Whether it's Apple's latest M1 chip made by TSMC or Nvidia's latest GPU made by Samsung, it's needs to be manufactured using ASML's EUVs.
The crazy part is, ASML was only able to ship 31 EUVs in 2020 and there was only about 100 EUVs in the world as of January this year.[1]
I do wonder if this is indicating a flaw in ASML’s business model more than anything else.
The entire (massive) CPU business hinges on these machines in 2021 and given that I’m quite certain ASML hold substantial patents it seems like they should be offering these machines for per-wafer fees rather than selling them directly to TSMC and Samsung for them to print endless mountains of cash with them.
AMD and Apple are utterly reliant on these devices, and no-one outside of tech even knows they exist. If no-one sees the massive value your company contributes that’s a problem with PR.
EUV lithography is a strange beast. It requires extremely clean rooms and a lot of practical know-how and skills from the humans around it. Few companies are in position to actually use EUV machines to produce chips.
So, in practice, ASML and the few companies are dependent on each other and probably found local optimum of coexistence. ASML does not want to push several top-level tech companies into search for alternatives and the top level tech companies do not want to waste their talent on replication of an already existing product unless necessary.
This is very true, some one who has interacted with the tech in the past. It’s mind bending physics that if you do the cal a on a napkin you could do it, but all the materials have 1000 tweaks to deal with imperfections.
They are also so far ahead, after uv it will be X-ray, and even electron etching. They are developing tech 10 years ago that still will get them further. At some point there will be enough power, for general consumer products to function just fine. Once a laptop can run email/web/word/excel smoothly, which I think is like 90% pc use, on a whole day battery. I think we don’t need more speed. It might be at that point, that some players step in. Try to build machines that work at similar quality. However, why not just get a second hand ASML? As far as I know once a new generation get bought the old goes to a lower tier processor manufacturers.
Once a laptop can run email/web/word/excel smoothly, which I think is like 90% pc use, on a whole day battery. I think we don’t need more speed
Thats not how it works, software is a gas that expands to fill all the available performance space until it all slows down and we need better hardware again.
But suppose some nation decided that they could devote more software development resources to designing efficient software, and that they could limit processing intensive applications (like online gaming?).
Conceivably they could run the economically important IT functions of the country using less powerful, easier to manufacture, less power hungry greener, and readily available semiconductors.
The M1 MacBook Air can run a whole day (and more) on one charge, but „only“ 5h when doing heavy development with Xcode. If you build a machine that lasts 2x as long with 10x the power, I‘ll pay at least 5x the price for that. Maybe 10x.
> However, why not just get a second hand ASML? As far as I know once a new generation get bought the old goes to a lower tier processor manufacturers.
It sounds like you'd be stuck producing old designs on those old machines.
During the time I worked at ASML, I spoke to some of the engineers about things like this.
Apparently, that equipment rarely gets sold off or excessed. Those machine keep producing for years and years and decades even, for the same companies, in the same fabs, in the same location. Even as they get older and older, they’re still so profitable to keep running that it almost never makes sense to turn them off or sell them to someone else.
They were actually quite proud of the fact that they had lots of old hardware in the field that was still running and producing wafers for the same customer so many years later. They were a little taken aback by what it cost them to keep those old systems around themselves, so that they could provide support for them in the field. But that’s what the customers were paying for, and so that’s what ASML did.
In an over simplify model you would expect that. But in reality there is so much going on having EUV doesn't immediately guarantee success, just ask Samsung about it. And at what point will Carl Zeiss do the same to ASML when they are solely relying on their optics? TSMC uses their own in house pellicle ( as suspected for some time and only recently confirmed in an investor meeting ) There also dozens of other companies in the Foundry supply chain that are also critical to the entire process that just hasn't reached mainstream media.
In an investor meeting, ASML told its audience that TSMC knows more about how to operate their EUV Twinscan then they do.
And no, Intel, Samsung, and TSMC are no longer ASML's shareholder, as somehow constantly being pointed out in various form of social media including HN. They only invested in I think 2011 / 2012 to help fund the development of EUV. They have since reduced their investment once this was done.
It is also funny you mention this, because I am sure if ASML were an American company they will absolutely do that to maximise their profits. But ASML is Dutch. And ASML got to where they are today with the help of TSMC. It used to a fight between Canon and Nikon. But that is another story.
Not really. It's that they have ZERO competition AND the problem of inventing/manufacturing the technology is so hard that no one can afford to compete and its barely viable for ASML as it is.
The other issue is that $150M for a stepper is more than an order of magnitude more costly than the previous stepper technology used for 20+ years. To make that work, very expensive CAD tools were required to grossly violate normal diffraction laws (but obviously not the laws of physics themselves). This is done with "multi-patterning" which uses multiple exposures of the same targets to attain sub-diffraction limit dimensions. It requires many additional process steps to do this and it's mathematically/CAD intense.
And finally for fabs, the throughput of ASML EUV is about 1/50th of the prior steppers (but the previous steppers required 5-10 additional multi-patterning stepper exposure steps in the process to attain similar results. The general rule of thumb is an additional process step is 1/2 day of throughput so the old technology would add a week or two of throughput but the lower throughput of EUV only saves about 50% of that so the throughput is worse or a wash.
As it is, the wavelength of EUV is 13.5 nm so the diffraction limit is still only 6.25 nm feature size (so 7 nm is clean without multi-patterning but below that you need to return to multi-patterning but not as many steps added (yet).
The "Net Net" is that yes EUV is the only way get deep nanometer dimensions (<25 nm) easily and cheaply BUT EUV is economically marginal for both stepper vendor and fab.
The former is why no competitor wants to jump in (plus patents as well). The latter is why it's certainly necessary if the commitment to deep nanometer is life-or-death for the fab. A LOT of fabs have decided that there are better markets above those dimensions that they can tap so they've written-off deep nanometer and switched orthogonal to different market value propositions.
Fun Fact 1: another common name of EUV is "soft X-rays". As such, conventional lens-based optics DOES NOT WORK, so the ASML system uses the only technology that works for X-ray optics: low-angle, glancing metal mirrors. A well known other application of this is X-ray telescopes in space.
Fun Fact 2: the optical power level at the mask is less than 10 watts which is barely enough to expose the photoresist. This requires far longer exposures with perfect mechanical stability to hold the mask-created image exactly still. This is why the throughput is so low compared to prior steppers. This is also where much of the capital cost comes from: stabilizing mechanical components far beyond what prior technologies required.
"ASML agreed that 55 percent of the components used in its commercial EUV systems sold in the United States would be sourced from U.S.-based suppliers, which Moniz said will keep high-paying jobs in the United States"
ASML doesn't hold the core patents. The U.S. Department of Energy allowed ASML to join the EUV LLC. ASML bought Cymer for exactly that reason.
ASML has a market cap of c $350 billion. If they wanted to raise tens of billions on the public markets they could probably do it. Not VC territory at all. They're probably limited by the size of the market.
My guess is that the limit is human resources. The machines are probably assembled by hand and they probably need highly-skilled and specialized engineers to work on it, such kind of profiles are rare.
The demand is not large enough to industrialize the production of such a complex machine so you have to rely on humans.
Each of their machines is custom and needs very specialised knowledge, so it's difficult to scale up. However, VC could encourage competition, which would help keep the chip industry healthy, as we all know what happens when a company has a monopoly in an industry.
I'm a firm believer that at certain scales, competition becomes impractical to the point of irrelevance. The incentive structure has to exist to draw competitors into capital intensive markets like this and it just isn't there.
As the article points out, this company has been pressured at geopolitical scales to deny China access. The fact the US government recognizes the value in denying China access because competition becomes impractical says a lot. China is filled with technical and engineering talent no different than the rest of the world and has its fair share of capital as an economic powerhouse. There's a lot of incentives for the government and people of China to develop their own competitor but it's recognized that this is impractical due to the level of proprietary knowledge, instruments, and skills needed to recreate the device. I'm confident they can produce a competitor if they choose but it'll be decades without any real competitor.
Now pretend you're not the PRC but Joe Sixpack in the US and you want to enter the market. We need to drop these ridiculous ideas of competition regulating the markets. Even injecting VC money here is probably at irrelevant scale. I'm all for competition but we need to look for new models of market regulation. Competition just isn't it in many cases.
it was started 37 years ago and has a mkt cap of 350 bn. Not sure that qualifies as a startup, and neither does it serve as a counter pointer to "Europe's lack of startups and unicorns."
and they are based on research done by the United States which they licensed, hence why the US is able to block them from exporting to China. All the UV lithography research was done by the US decades ago
They are often mentioned in these context, along with Spotify, Adyen, SAP and a few others. It doesn't really fit the US-based narrative so is mostly ignored.
For reference, until quite recently, chips that were referred to as being "14nm node" were actually built using light that was 193nm and lots of special techniques like multipatterning, immersion, crazy lenses, etc. So when EUV arrived and is now using 13nm, you can bet that 13nm will allow quite a bit of runway for creating tiny features -- one concern I'm sure is the use presently of mirrors for all of the optics, so not sure how that would work for something equivalent to immersion -- seems like not possible maybe?
AFAIK immersion is very difficult because EUV is kind of like a disruptor beam. hc/13 nm is about 95 electron volts; that's enough to ionize just about anything, though they more often boost inner (non-valence) electrons up to outer shells instead. It's just as accurate to think of them as super soft X-rays as it is to think of them as "extreme ultraviolet." Sometimes they're called "vacuum ultraviolet" because air is opaque to them.
Refraction is the key to immersion's resolution improvements, and I don't really know how it works. I have the impression that the higher permittivity of the refractive medium means that the electric field of the light propagates more slowly into it because the material is electrically polarizing, but permittivity varies with frequency, normally decreasing at higher frequencies, so hard X-rays barely refract at all. That would suggest that EUV might not refract much, quite aside from being rapidly attenuated.
However, at least within and near the visible spectrum, refractive index increases at higher frequencies. So clearly there's a lot I don't understand.
If even air would absorb/scatter the EUV beam, immersion in liquid would seem to also cause a LOT of scattering. That'd reduce the contrast, which sounds undesirable. (Speaking as a complete noob here)
I've recently bought several shares of ASML at about €670, as I can't imagine them going down anytime soon.
They seem to have a pretty much monopoly on the most advanced chipmanufacturing process and respective machines.
ASML stock is also already going up for years and years. I wish I'd bought them earlier.
Although ASML should be our Dutch pride, it's a company we hardly hear about in the Dutch press. Probably because their tech is relatively hard to understand.
They don’t have a monopoly on EUV, although they have the most mature production systems, albeit still at low volumes. The core technology and research behind EUV was actually performed by US Department of Energy labs, who then shared their work under agreements with a small number of industry partners, including but not exclusively ASML, back in 1999 (https://www.eetimes.com/u-s-gives-ok-to-asml-on-euv-effort/). This is also why ASML has a US presence and must source components from the US. There are other companies playing in this space, and while several are pursuing EUV, some are also using alternate approaches to fulfilling the same demand (like Canon’s nanoimprint). I would characterize ASML as temporarily having a long lead on EUV, but that can change in the span of 5-10 years.
There is not going to be a competitor to ASML in 5 years. There might be alternatives on the horizon in 10 years, but they will very much still have a monopoly on EUV lithography.
Living in Eindhoven & bought a couple of years ago as my first ever stock purchase, went to the yearly shareholder meeting in Veldhoven which was super low-key with a very homely vibe & some super long-term holders (20+ years). Still my most profitable purchase as well.
I've heard of ASML for over a decade now - Dutch mainstream press might ignore them, but specialized sites like tweakers.net and partners (it's been one of my daily visits for nearly 20 years) have reported on them for ages (https://tweakers.net/zoeken/?keyword=asml#filter:q1bKTq0szy9...)
As for the stonks, it looks really healthy, only ever having gone up over the past 20 years. Significantly in the last two years though. But I think it's a safe bet given there's little competition on the market at their level. I'm sure e.g. China is working on a competitor, but they have their work cut out for them - even if they have the designs for these machines.
It's priced in at the confidence level the market has. If you're more confident than the market then it's an opportunity to make money. (You also have to be right, of course.)
"Don't even ask the question. The answer is yes, it's priced in. Think Amazon will beat the next earnings? That's already been priced in. You work at the drive thru for Mickey D's and found out that the burgers are made of human meat? Priced in. You think insiders don't already know that? The market is an all powerful, all encompassing being that knows the very inner workings of your subconscious before you were even born. Your very existence was priced in decades ago when the market was valuing Standard Oil's expected future earnings based on population growth that would lead to your birth, what age you would get a car, how many times you would drive your car every week, how many times you take the bus/train, etc. Anything you can think of has already been priced in, even the things you aren't thinking of. You have no original thoughts. Your consciousness is just an illusion, a product of the omniscent market. Free will is a myth. The market sees all, knows all and will be there from the beginning of time until the end of the universe (the market has already priced in the heat death of the universe). So please, before you make a post on wsb asking whether AAPL has priced in earpods 11 sales or whatever, know that it has already been priced in and don't ask such a dumb fucking question again." [0]
Counter point: beginning of 2021, for weeks I was wondering why the markets were not reacting to Covid.
Only when travel restrictions were introduced, and only when they were introduced to the US, the market started taking notice.
Does the market extrapolate future earnings and macro economic trends, sure. Do some people have secrete insights beyond that? Sure. But I'd say the market as a whole is pretty stupid and reactive.
Agreed, I was trading VIX futures calls and /ES puts around that time and was equally mystified that markets were as frothy as ever. It was definitely a very wild time to be speculating.
efficient market theory is astrology for males, in a world where astrology and trading was gendered.
the commonality being that neither produce peer reviewed reproducible results.
nothing is priced in. things can be undervalued with known public information. market appetite can also further overvalue something with a new normal for price to equity ratios.
Historically speaking these one off stock recommendations from HN have really been fantastic. I just sold them too early and didn't put enough down.
Like that one comment I remember in 2011 recommending to buy and hold Bitcoin for 10 years. Right. I bought. Didn't hold. Same deal with Tesla.
I want something that takes $5,000 from me, invests it, then makes me forget about it, then 10 years later sends me a letter and says "you forgot about this, 10 years ago you did such and such, here you go"
You can do fractional buy from the US on that instrument in Sofi (not in RH) btw for those who were looking to put only like $50 in. If you are doing an onboard there to invest in this, email me (check my bio) and I'll send my referral bonus link for signup (it's $25 for stocks + $10 withdrawable cash currently). Free money...
> I just sold them too early and didn't put enough down.
Mood; I bought ETH and sold when I thought it couldn't go any higher. I've bought and sold Tesla and Apple a couple of times thinking it couldn't possibly go any higher. Still made my money back and then some though, so I can't complain.
As for your wish, index funds are your friend, low risk as well. I've got some money safely in a savings account (effectively no return on investment), some in 'play' stonks (-2.19% to +70% return on investment per year, 2020 was really good), and some in a managed investment fund that just spreads things around according to a risk profile (don't have a lot of numbers for that one, but +13.29% so far this year).
This is not financial advice, I'm confident we'll have a market crash again within a year or two, and when investing you need to manage risk; index fund (following e.g. the S&P 500) are relatively safe investments, trading stocks is fine if you don't panic sell, but options and going 'all in' on single stocks is dangerous and not recommended. Don't invest money you can't afford to lose. Don't beat yourself up for not getting in on things earlier.
I guess it's a combination of cheap money, government buying debt, government giving away money like a drunken hedge fund guy who's date backed out. (couldn't use sailor), and no one thought businesses could survive through Zoom?
I have watched professionals wait until the last Retail investor puts in his last dime, and poof--they pull out.
What's weird is I don't even think we are close to that pullout by the big guys.
The wealth created in this bull market must a world record? (percentage wise)
Personally, I wouldn't put ASML in the buy bucket for very long though. The stock has already more than tripled during the last year, mostly off the back of the big producers announcing enormous expansions of production capacity which is great for ASML. This Industry is cyclical though so in a couple of years once those factories are opened, there will be a large overcapacity leading to scaling down capex.
I wouldn't be surprised if the current chip shortage becomes chronic for the next decade. The number of uses of chips are skyrocketing as products that are now "smart" compete to be even smarter.
AI applications are processing heavy and applications are growing exponentially.
The competition in the highest end chips is fierce now that Ford, Toyota, etc. are competing against Tesla which has its own custom chips.
Augmented reality in the next several years will severely push demand for the highest performance, lowest energy using chips. Which tend to be the top of the line chips.
Governments around the world have realized chip production is an existential capability. The US alone will be pouring money into catching up and matching TSMC, Samsung, etc. And those companies will respond by prioritizing progress even more.
My money is on their not being a downturn in chip demand, or over capacity, for a very long time. Maybe not for decades.
That's a strong statement, but I think computing has turned a corner.
The shortage is not in high end used for AI or self driving car computation, it’s in low end chips like power management, audio codecs, and microcontrollers
I think the competition to create smarter products is across the board though. Once a product category has chips, it can be made smarter than the competitors. So its game on at every level of manufacturing.
"Smarter" not necessarily meaning AI, but can be AI.
You could have said that about Microsoft stock in 1990, Amazon in 1999, Google in 2011, Apple in about 2005... And about lots of forgotten flops as well.
The future is unpredictable and the best payoffs are going to inherently be controversial and have a larger risk
I’m feeling gold. It hasn’t moved for a decade and has all the reasons to move in the next one as people escape the currently overvalued overprinted mess we made for something real and tangible again.
I want something that takes $5,000 from me, invests it, then makes me forget about it, then 10 years later sends me a letter and says "you forgot about this, 10 years ago you did such and such, here you go"
I can help you here. Let's discuss. I accept eth or bc. 3f42d
If you find this interesting, there's an excellent history of precision in manufacturing by Simon Winchester called "Exactly" that includes a fascinating chapter on this. The book's chapters progress through various key technologies in history whose manufacture led to leaps forward in precision engineering, culminating with modern smartphones.
ASML get a lot of coverage since last year, I've had never heard of it before, and I suspect just like many.
It looked liked it emerged due to concerns about TMSC getting too much power and advance with billions on investments for semiconductors, and all of sudden all was alright since TMSC heavily relied on a dutch company.
Maybe there's more of it, I'm starting to suspect there's a submarine in there.
Both needs each other as to what I'v concoluded when I researched this topic.
Making the machine is one thing and using it for fabrication is different thing, it might sound naive to say so, since typically one might think generally of machines as they comes with manual, and that's all you need for opreation..etc, but that is really far from be the case for the EUV machines or the fabrication plants...
Lastly, to add to the point, if it was only about the machine, why Intel haven't started using it for fabricating <7nm...
From the text: Amid the recent chip shortage, triggered by the pandemic’s economic shock waves, ASML’s products have become central to a geopolitical struggle between the US and China, with Washington making it a high priority to block China's access to the machines. The US government has successfully pressured the Dutch not to grant the export licenses needed to send the machines to China, and ASML says it has shipped none to the country.
Having an EUV scanner in your possession is not even half the battle. Building the mirror systems alone is an enormous effort taking thousands of engineers at Zeiss. Not just the figuring and coating of the mirrors but the metrology for them takes a bunch of bespoke systems developed by Zeiss.
Two important facts that lots of folks seem to miss. The extreme ultraviolet process was created by and remains owned by the US DOE and LLNL. The US wields enormous control over ASML because of this fact. https://www.spiedigitallibrary.org/ebooks/PM/EUV-Lithography...
The EUV process is an insane rube goldberg physics "hack" that makes these machines very far from something you can just churn out model T style.
I knew I saw this on HN a couple days ago. I figure that's what "inspired" this article, though it doesn't really have any more detail than the tweet thread...
> “To tell you the truth, nobody actually wants to use EUV,” says David Kanter, a chip analyst with Real World Technologies. “It's a mere 20 years late and 10X over budget. But if you want to build very dense structures, it’s the only tool you’ve got.”
The Landauer limit is about 2.8 zJ at room temperature. Modern chips are ludicrously far away from that being relevant. A chip using 1W would need to do 10^20 operations per second to even get close; right now they do more like 10^9.
Stock prices aren’t necessarily linked to a company being important, unique, or revolutionary.
ASML’s success is highly linked to their ability to stay cutting edge with a very specific technology, and their ability to keep their literal handful of customers.
Amazon, Google, and Apple can kill a bunch of products, lose a bunch of customers, fail at a few projects, do a few things worse than other companies, and still turn a big profit.
Whenever I see a big complex looking machine like that, my first thought is "That's impressive, these guys probably don't have much competition" then my second thought is "Surely it's possible to build a smaller, simpler machine that does the same thing in a much simpler way."
Can we make Moores law stop?
I think it’s accelerating too much.
When there is acelleration there is novelty.
When there is novelty there is uncertainty.
When there is uncertainty there are problems.
We need to make research into semiconductors and progress in hardware illegal.
Only so will we be able to prevent social harms caused by artificial intelligence because the most advanced AI architectures are very resource intensive.
When there is resource consumption there is scarcity.
