[CookiesNCream23]

I work in the industry and have worked in a fab. Here is some insight.

We are taught that fabs are not treated as factories but as hazardous chemical storage plants. On top of that, we work with high pressure and high power systems.

Fires have accounted for the most damage to fabs over the years; however, this situation is different.

The site is not a fab, it is a ASML manufacturing plant. This plant does not produce chips. It produces parts for the ASML machines. It makes the tables the wafer moves on and the frame the mask moves on.

https://www.youtube.com/watch?v=jH6Urfqt_d4

https://en.wikichip.org/wiki/mask

Downstream effects of this fire will reduce the uptime of the machines and the delivery of promised machines to our customers.

To plug ASML. Speaking as a new grad. If you are in hardware, physics, nanoscience, simulations. ASML is the best company to work at if you want to learn. I get exposure to maybe the most complex engineering system is the world. The scale, complexity, details, and just hardcore technology is mind-blowing. I am plugging ASML because it is not widely know and I would love if fellow engineers had the opportunity to work here. I absolutely love the work I do.

[collegeburner]

This definitely doesn't match how I heard some software is done there: https://news.ycombinator.com/item?id=18463181

[skrebbel]

FWIW (i wrote that comment, also btw woa I'm super flattered to be quoted like that!): note that it's based on my, by now, pretty dated inside knowledge (~6y). They might've improved since (though judging by the sibling comments, maybe not).

[nebula8804]

I find it funny how the other commenter talked about how ASML could be disrupted by a competitor with better software practices. Completely laughable in todays chip shortage world.

I have had to learn this lesson the hard way when I assumed that Disney+ was going to flop when a lot of the senior engineering talent supposedly quit during the ramp up. What I didn't realize was that technology isn't as important as other things when you have a moat. Disney+ ended up relatively successful despite their technological hiccups. Same reality with ASML.

[CookiesNCream23]

Definitely two types of software here. Everything about that post is spot on.

Internal software is messy. Testing is also messy. Lots of things need to be automated. End of the day, it’s a bunch of hardware people.

The exciting stuff are the physics simulations. Modeling how lens aberrations affect the light that passes through and how to correct for that.

[ojl]

Software wasn’t mentioned in the comment though. It might be an awesome place to work at within those other disciplines which were mentioned.

[the_only_law]

This is the sort of thing that shys me away from every working at a hardware company doing software.

[Heliosmaster]

yep. accurate comment. How do you like 40M LOC with 50% code duplication?

[throwoutway]

That sounds like a train wreck waiting to happen. I’m amazed that hasn’t put an egg on their face. Are their customers complaining about issues?

And how can the risk managers know so much about potential regressions? Are they a panel of former engineers?!

So many questions

[Heliosmaster]

keep in mind that machines like those are so complex that you don't simply buy them and install them.

It's more like getting custom-made artisanal goods: each machine has a "father/mother" from its 'conception' til delivery who treats it as their (work) baby. Then a special plane will fly the machine and a team of dedicated engineers will spend weeks in your fab doing install / validation. The process is long and complex.

Downtime are really painful and expensive so the philosophy is "better copy paste the code rather than change what's working". downtime are not very frequent but people are just the least expensive / worrying thing there. It's not like you can easily start a competitor...

[r00fus]

It almost sounds like there should be states funding creation of a competitor simply to prevent the "fire impacts entire global industry" kind of news we're seeing today.

[sharken]

ASML has 16 locations around the globe so while it will have some impact, it could be a lot worse.

https://www.asml.com/en/company/about-asml/locations

[Melkman]

Not every location does the same thing though. ASML has bought other companies that were key suppliers to them. Like Berliner Glas Group in 2020 which is located in Berlin and is where the accident happened. No other location of ASML will be able to compensate for production losses there.

[Heliosmaster]

a competitor won't be enough, the whole supply chain is made by companies which are single point of failure (the reason why europe can't afford to say "ASML only sell your machines to EU fabs" is that one of the key ASML supplies is US.. and without them no machines...).

[nebula8804]

Well China is trying are they not? We will probably have to wait another 5-15 years to see the results...

[zogomoox]

The customers do complain, but they only care about wafer yield and throughput. They will accept any dirty software hack or manual procedure to circumvent issues _now_.

[jokteur]

I am currently beginning a PhD in physics (mathematical modeling, simulation on HPC) in Switzerland, and I am afraid that after my PhD, I won't find a job that really pleases me. These last years, I came under the impression that you currently can only find ML driven software engineering jobs (I am exagerating of course), and having done quite a bit of ML and software engineering, I now know that I don't want my future job be like this.

A company like ASML gives me hope. Do you know by chance if there are similar companies (complex engineering, with physics, maths, chemistry, computer science, ...) that are located in Switzerland ?

[Youden]

> Do you know by chance if there are similar companies (complex engineering, with physics, maths, chemistry, computer science, ...) that are located in Switzerland?

I've only worked with their products and not a whole lot with the people but Sensirion [0] is very close to Zürich and is ranked one of the best employers in Switzerland while seemingly being at the intersection of the things it seems you're interested in.

Switzerland is also full of manufacturing equipment suppliers (e.g. Schaublin and GF Machining Solutions, the entire "Watch Valley") and manufacturing is full of intersections between physics/maths/chemistry/software engineering.

Plus there's the pharmaceutical companies.

For a small country, I think Switzerland is about the best place for you to be.

[0] https://www.sensirion.com/en/

[jokteur]

Thanks for the tip.

[ChuckNorris89]

>Do you know by chance if there are similar companies (complex engineering, with physics, maths, chemistry, computer science, ...) that are located in Switzerland ?

AFAIK, Switzerland is full of such companies. Especially since CERN and many top universities and research institutes are there.

[jokteur]

Well, I am already beginning my PhD at a top university. I would be more interested to hear about what kind of jobs are in the industry in Switzerland.

[ChuckNorris89]

Sorry, I don't remember exactly which companies by name, otherwise I would have spelled them out for you, since I did this research 10 years ago when I was considering moving to study in Switzerland, so, I hope I cause you no offense, but I feel researching the job market is what you should have also done your due diligence on, before starting your PhD/studies ("I wanna study X, but where can I work later and use X?")

[jokteur]

No problems. You are right, it is always good to research the job market before beginning the studies. I was just asking out of curiosity, because the job descriptions at ASML really peaked my interested.

[4gotunameagain]

A lot of FLUKA radiation simulations at CERN if you like that !

The environment is nice, but the area personally didn't work out for me.

[tullatulla]

If you don't mind moving to Germany, check out Carl Zeiss (SMT).

We supply the optics to ASML and also make our own semi equipment directly sold to fabs. Lots of growth and interesting challenges at the cutting edge of physics.

[microtherion]

Synopsys still has an office in Switzerland, to the best of my knowledge, and this sounds like it might be along the line of work you're interested in.

[amelius]

This comment makes me wonder how much ML is used at ASML or if there is an growing trend ...

[CookiesNCream23]

Can’t tell you the exact details but you should be able to find details from the patents, papers, and tech talks from ASML.

[zogomoox]

I recommend you check out the ASML / Brion campus in Silicon Valley.

[throwawaybutwhy]

RUAG?

[traceroute66]

> Downstream effects of this fire will reduce the uptime of the machines and the delivery of promised machines to our customers.

Given that it is the factory acquired from Berlin Glas in 2020, surely its capacity would only "just" be coming online at ASML and they have pre-existing capacity elsewhere ?

[postingawayonhn]

It sounds like Berlin Glas was already supplying these components to ASML and ASML just decided to buy them out.

[aemreunal]

They may have sold that production with the assumption it'll come on, which will now be delayed.

[CookiesNCream23]

Can't talk about that :)

[Heliosmaster]

As a person whose first job after graduation was at ASML, and left after 6 months, if you like software I don't think it's the best company to work at.

[varispeed]

After working at ASML would you get enough knowledge to start competing company? Or is the job so protected with NDAs that basically you wouldn't be able to translate or use the gained knowledge anywhere else? I am thinking, for example, when you realise the working conditions are not great or you are paid not enough, do you have any leverage? If ASML has no competition, then do you have any negotiating power?

[traceroute66]

> After working at ASML would you get enough knowledge to start competing company? Or is the job so protected with NDAs that basically you wouldn't be able to translate or use the gained knowledge anywhere else?

Bit of a strange question ?

Competing with ASML in any sort of serious form would need more than just stealing some IP, it would need $$$ measured in the many billions.

ASML does have competitors, Canon and Nikon being the obvious two. But ASML has better technology than them, which is why continues to dominate the market.

ASML's model is not dissimilar to Apple's. Throw money at R&D like nobody's business, make some quality company acquisitions along the way, reap the rewards.

ASML basically have a 100% market share in EUV, 95% market share in ArF tools, 20% in Dry tools.

My limited understanding of the market is that basically ASML will be untouchable until at least the 2030's.

[nine_k]

To start a competing company, you'll need billions, and an idea of a market niche.

There are about two companies like ASML in the whole world, and there are maybe a dozen consumers of their machines, companies like TSMC or Samsung. ASML is in an effective monopoly position.

It's much like asking if you can open a competing company to build nuclear submarines. Technically you likely can.

[Cthulhu_]

And the accumulated knowledge of decades of lithography from hundreds, if not thousands of people.

That said, it's not impossible; SpaceX came out of left field and within a decade became one of the most prominent space rocket companies, they managed to find rocket scientists and engineers.

[ethbr0]

* came out of left field ... with the support of NASA contracts and knowledge sharing.

Which isn't in any way to take away from the pioneering things that SpaceX engineers have done (and the way they've done them!). But is to say they definitely had help in getting over the "high-capital-investment industry" starting bump.

[skrebbel]

There's non-competes but they're easily circumvented. I know of multiple people who started a sole proprietorship (ie contracting) and took a competing company (for some distant value of "competing") as their customer. Dutch labour law prevents most non-competes in practice.

I once freelanced at a startup that was building a machine that could compete with ASML's 30 year old machines. Those are still being used, resold and refurbished so one of those refurbishing companies (made up of ex ASMLers ofc) thought they could build new ones from scratch.

It worked (after a series of bankruptcies, restarts and acquisitions I might add - it was a rocky ride but not for technical reasons). That doesn't mean they will ever be able to catch up to ASML, or even intend to, but it's clearly possible for ex ASML people plus some young new engineers to build a lithography product. I bet with a few extra years patience you can do without the ex ASML people altogether.

This means that if eg China wants to have their own chip industry, and the US keeps preventing ASML from selling their best machines to China like they currently do, all the Chinese government has to do is encourage a few dozen lithography startups to happen, aimed initially at the low-end of the market. Add some protectionism (eg import tariffs on any chips that could be made by Chinese machines), wait a few decades, and done. I'm pretty sure it'll happen.

[CookiesNCream23]

Not a chance. The machines are so complex that nobody knows the whole thing. People dedicate their lives to one small component.

What I value is the opportunity to be see technology and scale of this caliber.

[testtesttestte]

Some people tried it and got sued for patent infringement (ASML vs XTAL)[1]. ASML patents all EUV-related technology, so you'd have to come up with a brand new way of generating EUV plasma at high energies at constant rate. Even ASML had to buy an American company Cymer to get a working prototype. Then, because EUV radiation doesn't pass through air or standard lenses, you'd have to come up with another optic column system. If you want to go 5nm and below, you need anamorphic lenses/mirrors with almost atomic precision [2]. ASML gets these columns and lenses/mirrors from Zeiss, but probably has patents on their practical application.

The sheer size of the supply chain you need to create to make a working machine is massive. It's one of the reasons why ASML got ahead of the Japanese companies that tried to do everything in-house.[3][4]

It took 17+ years and €6 billion in R&D to get EUV working [5] (as such EUV litography became feasible in the 90s), so ASML's leading position on the market is justified. They have to recoup all the costs, so they will make sure to keep the market leadership through all legal means.

My guess is that if someone would come up with a more efficient EUV plasma generation or better optic systems, ASML would acquire the company and take over the technology.

Regarding the negotiating power, some companies like GlobalFoundries couldn't afford ASML's EUV machines, so they parked their EUV plans [6]. It's only big players like TSMC, Samsung and Intel that can afford this tech and all these companies know their worth and prices. ASML needs money to pay for the R&D and can't afford to lose it's biggest customers that are also buying their non-bleeding edge technologies (DUV litography).

I don't know the business side of things, but it's definitely going to be interesting to see what happens with competition in the EUV litography in the upcoming 10+ years.

EUV litography and silicon-based chips have their physical limits, so there might be other ways to continue improvements.[7]

[1] https://www.reuters.com/article/us-asml-holding-xtal-court-i...

[2] https://semiengineering.com/extending-euv-to-2nm-and-beyond/

[3] https://ideas.repec.org/p/eti/dpaper/05007.html

[4] https://youtu.be/SB8qIO6Ti_M?t=673

[5] https://www.asml.com/en/products/euv-lithography-systems

[6] https://www.anandtech.com/show/13277/globalfoundries-stops-a...

[7] https://semiengineering.com/making-chips-at-3nm-and-beyond/

[zogomoox]

EUV and higher energy photons are difficult beasts, I would put my money on electrons next, so perhaps the electron microscope companies hold the interesting patents there.

[CookiesNCream23]

E beam lithography in credibly difficult. Recommend chapter 13 in "Principles of Lithography" https://spie.org/Publications/Book/2525392?SSO=1.

My opinions is that this whole comment discussion about creating a competitor is in the wrong direction. It is impossible to recreate and outperform what ASML has.

A better focus would be to tackle problems in the <2nm processes. If someone found a method to stop electron tunneling at smaller nodes, this will net much more revenue and create more value for society.

[DoingIsLearning]

I've seen some interesting positions from ASML but like many Dutch companies they seem to be office/lab centric. Do you know of any software teams within ASML that are remote friendly?

[ChuckNorris89]

Most semiconductor/electronics companies are not at all remote friendly, in the way you'd imagine, since, like you said, their core business is lab centered and focused on selling products you have to physically interact with, unlike the pure SW industry.

Some are however, remote friendly in the way that they do offer WFH options, but that means you live around the office/HQ for legal and security reasons and can WFH every now and then, with the agreemant that you have to show up at the office when required, not you get employed form anywhere in the world you happen to reside in and keep working from your living room, that's a no-go from the start for any established company in the semi field.

Also, most of them believe in a culture where shoulder-to-shoulder collaboration is key to development success and innovation, and having worked in the industry I have to agree. When the industry was forced remote kicking and screaming in 2020 "thanks" to Covid, morale and junior onboarding suffered a lot since the whole processes and culture was built on decades of shoulder-to-shoulder work where the only way to learn was, besides the necessary years in academia, to hang around seasoned graybeards and get your hands dirty with them vs self study with Googling the answers on stack overflow, so this process couldn't (management also didn't want to) suddenly uproot it and convert it to fully remote.

Basically, a material scientist, optics engineer or physicist would be at the core of the main money making products (lithography machines) and therefore be way more valued than SW devs which tend to be treated more as cost center in this industry (in general but can't speak for ASML), so if you like being at the core of the product and be treated like a rockstar, then I'd stay in the SW industry and avoid the semi industry all together (I worked 7 years as a dev in semi, before leaving it for good for the SW industry, but the graybeards I worked with in analog design with deep domain knowledge in RF were making bank and were pampered like rockstars, whereas SW and FW devs were treated like replaceable cogs that could be easily offshored without any losses).

[zogomoox]

Morale also suffered because of the push to implement Scaled Agile (SAFe) in the software teams.

[ChuckNorris89]

Yea, Agile in semi/HW industry is pure cancer. Innovations in this industry come as you let your engineers take their time to try out a million things that don't work in order to eventually, after months or even years, get to that one thing that does work, not by shoving them in the Agile meat grinder of daily standups, sprints, TPS reports, velocity charts, grooming meetings, etc.

Agile in HW only "works" when your company is a sweatshop, operating on low margins, bind by churning out commodity stuff in volume to demanding customers (i.e. Apple) on a strict cadence with the fear that if you don't deliver on time and under budget, then they go somewhere else, so you hire some Agile consultants to act as whip-crakcers and henchmen of management to make sure your devs are busy bees and ship, ship, ship.

I assume this is not the case with ASML (and maybe also with the likes of Nvidia, etc.) as they're the tip of the spear and so far ahead of the competition that they can take their time and charge however much they want and their customers have no other choice but to wait and pay up as there's nobody else to go to for competitive alternatives in the bleeding edge space.

[Someone]

agile with lowercase ‘a’ was defined for software development (https://agilemanifesto.org/), so you cannot blame it for not applying fully to hardware development.

The manifesto is extremely flexible, though (it doesn’t mention standup or fixed-length sprints, for example), so I think it can work for hardware, too, but you likely have to make the sprints longer and be more flexible in lengthening or shortening them to match reality, maybe do more documentation, etc.

I think ”half year sprints” loosely is how the Apollo program was run (https://en.wikipedia.org/wiki/List_of_Apollo_missions), certainly after Apollo 1, when they realized the schedule wasn’t the most important thing in the world.

https://en.wikipedia.org/wiki/Apollo_1#Program_recovery: “We were too 'gung-ho' about the schedule and we blocked out all of the problems we saw each day in our work. Every element of the program was in trouble and so were we.”

The dogma of Agile with a capital “A”, on the other hand, IMO, doesn’t work for anything, be it software or hardware.

[rektide]

Parent said Scaled Agile Framework, which is rather alike waterfall- top down management- with agile ceremonies.

[thrwyoilarticle]

When I was a grad at a similar tech company, I avoided alluding to it on HN for fear of appearing to be representing the company (after training mentioning this). Are ASML more forgiving?

[alexpetralia]

For those interested, I very much recommend this brief primer on ASML: https://youtu.be/jJIO7aRXUCg

[McBeige]

Sounds fascinating! What kinds of simulations?

[CookiesNCream23]

I am not an expert in this area. Most of it is physics simulations.

An example I know of is that we use feed forward control to make sure layers are aligned and printed with nanometer precision. Simulations are needed in this instance to model effects like thermal expansion, pressure waves, and much more.

There is also more work in metrology (measurement) with stuff like scanning electron microscopes, lasers, flow, etc.

You can find more about this by looking at the companies ASML has acquired or ASML's job postings.