundreds of companies that are um companies you would have never heard of that are part of this ecosystem.
46:27
And so, I I think it is it is possible to to bring them over, um but just to do it efficiently and um it it's just going to take time, you know, time and a lot of capital.
46:35
But I do think Um I mean, there's interest both from like from a federal level, they have the CHIPS Act to Britain's point.
46:41
But also, I mean, TSMC I would think that TSMC's biggest customers also would like to have a fab here just so they don't have to Yeah.
46:49
just so they can sleep at night, right?
46:50
I mean, like if they can't launch the next iPhone because TSMC is, you know, having issues with with with Taiwan sovereignty or whatever it is, then that that is a big problem.
46:57
And so, um I think that's that's another kind of vector is you should get push from TSMC's big customers over time.
47:03
It's interesting to me that like COVID happened and we're like, "Oh my god, we don't have domestic mask capability uh to do this."
47:10
But nobody's thinking about domestic semiconductor capability.
47:12
And so, why isn't it like Intel, Apple uh why aren't these guys combining together making their own fabs?
47:19
Even though they're they're all using different architecture.
47:22
They're all using sort of like their own engineers to design the chips.
47:25
Why aren't they making a TSMC competitor?
47:28
We've wondered the same thing.
47:28
I mean, one reason is because just TSMC is so wonderful.
47:33
I mean, they're phenomenal.
47:35
They're the closest partner in many cases. They love them.
47:36
They they're enabling their products and they're a dream to work with from everyone we've ever talked to about TSMC.
47:43
Um so, they don't really see a problem, but geopolitically is it a problem? Absolutely.
47:49
So, we've asked this question, too.
47:51
I don't know really the answer.
47:53
John, do you have a good speculation?
47:54
Well, I I do think it's it's actually a really good question cuz I think that's definitely their ambition.
47:58
And there's actually there's an open level open letter from Bob Swan, Intel CEO, that they had ambitions to um you know, be more of a player in in domestic semiconductor manufacturing.
48:09
But the problem is is that they've never done this this model well of being a foundry where you're I mean, Intel forever has just built Intel chips, you know, and on the very very much on the margin they've tried to build you know, kind of be a third party manufacturer for other companies.
48:20
But they have just never been able to do it well.
48:22
And so, I mean, the the most visible example is they actually um worked with this company Altera, who they then acquired um you know, in 2016, but they could never get the manufacturing right.
48:31
And that actually ended up being a really difficult acquisition for them.
48:33
And then the other thing is is Intel's technology is actually inferior to TSMC's now.
48:37
So, if you're Apple and you've been working with TSMC forever and they've got the leading edge technology pretty hard move to to go back to to Intel.
48:44
And so, I think maybe if it's something that is really of national um you know, strategic importance in terms of something for aerospace and defense or something that really just can't have to be built in the US, then you can make that argument.
48:56
But um it's just it's so hard to replicate the model for for Intel.
48:59
Even though I I if I were Bob Swan, that's what I would be trying to do.
49:03
I think it's such a good question.
49:04
And the Taiwan government is a major stakeholder in TSMC.
49:06
So, yes, they're public, but it's not not quite as easy as just buying them and moving them.
49:11
So, so who is the guy that did the TSMC?
49:13
Like they funded it and they brought it to Taiwan.
49:15
They said like create all this stuff. Who was that? What was his name? Morris Chang. Morris Chang.
49:20
So, who's the Morris Chang that you could do that to in the United States or Canada or Western Europe or like there's got to be somebody who's on par with that where you could be like, "Here's $100 billion, $200 billion. Like let's let's go.
49:32
It's going to be a 20-year project, but we got to do it."
49:36
We probably have a few that come to mind.
49:38
Um Morris turns out was just incredibly special and Oh, of course. Yeah.
49:43
in history, but yeah, there's a few that come to mind.
49:45
Um I would say say the Canadians who listen to this.
49:48
Like we need to find somebody to fund.
49:50
That is exactly what China is trying to do.
49:52
Like they have hired some um you know, senior like operational people out of TSMC, you know, and and thrown you know, crazy numbers at them and probably said, you know, this is a 20-year march and you've got I mean, China has raised these funds, you know, multiple that are tens of billions of dollars.
50:06
And so they they could put 50 billion behind them.
50:09
Actually, I don't know them by name in terms of who kind of the up-and-comers were at TSMC that they've been able to um pluck out, but it is like all of the I would say that the the very best um people to build a business like that um in in China or the US or Canada would be um the manufacturing folks from from TSMC.
50:26
And then the then the only other people I'm guessing this is what we're Britain might have is is just to take like a visionary that knows the ecosystem really well.
50:32
Like um Lip-Bu Tan as the CEO of Cadence Design Systems who's been one of the leading VCs in semiconductors actually for for decades and runs like one of the most important companies in the whole ecosystem and he's on the board of Samsung. He knows everyone.
50:47
Um and so someone like that I think that really um probably would get the playbook down and would at least know who to call um would be um who who I would call.
50:55
And I think Britain might have mentioned Rich Templeton also who's the CEO of Texas Instruments and is who we think one of the all-time great leaders in semiconductors um would be the other one that would would have a a shot at it because TI actually still runs their own manufacturing.
51:10
And so he gets manufacturing.
51:12
I mean, it's different because it's it's a lot older and isn't like truly bleeding edge, but um I think he understands the model and at least how you would go about doing it.
51:19
Talk to me a little bit about TI.
51:21
I mean, I keep hearing their name crop up over and over and not to mention like it seems like they're really good at investing shareholder capital. Yeah, they're amazing.
51:30
Um you know, there's really two ends of what we're talking about.
51:33
We've got the leading edge companies, companies that are developing the most advanced chips in the world.
51:37
That's sort of the Intel, Nvidias of the world.
51:39
And then you've got um the catalog companies which are sort of the Legos that go around that leading edge chip that make it all work seamlessly with products.
51:51
And those catalog companies may have tens of thousands of parts.
51:55
TI is what we call a catalog company.
51:57
They've got lots of parts.
51:57
Uh they take a long time to develop, but once you develop them, they turns out they have 30 or 40 year shelf lives.
52:06
Um so TI did develop leading edge parts for a long time.
52:10
And like John said, they manufactured it themselves.
52:13
Turns out that business that they were in, mainly making cell phone chips, um got really competitive and the margins were rooted out of it.
52:20
And so Rich uh Rich Templeton, the CEO of TI, made this strategic decision to get out of that business.
52:29
Turns out it it was you know, the majority of the revenues when they decided to get out of it.
52:32
Um and uh it they couldn't sell it.
52:36
They tried, but no one wanted to buy it.
52:37
They actually bled it down to zero.
52:39
Um and then they doubled down on the catalog business.
52:42
So to your point about being very good with shareholder capital, these were massive capital decisions and they were uh incredibly good for for TI shareholders.
52:53
So so the just add to that, so the beauty of what they do, um they're the biggest analog chip company in the world.
53:00
Like if you think about like there's basically two different kinds of chips to Britain's point.
53:03
There's digital which is like zeros and zeros and ones, you know, big you know, thinking in binary, big um you know, heavy digital chips from Intel or Nvidia and those can cost, you know, a few thousand dollars.
53:13
And then there's analog chips and analog really like communicates with the outside world.
53:16
So analog chips can um you know, measure signals or sound or vibration um or so it's it's not just zeros and ones.
53:23
It's it's really um interpreting, you know, the physical world.
53:27
And the the beauty of of analog is that actually the average um cost or the average selling price of an analog chip is around 30 cents.
53:35
And so it's like these guys sell like billions and billions and billions of you know, between whatever 10 cents and a few dollars worth of um silicon to um you know, to tens of thousands of companies.
53:47
And so they really play well in like the industrial market which is really like 14 different kind of submarkets.
53:52
Like that would be you know, heavy equipment or factory automation or building automation or medical devices.
53:58
Um and then the automotive market.
53:58
I mean, cars have tons of analog just because there's a lot of like power management that has to be done.
54:03
There's a lot of different signals that have to be processed.
54:05
So TI has really doubled down on these markets.
54:08
They're going to change really slowly over the next um you know, like 10 to 20 years, but they're they're just like playing for duration.
54:14
And and analog semis to to to what Britain mentioned also is you know, you you kind of win a part in a in a car model and that's really not going to change the next like 10 to 20 years.
54:22
Um and it can go even longer if you win it in like a John Deere tractor or like a you know, a Caterpillar piece of heavy equipment.
54:29
And so um it's almost like a software model where once you're in, it's like a recurring business for um you know, for 20 years.
54:35
And that's why I mean, TI actually has some of the best margins in the S&P 500 and has been a phenomenal cash generator and a and a great stock because they've um they've built this business that is just like built on durability and being in these long markets that um you know, change very slowly, but um they can just like monetize and especially as silicon kind of pushes deeper into all of them over time, um that's just a great place to be.
54:58
Are there a lot of natural competitors to TI in that space?
55:05
Yeah, so there there are there are a few.
55:06
So the the analog market is still really fragmented.
55:07
Um Analog Devices is kind of their natural competitor and they're they're actually just announced, you know, two months ago or so that they wanted to acquire um the number three player in the space who's Maxim Integrated.
55:19
And so um the beauty of it is I think the combined, you know, TI plus um you know, the new Analog Devices after they acquire Maxim will probably be about a third of the analog market.
55:29
And so even though you kind of have these like two 800-lb gorillas, there's still just like so much market share to go after.
55:34
Um and I think I do think it's one of these markets where the big will get bigger over time um both because you just kind of have more breadth of product um but it's also like TI's reach is unmatched.
55:44
They actually go direct to their customers now instead of going through distribution.
55:47
Um you know, they have tens of thousands of parts.
55:49
They can release new more new products every year than their a lot of their competitors um probably half the market could combined, you know.
55:56
So there there are natural competitors and all the companies have, you know, amazing margins and aren't the beauty of like a 30 cent part is you're not going to go in there and offer 25 cents to like get the business, you know, with and the customer customer doesn't even really care about that that much because it's probably a electronic system that has hundreds of dollars of content in it.
56:11
And so um they're all competitors, but it's not really like one of these like really price driven markets that you could might see in the you know, in the smartphone market or something like that.
56:19
And it takes forever to introduce a relevant catalog of parts.
56:22
I mean, TI with all their engineers may may introduce 5%, you know, new parts per year.
56:28
So it's 20 years to to get that the catalog.
56:31
So it's just one of those businesses you can't come into and make a better, you know, widget and then gain a lot of share.
56:37
It doesn't work that way.
56:39
I want to try something never really done before, but never really done something with three people on a podcast.
56:45
So like uh I want to take the view of okay, if you're John, you take the view of the United States.
56:53
And I want to hear what you would do what you would do to like what is this strategy here?
56:58
We want to bring back domestic capability. Maybe it's not.
57:01
And Britain, I want you to take China and you play China.
57:03
And I want to hear you guys just riff on like what do you see as and these are opinions obviously, but like give us your opinions on bit of chess.
57:13
Yeah, a little bit of chess here.
57:15
Like give us some back and forth on uh on things.
57:20
I think I I probably have the easier one because um I think the US is in just such a better starting position, but um if if I were um you know, designing our semiconductor approach for the next decade in the US, I would go to all the major designers um to Apple and Qualcomm and Nvidia and Amazon and Google.
57:40
And by the way, this is like, you know, trillions of dollars of market cap and of economic profit that that you could walk up to and and just say we need to have a consortium to support domestic manufacturing and the government will support this.
57:52
But this really it's like we have the whole ecosystem here.
57:54
We own the the um equipment.
57:55
We've got um all the design guys here.
57:58
We've got the software here, but we don't have is manufacturing.
58:02
And so, you know, we probably need to put together a hundred billion dollars um to spend over the next, you know, decade to really bring manufacturing here like strategically.
58:11
And so that um honestly, I think that's something that that, you know, if I were Apple or Nvidia or anyone else, I would sign up for because it's so strategic and they've got um the capital to do it, but that's that's where I would start.
58:22
So you'd limit technological uh use in China to any like you would cut off all US products, all US like So that's a great question.
58:31
What I would do is um there are technologies in China that they've developed that are totally legitimate.
58:37
And they they've acquired IP that is legitimate.
58:38
I mean, they're they're entering entering into the NAND flash memory market and that's um technology they've built domestically and also acquired like totally, you know, above ground.
58:47
And so I just think um for as in terms of like how we're perceived on the global stage, I don't really think that we can just ban um shipments of equipment to companies that actually have legitimate IP.
58:56
But I would not support and the US hasn't done this.
59:00
Um you know, but I would not support anything where the IP is in question.
59:04
And so I think that I think an outright ban is probably the the not the right way to go.
59:07
It's just making sure that I mean, the beauty of the chip business is there are only so many prop big projects.
59:12
You kind of know who they all are and where the IP came from.
59:13
It's like there's, you know, three or four memory projects in China and two foundry projects.
59:17
And um so you kind of know where it came from and if it's legitimate or not.
59:21
And so if it's not legitimate, then you can cut them off.
59:24
And we we've actually done that.
59:24
There was a there was a DRAM project that um the IP the IP's origin came into question.
59:29
This was about two years ago now.
59:31
And we stopped selling them them equipment and everyone just walked off the project.
59:34
It was like a $5 billion project and it's now it's just like an empty factory sitting in China which is just amazing.
59:40
But let's say let's say John did cut the equipment off to me, you know, China and and I wanted to get that equipment back.
59:48
I think I've two roads open to me.
59:48
One is of course the piece of equipment I want most is made by company in the Netherlands.
59:54
I can try to play nice with them and sway them away from following US policy.
1:00:01
So there's that diplomatic road open to me and then there's a much more immediate road which is I just say okay Shenzhen used to be a special economic zone.
1:00:09
I'm going to change the law and I'm going to going to put locks on all of Apple's factories.
1:00:15
Every iPhone is made in China.
1:00:17
I could stop that tomorrow.
1:00:17
There would be no more shipments coming out of China.
1:00:22
So that's my first move if I'm if I'm really getting crazy, right?
1:00:25
And then the second move is I move to take Taiwan either uh through diplomacy which doesn't seem to be working at all because the pro-Taiwan government got ousted in the last election or by force.
1:00:40
The third move is I could say well I make precursors to a lot of your drugs.
1:00:46
You can't make those drugs without what I make in my country.
1:00:48
Those are now no longer available to you. Right?
1:00:53
So we as as China when I say we, the Chinese government because that's what I am.
1:01:00
We have leverage points like you do.
1:01:03
Maybe not high technology leverage points but in the supply chain we have massive leverage points too.
1:01:09
And we can cut those off and that can cause you a lot of pain in the short term.
1:01:14
I can't imagine how Apple how Tim and Tim Cook is sleeping at night knowing all of his factories are in China and we're kicking an anthill there.
1:01:25
John, how would you respond to that?
1:01:27
Yeah, I mean I think this basically is how it turns into a big game of chicken, right?
1:01:30
I mean it really is like who can you know keep a straight face before we you know clash into each other.
1:01:36
But um if I were like if I were Tim Cook, what I'd be thinking about is I do think that Apple is like one of the few examples of China allowing a foreign company to come in and flourish and I do think like on a 20-year horizon China wants to have be able to hold examples like that up to the world.
1:01:52
I don't think they want to be a completely closed off economy and I think that can change.
1:01:55
But I think that is kind of you know if you're playing chicken, I would say okay go ahead and shut that down and by the way you know I've got some manufacturing in Vietnam and it might I might miss an iPhone cycle but you know in 2 years I could probably spin up 80% of my manufacturing capacity in Vietnam and the rest in Mexico.
1:02:12
And so I do think there are other parts where a little bit more pain into a corner and I'm like honestly like rare earth material rare yeah rare earth material is another one.
1:02:18
But um that's that's kind of how I would take it from the from Tim Cook's perspective.
1:02:26
Any final thoughts Brent?
1:02:26
You know it gets really ugly really fast and really it comes down to what do you want to do with Taiwan?
1:02:32
Uh uh you know of course if China does move to take Taiwan, they they would have TSMC but we still wouldn't we wouldn't it's not like we'd be shipping equipment over there anymore.
1:02:44
So they still have some somewhat of the same problems.
1:02:47
Um So you know could they take that through diplomatic means? I don't know.
1:02:52
But it it just gets really ugly really fast.
1:02:56
And and for us it's it's sort of crazy because China's a country we've been going to for 20 years.
1:02:59
I've never thought of them in these terms.
1:03:01
I've always thought of them more as a partner. Do they do dodgy things? Absolutely.
1:03:08
But does everybody sort of sign up for it to get access to their markets? Absolutely.
1:03:11
They know what's going on but they want access to the Chinese consumer.
1:03:14
So they'll they they allow it.
1:03:18
But the sort of the head that we've seen this come to in the past 4 years is not something I expected.
1:03:23
What what are the that was really fun. I liked that.
1:03:25
Thank you guys for indulging me in this little back and forth here and hypothetical.
1:03:29
But what what are the the sort of like misconstrained rare earth materials that go into the manufacturing process?
1:03:37
I don't know the answer to that. John, do you?
1:03:39
No, I'm I'm I'm not an expert I should know that the number the like percentage of rare earth materials that go into you know some key technologies including smartphones is like 90% comes out of China.
1:03:50
You know and I do think like in terms of the the list of their the levers they can pull, you know that one's up there.
1:03:55
You know I think we just have to be mindful of what I think we've got to focus a lot on what we have that's strategic that they really need but they they have some of that stuff too.
1:04:03
I I mean it's great when when we can sort of like have this you know no one can sort of pull the cord on anybody else cuz we're symbiotic and we're mutually dependent.
1:04:12
But if you don't have that mutual dependence, all of these these sort of like globalization actually breaks down really quickly in uh escalation or hostility, right?
1:04:23
If there's no mutual dependence or no mutually assured destruction um then there there's all of these other considerations that go into it and I think maybe we we we are at that level of mutually assured destruction with a weapons here just from a trade perspective.
1:04:39
But when you start looking at all the things that are sort of outsourced that are critical infrastructure, you know, masks, antibiotics, semiconductors and you start looking at things through a different lens.
1:04:50
It becomes a lot more apparent that some of these things you want to have domestic or at least guaranteed friendly allies who are controlling it, right?
1:05:01
Like there's a big difference between China doing antibiotics and Canada doing antibiotics from the US point of view, right?
1:05:07
The same is probably Western Europe and I think that that's just a really interesting and fascinating lens to look at things through. It's it's eye-opening.
1:05:16
It's very eye-opening and I have to believe that the rest of the world is watching.
1:05:19
You know, our ability to to really set you know one of the largest countries in the world you know one of the most powerful countries in the world back uh 10 or 15 years is is should be shocking. Shocking for everyone.
1:05:38
Um that that one one set of companies one handful of companies really holds the power to do that.
1:05:45
And then part of the problem is it just worked so well for so long.
1:05:46
Like the the equation for a company like Qualcomm or Nvidia that they don't have to build factories and put you know an advanced semiconductor fab cost you know tens of billions of dollars.
1:05:56
So be able to outsource that and have it conveniently made in in Taiwan, you know, I I think that just worked so well until we honestly like got to this point and everyone looks around and you know it's like less than 15% of semiconductors are built in the US now.
1:06:10
And so I think that's why we have this obviously with all the geopolitical tensions also you just had this realization and that's why I mean the industry has really made this hard push to to lobby for with the Chips Act for for more government support to to try to bring incentives to to build you know more semiconductors here in the US.
1:06:27
But like John said, it's a whole ecosystem.
1:06:28
I mean the way it works now is really simple.
1:06:30
It's like you you know you you do that design, right?
1:06:31
You send it over um electronically. TSMC fabs it.
1:06:36
They ship that chip to Shenzhen.
1:06:38
Foxconn puts it together.
1:06:40
There's a whole host of other equipment makers that are in that same area Longhua which puts it together in in the Apple fab if you're talking about an iPhone for example or an HP computer or whatever.
1:06:51
And it and then you ship it straight to the US and the the company actually never touches it, right?
1:06:56
So it works really well but to bring that back to the US is just it takes a whole ecosystem and that those ecosystems move in decades.
1:07:07
This was a big hairy topic that we dove into today over the past you know 75 minutes.
1:07:11
Is there anything we missed that you guys want to cover?
1:07:16
I think one of the things that John has said repeatedly is interesting.
1:07:18
I mean semiconductors of course are fascinating in and of themselves and of course the geopolitics and how we got to a handful of companies that really control the fate of the world.
1:07:28
That that's all interesting.
1:07:29
But for a long time it was really dependent on what came next.
1:07:31
You know, was it a was it a the PCs were growing, you know, did the PCs need more DRAM or something like that or then the of course the cell phone market came up which was a massive market that drives semiconductors.
1:07:44
Today it's really different.
1:07:44
It's not like we're looking to the cell phone market.
1:07:48
Um John has this great line that says you know the next big thing is there's no next big thing.
1:07:55
Everything's the next big thing.
1:07:56
And and John I should just let you say it.
1:08:01
Yeah, I mean I I think that it's it's there is no next big thing basically in the industry and the the reason that we maybe going back to the beginning of the conversation about really the history of the industry.
1:08:08
It's always been kind of driven by like one big application.
1:08:11
It was like mainframes and then desktops and then you know laptops and smartphones and that's like 40 years of computing basically.
1:08:16
And now it's you know the future of the car.
1:08:20
It's the future of the data center.
1:08:22
It's industrial equipment having intelligence. It's medical devices.
1:08:24
I mean it really is I think for the first time there's not at least to us an obvious kind of like consumer business application that's just going to like be the next exciting application.
1:08:34
And so I think just like for the durability of the industry, it's great having a lot of exciting growth drivers at once and it's brought a lot of you know capital and excitement to the industry for the first time in a long time.
1:08:46
Like there there actually startups in semiconductors for the first time in forever and you're seeing you know huge companies like Amazon and Google getting into the market for the for the first time.
1:08:53
And so um it just shows just that there's like a lot of it's kind of this renaissance that honestly wouldn't have anticipated 10 years ago which is really exciting to follow along. Excellent.
1:09:04
Thank you guys so much today.
1:09:06
This was an awesome conversation and I really enjoyed it. Thanks so much Shane. It was great.
1:09:12
Thanks for having us Shane.