This is Matt Russell, and today we have a replay of our 2023 episode on ASML.
Our guest was Tom Walsh from Bailey Gifford, and I will often share this episode with the.
Tom takes the complexity of extreme ultraviolet lithography and makes it incredibly digestible for anyone who is in the investing space and really brings to life what has happened and the evolutions in this technology, to put it in the place that it's in today.
Now, since the episode, EUV has only cemented its place as being core to AI infrastructure.
And if we just take a look at some of the numbers that Tom shared, the 21 billion euros in net sales in 2022 has grown about 50 to 32 billion in 2025.
Gross margins have remained steadily north of 50%.
And as Tom alludes to, they were in the 30s not that long ago.
So the story here has played out in a very interesting way.
And amidst all the volatility, the stock returns have proven to be strong as well.
So this episode is timeless for a lot of what you'll learn about the business and the technology, but also interesting in this moment in time to understand how ASML fits into technology.
DAI landscape.
This is business breakdowns.
Business breakdowns is a series of conversations with investors and operators diving deep into a single business.
For each business.
We explore its history, its business model, its competitive advantages and what makes it tick.
And we are here to bring them to you.
This is Matt Russell, and today we are back covering the semiconductor value chain.
ASML was once a forgotten subsidiary of Philips.
Today, it's one of the most important technology companies in the world.
To break down ASML, I'm joined by Tom Walsh, a portfolio manager at Bailey Gifford.
Tom helps explain what exactly is happening inside an extreme ultraviolet lithography machine and how ASML came to pioneer this technology from the Netherlands.
It was a non-traditional path to say the least.
Now this breakdown pairs very well with our breakdowns on AMD, Qualcomm, and Cadence.
And I'd also highlight the Founders Podcast episode number eight on the Intel Trinity.
Please enjoy this breakdown of ASML.
All right, Tom, thanks for joining us on Business Breakdowns here.
I'm excited to get into ASML.
We've covered some other names in the semiconductor value chain.
But ASML, for my research, has a really interesting story in terms of how they become the giant that they are today.
Maybe you can just start us off with a little bit of the backstory and how they came through as this scrappy outsider in this major industry.
I think one of the really interesting things about ASML is how it came to be so dominant in its industry.
And often you think of the dominant players having sort of started with some massive competitive advantage.
And that just isn't the case with ASML.
It started back in 1984 when it was spun out by Philips.
And some spin outs are sort of destined to be great companies.
They're clearly ready to fly the coop and take on companies like PayPal as they emerged from eBay.
They clearly got a strong position and ready to capitalize on that.
And ASML absolutely wasn't that thing.
It was problem child of Philips.
They'd been battering away for over a decade trying to get into the photolithography business, and they just hadn't succeeded.
In fact, the joke within Philips when it was spun out was that ASML being created was essentially a sort of delayed layoff process for the workers that were being transferred into it.
So it emerged into the industry as number 10 of 10 lithography players in the early 1980s.
It had no revenue whatsoever.
It had no commercially credible product.
It had no offices.
In fact, its first offices were some wooden barracks that they built in the grounds of the Phillips facility in Eindhoven.
And a good proportion of the workers didn't want to be there.
They'd been shunted there out of Phillips and they thought the business was going to die and they just weren't interested in being involved.
So these aren't really heartening beginnings for any company.
But what they did have were, I think, two quite specific but industry-leading technologies that had been developed within Philips and could be incorporated into future products.
They had a handful of really tenacious and quite brilliant engineers that were determined to make this work.
And they had a window of opportunity within the lithography industry, which was going through a big technology transition, around that time.
And the CEO who came in to lead the company, as it was spun out, could see this opportunity in the technology transition for ASML to jump ahead, because none of the incumbents actually had fully prepared for that transition.
Now, the early years of ASML were unbelievably challenging, and it was really the first decade just about survival for the company.
They survived by scraping through some savage industry cycles during the 1980s.
They were kept alive by further cash injections by their backers, Philips and ASM.
But ultimately, they started to make progress.
Their competition started to falter and fall away.
Their equipment started to get better and kept getting better.
And by the 1990s, by the mid-90s, they'd emerged as one of three really clear leading players within the lithography industry, the leaders being Nikon, Canon and ASML.
And they kept pushing.
In 2002, they finally overtook Nikon as industry number one.
That's a goal they'd set themselves on SpinOut.
It took them nearly 20 years to get there, but finally they overtook them.
But they didn't rest on their laurels at that point.
They got there... foot on the accelerator even harder.
They kept delivering market leading innovation and investing in moonshot technologies that many believe would never actually work.
And ultimately that led in 2019 to the first use in high volume production of extreme ultraviolet lithography equipment, the most advanced lithography equipment available in the market and the only equipment able to make the most leading edge semiconductors in the world.
And that's put ASML in this position where it has essentially 100 of the world's most advanced lithography equipment and a pretty dominant position, 90 plus of the next generation of lithography equipment as well.
Can you try to explain in the simplest terms possible lithography equipment, what photolithography is just?
However, you can break that down in layman's terms.
If you take it all the way back to semiconductors, what are semiconductors?
The way they work is basically having a series of really, really tiny electrical circuits embedded on the surface of a small piece of silicon.
And the question is how do you get those electrical circuits in?
Bearing in mind, these are microscopically small.
They're so small, they make the width of a human hair look giant.
And back in the beginning, they would do this with effectively having to touch the surface of the semiconductor.
But that had its own problems.
It had real problems in terms of defects that were introduced to the surface.
In the 1970s, they finally struck upon the idea of essentially using light projected through a mask.
So essentially projecting light to create a pattern on the surface of the silicon chip.
And that's done by essentially coating the silicon chip with special chemicals, something called photoresists, and then projecting the pattern of electrical circuits onto that, such that you can then embed those circuits onto the surface of the chip.
That in itself sounds relatively simple.
In concept, it is really simple.
In practicality it's unbelievably difficult because the dimensions that they're doing this at are so small.
They're measured in handfuls of atoms rather than in millimeters or centimeters.
So in concept, very simple.
It's much the same as if you go to the old-fashioned cinema projector.
You shine light through an image, use a lens to focus it and you get an image projected onto a screen.
The difference between the old-fashioned cinema projector and a photolithography machine is the image is obviously not Spider-Man.
This is the circuits of a semiconductor, and the lenses are being used to concentrate and make that image even smaller than it starts out, rather than blow it up big onto a screen.
And I guess this explains when you mentioned competitors or other peers in the space Nikon and Canon, I think of cameras and associated lenses with those businesses.
Was there a camera angle to this at the early stages, or was it always with the end goal being related to the semiconductor space?
So the first person to think of using photolithography was actually back in the 1940s and it was an American company. who conned onto the idea that if you used a microscope, you could make things look large.
If you turned the microscope upside down and projected light through it, you could take something quite large and make it much, much smaller on the surface of another object.
So that idea had been kicking around for a while.
And the first companies to really master projection photolithography ASML does today were actually American companies.
And they emerged in the 1970s and they built very successful businesses on the back of that over the next decade.
But the Japanese saw what was happening and thought this was an interesting business to get into.
They could see that the semiconductor industry was going somewhere.
And obviously they started from a position with expertise in-house in the manufacture of lenses, which was one of the key challenges of photolithography.
And they were able to come through, starting initially really essentially by copying American machines that were on the market, but rapidly iterating to the point that their machines were simply better.
And that's how they managed to end up dominating the industry.
But in terms of, did they start in photos with a view of getting to semiconductors?
I suppose not, but they could see some transferable skills.
They saw some transferable IP and they really made the most of that.
Yeah, it's a really interesting area or product to be borrowing from with a completely different use case.
You alluded to it in terms of them being the leader in the space today.
Can you give a sense of how big they actually are?
Whatever measurement you want to use, whether it's revenue or something else,
In 2022, they generated revenues of about 21 billion euros and operating profits of about 65 billion euros.
And they've got a market capitalization of, as of today, just over 250 billion euros as well.
So they're a fairly large company in terms of revenues and in terms of market capitalization.
But when you actually look at the company, it's not actually selling many machines.
It only sold 345 machines. photolithography machines last year.
It's a mass produced item, but it's a very low volume, very, very specialized item.
The most expensive piece of equipment sells for north of 150 million euros a piece.
Yeah, that's sticker shock.
I was surprised when I was doing some research to see the number of units sold and I was scratching my head in terms of adding up that versus the revenue number, but that explains it well.
You mentioned they were somewhat of the orphaned company, the good co, bad co spin out.
They were the bad co in this particular instance.
What did the management team look like Then?
Versus today, have there been major visionaries in terms of their history who really drove this business to go from being the forgotten child to the leader in the market.
Right back at the very beginning, someone was actually an individual was brought in to lead the companies.
It was a spun out of ASML.
He stuck around for a few years and then there were a series of different CEOs over the history of the company.
The current CEO, he's been there since 1999.
He was CFO for a number of years before stepping up as CEO in 2013.
And he certainly got to take a lot of the credit for the success of the business in terms of steering the business through some pretty challenging times.
If you think about how tough things were in the semi-industry in the aftermath of the tech bubble in 2000.
He was the CFO there.
He had to keep the ship afloat.
And he ensured that the money kept going into R&D so the company kept extending its advantage.
As he stepped up to CEO in 2013.
It was by no means a given that EUV their leading edge products that are now the dominant part of their revenue base and their future revenue base.
It was no given that that would happen, but he was able to steer the company through that.
Again, keep it focused.
Not get distracted and try and branch out and diversify into other activities, but back as engineers to eventually get that to work.
So he's been key.
But if there were one other person who really is associated with the success of ASML, then that would happen to be a man called Martin Vandenbrink.
And he's the CTO and he's the president of ASML.
And he's been there from the very, very beginning.
So he actually applied to join Philips from 1983, just towards the end of 1983.
But by the time he arrived at Philips, he arrived into the ASML division, as was right at the time that it was essentially being spun off.
And the funny thing is, when he joined the story goes that the old hands at Philips who were working in this semiconductor part of the business were sort of chuckling to themselves that he was joining a part of the business that was being sent off to die.
And they were staying behind to continue to develop the e-beam lithography division that was going to be the future of semiconductor manufacturing.
Nonetheless, he stuck with it.
And he had a pretty trying time in his first couple of years.
And the company lurched towards the brink of disaster as it tried to bring its first commercial product to market.
They were struggling.
And actually after two years, despite coming out of education into ASML as his first job, just two years into that job his boss was moved off.
And he was moved to head up the development of the company's make or break product.
So there's a guy two years into his time working in the orthography industry.
And he steps up to run that.
Now, at the time, it was supposed to be a stopgap measure, as I understand it.
They wanted to bring someone in the head.
They just needed to get this thing of the line they needed there.
But they never found someone that was better than him.
He brought that product to market.
That product was a success.
The company went from strength to strength.
And in 95, he stepped up onto the management board and was appointed CTO.
And he remains there today.
And he's close to a genius as you'll meet if you ever get a chance to meet with him.
And he's been absolutely instrumental in the success of not just ASML but in the continued progression of Moore's law and semiconductor technology the entire industry.
We've talked about this on past episodes, but can you bring in how photolithography plays into Moore's Law, how they've been able to shrink things and how key this component is to that overall process of being able to shrink the chips?
So in very simple terms, the key gating technology is for the advance of Moore's law.
So the idea of making transistors smaller and cramming more of them onto the surface of the chip, doing that essentially doubling it every two years.
The key gating technology since 1970s has been photolithography.
So whenever photolithography is stalled, Moore's law is stalled.
It happened in the late 1970s and to a degree it happened about 10 years ago, when it was struggling to get extreme ultraviolet technology to work.
So photolithography is absolutely vital.
And what does a machine actually look like?
What's the size?
If you could just give us some type of picture of it.
So these things are huge.
The current leading edge machine, the extreme ultraviolet machine that, when constructed, is about the size of a double-decker bus.
You can, if you're lucky enough to get the chance, to go and visit the clean room in Veldhoven, which is where the company's headquartered, and you walk around the clean room and see them in manufacturing process.
These are huge, huge machines.
And they're built in component parts.
They're built in modules that ultimately are assembled at the factory of the semiconductor manufacturer themselves.
But they are really vast bits of kit.
And when they transport them to the end foundry, the semiconductor fabricating facility, it takes three jumbo jets to get them there.
So these are really extraordinary machines.
And they're going to get bigger.
The next generation of machines that comes out in hoping to go into high volume manufacturing in 2025 26.
They'll be bigger still because perversely, as you make the pattern smaller, you have to make the lenses bigger.
And that means the machine has to get bigger.
So smaller transistors, even bigger machine.
Yeah, that's a funny juxtaposition there, in terms of one thing getting smaller and the other getting quite bigger.
You mentioned...
EUV that, having a big role 10 years ago, it seems like a technology that's really owned by ASML today.
What's been the evolution there?
And how has ASML become the dominant player with that particular technology?
Depends on why you want to go back with Evie, because I could bore you for hours on it.
But the essence of why extreme ultraviolet was necessary is when you're shining light waves through a pattern onto the surface of a chip.
When they started off doing this, they could do it with just natural light.
And that worked because the size of the transistors was relatively large.
Even in the early 1970s, you were cramming nearly 200 transistors onto the surface of a pinhead.
But it's still large enough that it didn't matter so much that light travels in waves.
It didn't travel in a straight line.
I have to dredge my memory for high school physics, but light travels in waves.
And as the patterns you're trying to draw get smaller clearly, the fact that light goes through a pattern then continues on and a wave becomes a problem, because what is presented on the surface to silicon is not the same as what is actually created on the mask or the pattern you're trying to project in the first place.
So over multiple decades, the industry moved through different types of light source.
So you have two elements really to a photolithography machine.
You've got the light source and you've got the lens.
The lenses keep getting better, the light source keeps getting better.
And they moved from visible light to ultraviolet light to in the 1990s and 2000s they were working with what they called deep ultraviolet light.
But that deep ultraviolet light was still pretty challenging to make.
By that point, it's not a light bulb, it's a laser that you're using to create these patterns.
But that had a wavelength of 193 nanometers.
And the leading edge chips.
Today, the dimensions are measured in single figures of nanometers, so the leading edge node is five nanometers today, so it's a bit like trying to write your signature using a snow shovel or something.
You can do it, but it's not very easy and there's a high chance of mistakes.
So many years ago, back in the 90s, they recognized this was a major problem that we needed a next generation fix for this And, after much deliberation within the industry, extreme ultraviolet was struck upon as the solution.
The interesting thing is, when it started back in the nineties, when this decision was made, ASML was not at the forefront of it.
It hadn't been focusing on EEV, it had been focusing on survival.
There was a bit of a stroke of luck.
The research into EEV was developed by the Americans.
They'd seen their semiconductor industry decimated by the Japanese.
They developed through state funding.
Department of Energy and DARPA put a lot of money into this, in a consortium with industry partners from across the US, to fund the development of the technology.
But they didn't want to give it to the Japanese and they didn't have a domestic lithography maker that was going to be capable of taking it forward.
So they invited ASML to join the consortium and take that technology on.
And that was a really massive investment step, i think, in the history of asml and of the industry as a whole, and at that time the plan was to introduce it in the mid 2000s.
They thought they could get this up and running.
They thought they would have to get it up and running to be making semiconductors and continuing moore's law from 2004, 2006 onwards.
Obviously it was late.
By 2010 it still wasn't there.
2012
ASML managed to persuade Intel, Samsung, TSMC to support them, to co-invest in the company.
It took 23% share between them and ASML.
We've got 1.4 billion euros of R&D funding into ASML.
And finally, in 2019.
It will take 13, 14 years after it was supposed to be delivered, these first EUV machines came out.
And it's all just about being able to make the patterns smaller, make them more efficiently and therefore make them at lower cost, but also to make them better.
They're high quality patterns because you're not having to use lots of tricks to try and get a really wide bandwidth of light down to a very, very tiny design size.
That's incredible in terms of the storyline for how they were trusted with developing this technology.
Did the Americans put all their chips in the ASML basket?
Were there any other businesses that they trusted to try to develop this?
Obviously none of the companies within Japan, but was there anybody else that they looked to or hoped would develop this?
So the idea of the consortium that was set up in the late 90s was just to research and produce fundamental research into the potential of extreme ultraviolet, for that technology to be available to anyone who wanted it.
So had there been a domestic US player that was strong enough to carry it on, then they could have done so.
The Japanese set up their own consortium and coordinated and collaborated there to try and develop extreme ultraviolet.
And they continued to progress.
And around 2007, Nikon had a prototype of an EUV machine.
But the sheer cost of developing it and the sheer technological challenge meant that ultimately, first Canon and then Nikon gave up on it.
They had to give up.
They couldn't get it to work.
They didn't think it would work.
And yet only ASML had the funds, the resources and the support in order to push through and bring this to market.
That's incredible.
And where they stand today.
Is there anyone else relatively in the ballpark of developing a competitive machine?
Short answer is no.
It's taken from the first time they discussed it at ASML through to today is 25 years plus.
They've sunk over 10 billion euros of R&D into the development of it.
The components that go into it are not the sort of things you buy off the shelf.
They're things that had to be created for this.
You need to use a light source that I described the transition from light bulb to laser.
The light source for an EV machine involves shooting a laser at at droplets of tin that are smaller than a dust particle.
And you have to strike each dust particle twice once to flatten it, once to vaporize it, to turn it into plasma which is 40 times hotter than the surface of the sun.
And you have to do that 50,000 times a second as these things are shot through a vacuum.
It's not the light source you find on the shelf.
It's so difficult to make.
About a decade ago, ASML had to buy in the light source provider, a company called Syma in San Diego.
So no one else can make that light source, nevermind the mirrors that it used to focus the lights, nor all the other components that go into an EUV machine.
I don't see how anyone would catch up.
Maybe in 10 years you might want to come up with an approximation of the current EUV machine, but by that point you're onto the multiple generations forward of UV.
So, of all the competitive advantages, of all the technological advantages that I've ever come across, I can't think of one that's more significant than AS miles.
There's some barriers to entry, to say the least.
In terms of the actual value chain and where they fit in.
You mentioned the cost of one of these machines.
Do you have a sense of what this represents in terms of the overall production cost of a semiconductor and how much this particular step in the process, this machine, how much that represents of the overall pie?
Yeah, so the way they tend to look at it is their share of wafer fabrication equipment.
So what we know is roughly what it costs to build a new wafer fabrication site.
And we know what the equipment share of that is.
And we can see there's industry stats produced.
The ASML tends to account for somewhere between 20% and 25% of that.
It varies according to the type of semiconductor.
So it's much higher in logic chips, which are the chips that Intel make.
It's lower in parts of the memory market where they've got a slightly different technology path.
In terms of what they've been able to unlock for customers with introducing this technology.
Obviously things have gotten smaller.
What have been the outcomes of that in terms of tangible results that customers have seen?
In simple terms... what they've been able to do is enable the continuation of Moore's law.
So, as I think other people have spoken to you about in the past, the nodes as we describe them now, there's five nanometers, three nanometers.
They don't exactly correlate to the size.
Once upon a time, these were actually the size of the smallest dimension on the surface of a chip.
That's no longer exactly the case, they're slightly marketing terms.
But the fact of the matter is that there is still this process of squeezing more transistors and smaller transistors onto a piece of silicon.
The more transistors you have, the more calculations you can perform, the more memory you have on that single chip.
So every time they're able to advance the manufacturing process that enables the production of more advanced semiconductors at the same price, which enables your iPhone to get that little bit better, which enables the processing chips made by NVIDIA to do that more complex algorithm that enables the things like chat, GBT and artificial intelligence to be done.
So, in terms of pinpointing single things, there's lots of single things you could point to, but essentially it's the whole of computing progress.
Anything that's happening at the leading edge needs advanced manufacturing, needs photolithography to keep progressing so they can keep squeezing more transistors onto a chip.
Jumping a bit into the business model itself.
When a company buys one of these machines, what's the typical life cycle like?
Is this something that it's incredibly huge, as you mentioned, in terms of requiring three jumbo aircrafts to get it over?
How long is that in operation for a customer?
So once these machines are in operation, they tend to be there for a very, very long time.
They recently gave out a stat that 90 of all of the lithography machines they've sold in the last 30 years are still in operation on the floors of wafer fabrication facilities around the world.
Some of these are still in the same place they were originally sold.
Some of them, when they reach the end of their natural life with one manufacturer, may be sent back, refurbished and then sold secondhand.
But these machines last forever.
Well, not forever, but they last a very, very long time.
That's the nature of the semi-industry.
Once.
The largest part of the cost of making semiconductors is in the capital cost of building the facility.
Once you've depreciated that, the cost of running these machines actually is relatively low.
The actual marginal cost of production is pretty low.
And what ASML has been very good at is ensuring that, once these machines are installed, that's not the end of it.
They're able to continue to service them and upgrade them in the field, so they get ever more productive.
So the machine that's installed 10 years ago can have an upgrade put through that markedly improves its productivity.
And that obviously is good for customers and that incentivize customers to keep buying from because they know they get a machine today.
That same machine can get better and better over the next 30 years.
And do you have a sense of the breakdown of the revenue that they generate from new machine sales versus whether it's refurbishing, upgrading or anything else?
Yes, their new machine sales account for about 75% of revenues.
And the balance is from service and fuel options, which is essentially maintenance and upgrades in the field.
And how cyclical is that revenue stream?
I assume that they're developing these machines, but there might be major breakthroughs at certain points in time.
Is there much cyclicality to it?
What ends up driving the sales.
Whether it's a constraint on amount of volume, they can produce demand or anything else.
The industry is cyclical and ASML and its history has been very cyclical.
If you go back and look at the company's history well, from the 80s onwards, but even just since it was IPO'd, it's experienced and it's gone through the pain of vicious cycles in the industry.
Those have tended to become less severe in recent years.
Partly that's down to the fact that they've emerged as the sole producer of lithography equipment.
There is no one else to go to.
There's no competition on price that they have to cope with in that sense at the very leading edge.
If you want to build a wafer fab to produce the next generation of chips you have to get your order into ASML because there's no other show in town to get that lithography machine.
And if you cancel your order because things are a bit tough this year, you're to the back of the queue when things start to look better again.
There's no avoiding that.
So what you found in recent years that ASML's revenues have become less cyclical.
That's not to say it won't see another cycle.
That's not to say you won't see machine orders pushed out.
You've seen, obviously across the semi-industry in the last few months in particular, a series of companies coming out and confessing that things are actually a little bit tougher than they were expecting.
TSMC recently cut their revenue guidance.
So I would expect that there will continue to be cancellations and oscillations along the way.
But these days it's much more about that secular structural growth, I think, in terms of that ability to keep producing leading edge machines that are needed for leading edge logic manufacturing and memory manufacturing.
I think you said they sold just north of 300 machines last year.
What could they do?
Is that the limit to the capacity that they could produce?
Do you have any sense of how much they can ramp that up?
They had a capital markets day last year and they've put out some pretty punchy numbers in terms of saying what they would like to be able to expand their capacity to.
They are investing so they can produce 90 EUV machines per year.
That compares to about 40 or so sold in each of the last two years.
And 600 deep ultraviolet machines for the next generation of equipment.
So a very material step up and also something approaching 20 billion high NAEV the next generation of EU machines by around 27 28.
So they're investing very hard at the moment to expand capacity.
And a lot of that is around ensuring that you don't get bottlenecks.
They're conscious that at the very leading edge of the semi industry, they are the bottleneck and they don't want to be the bottleneck.
They don't want to incentivize their customers to start thinking about different ways of making semiconductors.
So they're really being very explicit out there and saying that we're going to do what we have to do in terms of making sure we've got the manufacturing capacity to produce as many EUV machines, as many EUV machines, as you need for as long as you can see forward into the future.
And when it comes to the manufacturing capacity, and the investment in that.
When you mentioned before, a lot of this is actually completed on site.
So they might ship it in several different parts.
When they invest in manufacturing capacity, are they actually investing in large facilities that are used to produce certain parts?
Or are there other things that are playing a role, just trying to bridge those two things together?
This is a manufacturing business.
It's not really a manufacturing business.
It's one of the quirks and one of the very special aspects of it.
They call themselves architects and integrators.
They actually produce a relatively small proportion of the components that go into the machine, but they assemble them.
It's a relatively capital-like.
What you really have is, when they talk about building out the capacity, it's building out their capacity to assemble these parts.
About 80 of the cost of goods sold is in components and materials that they're buying from outside the business.
And the remaining 20% is coming from labor actually on site.
So it's very much about components that others are making a very specialist list of suppliers that they then integrate and assemble into a single machine.
It's a core strength of the business.
But it's a core strength that actually has become a virtue, but it actually was born out of necessity at the company's beginning.
Back in the beginning, everyone thought the best model was the Nikon and Canon model, where everything was vertically integrated.
They had lens manufacturing in-house and they could do the whole thing.
That meant they made better machines.
But ASML didn't have enough money and they didn't have enough time when they started out to do anything like that.
So they went down this path of using suppliers, of being an assembler and building a relationship with their suppliers that, I think is probably second to none.
That has enabled the whole ecosystem to move forward together.
It is a fascinating industry in that sense that there are so many fragmented pieces within the value chain.
And it's something I've learned through these conversations.
And I think the way you describe the architectural design and how key of an element that is in this overall process, it's really something I didn't expect from the outside until I went a bit deeper.
We've touched on cost of goods sold a little bit and pricing a little bit.
Can you talk about the margin profile of this business, how they go about setting price, particularly in a product where they seem to be the only supplier of that product?
What does that look like over time?
And what does that look like today?
The margin profile has, as you might imagine, improved over time as they've gone from scrapping away from market leadership to becoming the dominant player.
So their gross margins are now around 50%.
That's up from in the mid-40s last decade and are averaging around 30 gross margins back in the 2000s.
That translated last year to operating margin of around 30%.
And that's likely to improve further from here as the next generation EUV machines start to come into production and the existing generation of EUV machines starts to increase in volume.
So there's room for those margins to improve.
But again, what's interesting, and I think what's been key to the success of ASML is, in spite of the fact it's been the dominant marketplace for nearly 20 years now, it's never looked to price gauge.
Its approach is incredibly collaborative.
I've spoken a bit about their relationship with their suppliers, but it's also the same with their customers.
This is an industry that is, I think, unlike any other in terms of the way that the companies collaborate in order to agree on and advance technologies.
And that only works really if you're honest and open with your customers and they're honest with you and you can see the sharing of risk and reward.
ASML puts billions and billions into developing new machines.
TSMC's business model, Intel's business model, Samsung's business model.
They don't work if they don't have the equipment to make these advanced semiconductors.
So there needs to be sort of a quid pro quo in terms of, well, I'm going to invest now.
You're going to buy these at a price that enables me to generate an appropriate return for all that investment and risk that I've taken.
But equally, you having supported me in developing these machines, the quid pro quo from me is I'm not going to gouge you.
I know I'm the only house in town for an EV lithography machine.
In theory.
Economic theory would say ASML could jack up their prices by twofold and take what they like.
But they know that if they took advantage of their position of pricing power, that just produces this massive incentive for their customers to look elsewhere, to look at alternative technologies, to look at alternative providers and to sink money into providing themselves with an option.
So you have this, what looks from the outside, quite uncomfortable position in terms of having a relatively small number of customers who are, in themselves, incredibly powerful.
But they're really codependent on each other.
And the success of the relationship has been born out of ASML not trying to take advantage of its position, either with its customers or with its suppliers, but seeking to price its products solely on the basis of the improvement that they deliver to their customers.
And they talk about trying to split 50-50 in terms of the profitability improvements they hand on to their customers.
They want to split the benefit of that 50-50 between themselves and their customers and they need to share that benefit down their supply chain as well.
Yeah, it's a fair point and seems like it's an incredibly important factor.
If it's not going to be a vertically integrated industry, there needs to be some level of cooperation on the supply chain side of things, where no individual piece in that value chain is bottlenecking the overall industry.
And on the pricing as well, you're not extracting more value than you're creating or not sharing in that value creation.
So it's a pretty fair point and an interesting point to make.
You mentioned the concentration of customers.
What exactly does that look like for ASML?
So their top two customers were nearly 60% of revenues last year.
I guess they don't disclose the third customer, but I guess if you put that in you'd be somewhere around 65 70 of revenues.
So it's a very concentrated customer base.
There are relatively few companies left in the industry that are capable of making leading edge semiconductors.
They are very large businesses and they have deep pockets.
They are the customers.
They're the only people who can afford to buy ASML's equipment.
And they're the only people frankly, that have the know-how of how to make the most of it and use that equipment to make the world's very best semiconductors.
I think this is probably a fairly obvious question, but in terms of being an acquisition target, is that just a regulatory, absolutely not happening scenario?
It seems like there's such an important piece of the pie.
What stops one of these players from trying to acquire the business?
I suspect the market cap provides a reasonable hurdle today.
It's a fair question if you roll it back.
It's not that long ago that they had a market cap of $20, $30 billion.
And why didn't someone come in and buy it there?
I think it is the nature of the industry.
Semiconductor industry has been successful because, over a series of multiple decades, it's been carved up into a series of niches, with companies playing their own role in different parts of it.
Companies that are involved in the design of semiconductors.
Companies that are involved in manufacture.
Companies that are involved in making materials, equipment that's used in the manufacturing, and companies that take those semiconductors and stick them into products that are sold to consumers or industry.
And the success of the industry has generally been around these pockets of expertise being carved out to companies that focus and become world-leading at it.
And then you've seen concentration within those niches, but you haven't perhaps seen so much in the way of vertical integration.
Samsung or TSMC tried to buy ASML in a decade ago.
I think the odds are they wouldn't have been able to develop EUV.
They certainly wouldn't have had cooperation from the other major semiconductor manufacturers in funding the RD that was necessary to make EUV work.
So everyone benefits really from ASML being an independent company.
Because Samsung is able to piggyback on the work that's being done by TSMC and work that's being done by Intel and the work that's being done by Micron and so on.
Back to the financials, reference some of the investment that they're making into manufacturing capacity.
What does the cash flow stream look like as an investor when you look at the business?
You have a certain margin profile.
How much of that is just immediately getting plowed back into the business in the form of capital investment or some other form of investment?
And how much is either being returned to shareholders or doing something different?
The company's cash generation is actually pretty fantastic.
Not least it's helped because customers pay a down payment.
They pay for this equipment.
Often there's only the development machines that are able to take deposits up front.
So that's been helped to support as they've had to spend heavily on R&D and building that capacity.
So the cash generation of this business has actually been fantastic over the years.
So pretty much net income converts to free cash flow generation.
And that free cash flow generation has generally been recycled either into dividends or which is highly unusual for a company that is growing as rapidly as it is or share buybacks, which the company's been pretty consistent about doing at modest levels if you go back 10 years or more, but at fairly substantial levels over the last five years or so.
So the company is able to continue to invest.
It spends every year about 15%, 16% of revenues on R&D.
It's investing as heavily as it ever has on CapEx, but it's still able to pay a dividend and give its cash back to shareholders to buy bags.
You mentioned capacity expansion, having leading edge technology.
These seem like drivers of the business over the next several years.
Is there anything else when you think about a bull case from an investor's perspective that stands out as an opportunity for ASML or anything else we didn't talk about in that category?
The opportunity really is.
The history of the company has been about smaller images, better resolution, faster throughput.
That's what they do.
Make things smaller, make them sharper, and make them faster.
And that's going to be the core of this business for a very long time, because that in itself is just this massive structural growth story, because the number of transistors we need, the number of chips we need in the world is going to keep growing, unless something very dramatic happens.
It's going to keep growing for a very, very long time.
So simply delivering on that, I think, is a massive opportunity out of them.
There are aspects that they are able to leverage more and more beyond the simple manufacturing I say simple beyond the manufacturing of lithography equipment.
And that comes around some of the computational stuff that's done around what they call holistic lithography.
You can't use an optical examination technique on leading edge semiconductors anymore because the feature size is so small.
You just can't see them.
So nowadays you need to fire electrons at the surface of a semiconductor to identify where the defects are arising.
And they're able to do this down to a resolution of a single nanometer.
That's four silicon atoms and that's not a marketing spin, that is actually four silicon atoms wide they're able to measure.
So the better they get at doing that, the faster they can do that with their e-beam metrology equipment.
The better information they can feed into the computers, The better they can adjust the manufacturing of the actual lithography machine, the better the semiconductors coming out of the other side will be.
And really that's the cycle that they will go through.
They'll keep working on that until Moore's law runs out.
I guess, in terms of monitoring, that as an investor, I can make the extreme comparison to something like biotech, where you have a drug and there's very clear stages of FDA approval where You get to a point where you know whether this is going to be an economically viable product.
Here, there's some level of technology that's being tested.
How do these milestones get measured or the events that determine which?
Yes, this has been a success and we're able to go on and sell this.
How has that been announced historically?
Is it something that's coming from the company or from customers?
And how do you expect that to evolve in the future?
I'm just curious with something like a drug approval.
These are kind of big announcements that have major impacts.
What does that look like in this market?
It's quite different, I think, would be the short form.
The one thing that was maybe closest to say a biotech process would be maybe the development of EUV, which was a major generational shift in manufacturing.
It was give or take 20 years on from the previous generation machines and took a fundamental design of the machine.
And there was a lot of uncertainty as to whether it would work, because you were changing so many different things.
You were changing the way that the light was focused.
You were going from using lenses to using mirrors to focus the light.
We had an entirely new light source system that needed to be invented and created and powered up and you had to do everything inside the machine in a vacuum because the light they were using was so delicate that it got absorbed by air particles.
So that was a massive risk.
We hope, and gradually they would report to the market on terms of test achievements.
They had achieved a certain amount of power in their light source, but one of the, I think, strengths again of the ASML businesses a lot of the improvements are incremental and they're able to do this because they have a modular approach to design.
And again, something that dates back right to the beginning of the company's history.
It had to take an approach that enabled it to get something to market quickly and it worked.
Rather than trying to make machine as one piece, they just break it down into components and they make each of those in parallel and then put them together.
And the brilliant thing about that is it means that they can take the current machine And they upgrade a single part of that.
Take one of the modules and upgrade that.
And there's risk obviously in upgrading that module.
But if it doesn't work it's not as if the whole machine falls over.
We'll be working on another part of the machine as well.
So a lot of the technological development, though incredibly high risk in the sense that you are pushing the boundaries of physics and boundaries of what most people would think was physically possible.
The actual process they've developed for introducing each of these new improvements is as de-risked as I think you can possibly do it.
So, yeah, it's about monitoring it.
It's about seeing things come through.
Yeah, we'll be able to get some readouts on how things are going with the mirror development for high NA, which is the next generation of EUV equipment down the line.
But a lot of the continuing improvement in terms of existing stuff that's in the market.
That's pretty gradual and incremental.
It's happening all the time.
And it's as low risk as anything can be when it's as complex as what they're doing.
They have de-risked it to a certain extent.
That's actually an incredibly helpful explanation.
I didn't appreciate the incremental progress you can make on the various components of the overall machine.
So that makes sense.
When you do think about the risks to the business, what stands out the most to you?
I would say there are probably three things, big things, that I would worry about in terms of the key risks.
The first one would be the ability of their supply chain to keep up with them.
The supply chain they would describe as their biggest competitive advantage, but it's also a vulnerability, because they need their supply chain to be able to keep up with the technology advance that they are trying to push.
And over the years, some of their suppliers haven't been able to do that.
And ultimately, they've had to buy them in.
So you saw that I mentioned previously in 2013 when they bought in Syma, their light manufacturer.
You saw that when they took a stake in Zeiss, their lens manufacturer.
They don't own it, but they have a stake there.
And that was just recognizing that the amount of capital investment required to bring the next generation of equipment to market was beyond the budget of Zeiss as an entirely standalone company.
So there is that challenge.
Each of their suppliers needs to keep up.
It's not just about what ASML are doing within their own R&D facility.
So that is always a challenge, but it's something they've managed, I think very, very well over the years.
And I think I would expect them to continue to manage it pretty well.
The next thing would be disruptive technologies.
Again, this really comes down to the idea that ASML's competition is not really another company, but it's Moore's Law.
So people keep buying ASML's machines as long as they're able to deliver the productivity improvements at a price that's acceptable.
So as long as it enables next Apple trip to be produced at an acceptable price that makes it commercially viable product, then those machines will still be bought.
If they struggle to do that, then the technology path might prove to be something different.
And you saw that a few years ago, about a decade ago, in NAND, one of the types of memory that exists today.
They could see the EUV was not happening.
It was not going to happen anytime in the near future back in 2012, 2013.
And new approaches were developed and started to be adopted.
And that was specifically going three-dimensional.
So instead of scaling it two-dimensionally across the surface of the chip, they started to build upwards on the surface of the chip.
So essentially they went three-dimensional.
And by doing that, it meant they didn't need leading edge lithography anymore in order to make more advanced memory chips of that type.
They could simply build it up by layering more and more through etch and deposition.
So the share of lithography capex within that part of the market fell quite precipitously as a result.
And it hasn't disappeared.
You still need the lithography there and it still grows as that market grows.
But it was a really big hit.
And I think it's a really good example of when ASML can't deliver something quickly enough.
The market has to keep moving.
They're not going to wait around.
So if that happens in another part of the market, if someone else comes up with a clever way of making transistors in a way that is cheaper and faster and more efficient than can be done using leading edge lithography, then the semi-industry will go down that path, because that's the way it's got to be in order for Moore's law to continue progressing.
So I think that's the risk.
Again, You've got to back ASML in the near future in terms of their ability to keep delivering and the relationships with their customers, to see what each of their roadmaps is.
But that is one of the long-term challenges.
And then the third one would be geopolitics.
That's probably more of a near-term risk.
But you can't ignore the fact that last year nearly 40 of sales were to Taiwan, nearly 30 of sales were to South Korea and give or take 15 of sales were into China.
Now you don't need to be particularly imaginative to work out how things could become difficult, given that customer base.
A very interesting and diverse set of risks there.
When we wrap up the conversations, we always talk about lessons that you can pull away from looking at this business and potentially apply to others.
What do you think stands out in that category for ASML?
I think there's probably three things in terms of takeaways that I've taken from looking at ASML.
The first one is the importance of looking beyond the cycle.
I think a lot of people look at ASML.
I confess when I first looked at it, there's this feeling that I know this is a cyclical industry.
I know things are going to go down at some point.
And you sort of worry about this.
Am I timing it right?
Am I going to look silly if I decide to take a shareholding now?
And of course, the industry does remain cyclical and you can't ignore that.
But that obsession on the perfect entry and exit point can blind you to the structural opportunity that's there, to blind you to the excellence of the underlying business and the long-term prospects when you stretch it out over 5 10, 20 years.
The other thing would be I guess this goes back to the very early days of ASML, but don't underestimate the role that luck can sometimes play in a company's early years in terms of how it comes to be where it is.
You look at the company's history before IPO.
Don't get me wrong The history of ASML is one of extraordinary ingenuity, tenacity and innovation excellence.
But in its early years there were certain things that played out pretty well for them in terms of in 1986, when they still didn't really have a product to market.
There was a massive recession and it killed a couple of their competitors and many others couldn't spend much in RD.
So they were able to catch up at a time when they were otherwise stagnant. on their knees.
Their business with the technology transfer on EUV, the fact that they were the only other player versus the Japanese.
If they'd been a viable US player, would they have been given such an easy road into that?
I don't know.
And then the third thing was this is the really biggest part of things about looking at ASML is not to underestimate the power of human ingenuity.
Because I've looked at all the research we've written on ASML going back over decades, And there's always been a huge amount of uncertainty about what this company is going to look like and what the semiconductor industry might look like five or 10 years out.
There's a great story back when Peter Wernicke tells this story when he joined the company back in 1999.
Before he did, he spoke to Martin van den Brink and asked him is Moore's Law going to still be running?
How long is Moore's going to run for?
And Martin van den Brink said to him, I think at least 15 years.
And 15 years later, you come back to Peter, to Martin and says, that's still going.
How long do you give it?
Martin says, I think about another 15 years.
And it's never been possible, even for the guy who's probably the smartest guy in this industry.
He knows more about it than anything else.
He's never been able to look forward to that.
And we struggled to look forward five years.
But what we don't appreciate is just how much is going on under the bonnet at ASML and just how much Innovation is taking place at that business and the drive there is to keep bringing through innovations that have kept advancing Moore's Law many years beyond when people thought it would die.
One of the things that I enjoy talking to you and your colleagues the most about is that deep history of notes and the deep history that you have with these businesses, where you can reflect not just on what's happened in the past three years or five years, but often the past 20 years.
And it shines such a light on how things often feel the same as they feel today.
And if you look at the path that the business took 20 years ago, when it was very much the same feeling,
I might just provide a little bit of extra perspective on today.
This was an excellent conversation, Tom.
I really enjoy learning.
And every bit of this conversation was me learning about this business and this industry some more.
So thank you very much for joining us.
Thank you.
To find more episodes of Breakdowns, ranging from Costco to Visa to Moderna, or to sign up for our weekly summary, check out joincolossuscom.
That's J-O-I-N-C-O-L-O-S-S-U-S dot com.
