So ... ASML is the only maker of high volume EUV lithographic equipment. They seem to sell >300 such systems per year. How many systems are needed for a high capacity fab (>100 000 wafer starts/month)?

I mean ... I get that making this equipment is ridiculously difficult, but this just doesn't seem viable, with a single company responsible for everything down the line, and with them having this limited production capacity.

I wonder what their profits were 5 years ago?
www.macrotrends.net/stocks/charts/ASML/asml-holding/net-income

Hmmm. Why do I feel like I am being gouged when I see sharp profit curves like that??

ASML = ASMR

mechtechI wonder what their profits were 5 years ago?

I too wonder why they charge so little?

zlobbyASML = ASMR


I too wonder why they charge so little?

Ya who knows. Maybe they charge $100 billion per unit ;)

R&D. EUV lithography has had a looooong road of development. The investment is now paying off.

ValantarSo ... ASML is the only maker of high volume EUV lithographic equipment. They seem to sell >300 such systems per year. How many systems are needed for a high capacity fab (>100 000 wafer starts/month)?

Rumours are, that from 5nm and later, the output capacity of EUV is only 15000 waffers/month.

piloponthRumours are, that from 5nm and later, the output capacity of EUV is only 15000 waffers/month.

Per system? So a "gigafab" type setup then requires 7 systems, or 10% of the quarterly output of ASML. That's not terrible overall, but I doubt all those machines are EUV machines, and I doubt all of them are going to general-purpose fabs like TSMC. Still, at least EUV volumes are pretty high - it wasn't long ago that they were shipping single digit volumes. But this market desperately needs a second actor IMO.

ValantarBut this market desperately needs a second actor IMO.

I wonder why nobody else thought of that? :D

zlobbyI wonder why nobody else thought of that? :D

I'm not claiming this is a new in any way - there's a reason most somewhat functional economic systems have regulatory agencies tasked with avoiding this type of situation, after all. At least Canon and Nikon have been involved previously, though neither of them operate at the cutting edge today. Capitalism tends towards monopolies, especially in high risk markets with long turnarounds for R&D and high demands for technical expertise, so the current situation is easily explained. But that doesn't make it any less dangerous. Heck, imagine the global scope of damage a single determined terrorist could do if they targeted ASML. Even outside of unlikely scenarios like that there are many, many ways in which this is really risky. And it could be alleviated in many ways through various regulatory actions.

ValantarSo ... ASML is the only maker of high volume EUV lithographic equipment. They seem to sell >300 such systems per year. How many systems are needed for a high capacity fab (>100 000 wafer starts/month)?

I mean ... I get that making this equipment is ridiculously difficult, but this just doesn't seem viable, with a single company responsible for everything down the line, and with them having this limited production capacity.

Keep in mind this is NOT 75 EUV systems per quarter. This is all systems so also DUV and maybe also older legacy systems. They mentioned capacity of 55 systems a year for next year (so this year probably lower).

Record is apparently 160 wafers per hour but this I expect is single layer and wafers require multiple layers. Assuming standard is about ~75% of the record you get ~85.000 wafer starts per month

Link here: www.asml.com/en/investors/financial-results/q3-2021

ValantarEven outside of unlikely scenarios like that there are many, many ways in which this is really risky. And it could be alleviated in many ways through various regulatory actions.

What would you propose? No one has been able to do what ASML is doing, and not for the lack of trying. So should the goverment come in and destroy their business by forcing them to open up their patents?

Eventually a new player will emerge if enough inc$ntiv€s materialize, for now we're stuck with ASML alone (btw they don't operate alone, they leverage several partners to build the machines)

ValantarI'm not claiming this is a new in any way - there's a reason most somewhat functional economic systems have regulatory agencies tasked with avoiding this type of situation, after all. At least Canon and Nikon have been involved previously, though neither of them operate at the cutting edge today. Capitalism tends towards monopolies, especially in high risk markets with long turnarounds for R&D and high demands for technical expertise, so the current situation is easily explained. But that doesn't make it any less dangerous. Heck, imagine the global scope of damage a single determined terrorist could do if they targeted ASML. Even outside of unlikely scenarios like that there are many, many ways in which this is really risky. And it could be alleviated in many ways through various regulatory actions.

If a single determined terrorist would target ASML, we would be falling back on DUV lithography, and its not like anyone is going to sleep any less for that. And those that do, we shouldn't really care about.

EUV is just a new node enabler. None of that is 'required' or 'mission critical' contrary to what people might say. We can do just fine riding 20-14nm nodes as we used to. Economically its going to create a huge market shift, but the market will survive because the demand is still going to be there.

Functional economic systems... yeah. I think the definition of functional is undergoing heavy inflation these days if we consider our current economy, fueled by hypercapitalism. It is that ideology in fact that causes you to say what you say - we've been born with the notion that economy is first and everything else comes second. Total BS. Its a system that worked for us, but does it still do that today? We're diving from one crisis into the next and we've been sticking bandaids on whatever is nonfunctional for how long now?

As for regulation... we're seeing the same slippery slope where regulators are understaffed and toothless, more and more. Shit has to hit the fan in a BIG way before we stop using the bandaids and start revising our systems. A pandemic wasn't enough, go figure, we're diving straight into the money pit like Lemmings again. In the sense of 'ASML and competition' though, no amount of regulation would have created a different situation. Its the same thing as what Intel has been diving into - they were adamant they had the grail with 10nm and they stuck with it. ASML was stubborn enough to keep chasing EUV even though the outlook was pretty grim and the margins for error super tiny. The risk involved in that investment... there was no single other company willing to go there. How do you regulate that then? What's happened here is just a design win.

Vayra86If a single determined terrorist would target ASML, we would be falling back on DUV lithography, and its not like anyone is going to sleep any less for that. And those that do, we shouldn't really care about.

EUV is just a new node enabler. None of that is 'required' or 'mission critical' contrary to what people might say. We can do just fine riding 20-14nm nodes as we used to. Economically its going to create a huge market shift, but the market will survive because the demand is still going to be there.

Functional economic systems... yeah. I think the definition of functional is undergoing heavy inflation these days if we consider our current economy, fueled by hypercapitalism. It is that ideology in fact that causes you to say what you say - we've been born with the notion that economy is first and everything else comes second. Total BS. Its a system that worked for us, but does it still do that today? We're diving from one crisis into the next and we've been sticking bandaids on whatever is nonfunctional for how long now?

As for regulation... we're seeing the same slippery slope where regulators are understaffed and toothless, more and more. Shit has to hit the fan in a BIG way before we stop using the bandaids and start revising our systems. A pandemic wasn't enough, go figure, we're diving straight into the money pit like Lemmings again. In the sense of 'ASML and competition' though, no amount of regulation would have created a different situation. Its the same thing as what Intel has been diving into - they were adamant they had the grail with 10nm and they stuck with it. ASML was stubborn enough to keep chasing EUV even though the outlook was pretty grim and the margins for error super tiny. The risk involved in that investment... there was no single other company willing to go there. How do you regulate that then? What's happened here is just a design win.

You seem to be reading pretty much the opposite of what I was alluding to into my writing. There's a reason I said somewhat functional regulation (because that is the best that anyone can make any claim to), and I made no claims to this being widespread or anything but in (rapid, planned and willful) decline. I'm also not saying this as an argument towards upholding late stage capitalist logic, but rather through the (grudging) acceptance that until we can put together and enact reform on a massive scale, the system we have is the system we have. And within that system, any breakdown of the (idiotic) race for the utopian ideal of eternal growth and progress leads to job losses, poverty, and human suffering. The system is in no way set up to handle disruptions - as we are seeing with the current supply shortages. So, while it's true that there would absolutely be alternative paths to stability in the case of a catastrophe, the short-term consequences before things got turned around (which might never happene - letting people starve is typically cheaper!) would be massive human suffering, as every industry reliant on these components (which today is every industry) would grind to a halt in some way or another.

You're entirely right that we can do just fine riding current nodes, but within an economic system where anything less than ~7% annual increase in GDP is pretty much a recipe for massive unemployment, that kind of stagnation won't work. I would love it if it did, as that would mean we were much further from the edge of the cliff than we are today, and much less eager to keep accelerating towards it, but that's not reality.

Heck, the continuous drive towards innovation and expansion is exactly why capitalism tends towards monopolization in the first place - it's always cheaper to be the only actor in a market, as it means you're not spending money on competing with anyone but yourself. So while anti-monopolist arguments are often geared towards increased competition and a drive for further innovation (which it can absolutely be linked to in some cases), there's also an argument for simply having more sources for what we need in the first place - which was the argument I was making here.