I was really wondering why and how the USA government managed to gain so much control over what a certain company in the Netherlands does, this clears it up.
The EUV was originally developed by US, the early efforts on EUV were done by NIST and DARPA and then a consortium of US DOE labs ( Sandia, Lawrence Livermore, and Lawrence Berkeley National Laboratories) and the private sector was formed (Intel, AMD, IBM, Micron, Motorola, SVGL, Tinsley,....), and more importantly the EUV light source which is the main and the most complicated part of the EUV system is designed and built by an American company, so US gov still have cotnrol over both the technology and ASML.
The truth is Intel became the top customer of many US equipment companies in the late 1990s. They then used their PICOS (Program for the Improvement and Cost Optimisation of Suppliers) strategy to continually beat them on price. I remember stories of SVG engineers essentially reporting to Intel reps on their Wilton campus. In fact, Intel would cherry pick every lens and then demand more qualifications than any other customer. As a result, Intel killed many of its suppliers. SVG, Gasonix, PRI Automation come to mind.
That's why Texas Instruments never went with SVGL. I was a Field Service Engineer Starting with Canon in the 1970's, moved to Ultratech Stepper and then to SVG, I saw the many changes and was always concerned about my job. Retired now and enjoy not having to travel.
This kind of fits with my pet theory that high-end lithography is something the semi-conductor industry simply can only afford to develop once (i.e. at one vendor). If there are multiple vendors and only a handful of sizeable customers, and each customer chooses one or at most two vendors to do business with, then the vendors end up each serving a very small number of customers. The customers then have a large amount of power, and short-term gain will drive them to squeeze the vendor for all their worth. I think here maybe US tendency to keep technology to itself by restricting export actually hurts it, because it shrinks the set of potential customers for vendors, thus making the vendors more vulnerable to over-exploitation by customers.
@@SylphDSThe US doesn't really keep technology to itself. There's lots of tech export bans in the news but that's only part of the story. Another side of it is everything is handled in treaties and technology sharing agreements. And yet another side is how companies everywhere now days use patents to prevent competition instead of the original intent of being able to collect a fee from licensing the patent. Bans are diplomatic statements that say someone made us unhappy. No more, no less. We've too many open ended technology sharing agreements for import and export for them to mean what they sound like. Half the stuff we ban wasn't even invented in the US, sometimes US companies don't even make the thing we banned. It's all diplomatic posturing, except for the times the EU and US act in lockstep. If that happens then it's actually important. If that happens it means both got together to figure out how to punish someone. Otherwise the bans don't mean much. Chips are about the only thing the US has a true monopoly on, and the US has that not because of bans but because of patents. Intel, AMD, Nvidia, and Qualcomm don't share their patents with anyone but each other. No one else can break into the market, they're looking at decades of patents by four companies. And I'm aware Apple designs chips now and for reasons too arcane to go into I'm not counting them as a high end designer. They're mid tier that gets to pull a very very old trick that no one else gets to do anymore.
I remember working on old GCA wafer tracks in the 1980s. They used paper tape to load their programs. Really, paper tape! 5:13 I remember also working on the Perkin Elmer Micralign, though I'm sure they looked a bit different to the one shown here, probably earlier models. I had a lamp explode in one once, that ruined my day. Later worked on the Canon steppers, we had one pre-production early model and it was possible to get the wiring loom wrapped around the precision mirror assembly beneath the stage, guess how I found that out. The software was ghastly but generally the machines performed well.
Who ARE you? We used Micralign tools at my first job as a process engineer at Harris Semiconductor... in 1979. I thought I was the last of a vanished race...
This is a great channel. I've been working in the electronics industry for decades, as an end user for finished semiconductors. When reading the industry "rags" (EE Times, EDN, ...), the semiconductor industry, is, and has always been, the poster child of volatility. Even compared to nonsense, such as crypto. But its amazes me that our modern technological world is dependent on the products of such a volatile business !
The best part is that Nikon and Canon were denied by the US gov to be a part of EUV development program, not only ASML was able to convince the US gov to join the efforts but to buy SVGL and a bunch of other companies to become the sole player in game, (the initial semiconductor equipment manufacturers were; SVGL, ASML, and USAL).
If you do not like this empire, you are free to choose the Polonium Empire. Or the Total Videosurveillance Empire. Seriously, this world is very complex and fragile. It is much more than simple slogans. Think about preserving international security and stability. Don't despise what you have, considering that it could be much worse. Be humble, see the good things in this world. Smarter people than you and me are working hard to save what we have. Just destroying the old order does not make a better world.
Dear Asianonmetry, Thank you once again for another well developed and informative piece. Of course the irony is that the very same voices in the American political sphere that moaned about national security and loss of technology issues, were the exact same voices that promoted low taxes, small government, and the idea that business needs to support itself without State and Federal Tax Breaks or outright Funding to support that very same technology. Add to this the ever present ( and primary ) American focus for quarterly earnings, and we have the reasons for this story.
Exactly. This attitude is how giants like GE and Boeing were brought to their knees, and how we lost so much of our industrial base over the past several decades.
@@souvikrc4499, I agree. When the CEO starts to genuflect to Wall Street's demand for quarterly earnings rather than keeping the employees on top and focusing on customers, you get the destruction of once mighty companies like Motorola and Sun.
@fluffybunniesandkittens8818, if my memory serves me correctly, the driving motivation for SVG's executive group during this period was quarterly earnings. Rather than invest in their people or their technical roadmap, they sought to please Wall Street. Everything else you mention follows as a result of this failure to keep their eyes on the ball and do the things that their engineering staff indicated needed to be done, but to chase immediate reported earnings. Execution depends on what the execution is aimed at, and the American semiconductor equipment manufacturers all aimed at pleasing the bankers rather than running their companies and organic growth through innovation. See also IBM as the captain of the Good Ship Financial Engineering.
ASML had access to all U.S patents and technologies, Canon and Nikon did not as it was seen as a global competitor. Add in significant funding from several key major fabs and chip designers, ASML becomes well armed to lead the R&D to fully commercialise and scale the production of would-be EUV machines
In addition to the patent issues already mentioned, at least one of the two Japanese makers (I can’t remember which) bet the farm on DPP source tech rather than LPP. That was the wrong choice and they paid the price.
We always hear how free market America has ostensibly been since Reagan, yet there the federal government was trying to pick winners in the lithography business during and after him.
America only likes "free market" when America is dominating the market place. As soon as it no longer dominates the market, it turns to government control. Example: Huawei sanctions. Just like it only likes "democracy" when the democratically elected leader of a country supports American interest. If he doesn't, he gets toppled by CIA 😂
Jai Hinduja. Please produce a story to clear up Huawei chip foundry business. Canadian rip artists had declared that 9000s is 7nm finfet but Japanese reverse engineering experts think that it is a 14 nm chip.
SUSS aligners are way easier to make the programs and use but require more carefully maintenance specially on the production floor since been a 1to1 aligner the masks have to be constantly cleaned to avoid exposure defects and that's a huge time drain.
For the record, Torossian is an armenian surname. According to armenian sources I found, the "Der Torossian" family was "famously one of the larger and more influential Armenian families who fled to Syria during the Armenian Genocide of 1915-1917". However, I could not find definitive confirmation whether he is armenian or if he considers himself armenian; or if he merely has the surname as heritage from his father's line.
Raytheon, Santa Anna, CA, had a fab right next to and visible (glass) from the main design room. This was in '67 when fabs were very small and clean rooms were only used for CMOS (ESD).
Utratech Stepper is why I'm living in Texas. National Semiconductor was opening a large fab in Arlington Texas and they couldn't convince any of their California reps to move. I took the job and left Canon. Moved from New York to Texas. BTW, the National fab has since disappeared. All traces gone. Old fabs don't die, they just disappear...
@@chetk4624 ...especially National Semi fabs! LOL Go try to find any remainders of the old Bell Labs fabs in PA and you will not find much either. You can't repurpose a fab easily. I did see one once in Boca Raton where they had turned the Motorola facility into public streets and someone bought an old fab. We were building a Postal Facility where the old loading docks were.
There is, actually, one remaining American niche lithography tool manufacturer: Onto Innovation litho division in Boston, formerly Rudolph Technology, which is focused (ha-ha) on Package and RDL related steppers that compete with (mainly) Ushio in h/I-line tools. Of course, the technology is pretty low and not really competitive with Asian tools.More of their revenue comes from metrology & inspection tool division in Milpitas, where they compete with nearby KLA.
Thank you for this very informative tale of how lithography went to ASML! Does not bode well for the US semiconductor industry onshoring efforts. Too many failures indicate that its time has passed.
Thank you for your very fine videos! A small correction: The name is DerTorrosian , who by the way, later was (is ?) on the board of Zeiss. The DerTerrossians are very much Armenian though Papkin grew up in Syria before moving to the US on an MIT scholarship.
@@ivoryas1696 I could go on forever...One thing that impressed me as a young engineer was that some optical designs barely met spec in simulation and had zero or negative margins for manufacturing. Yet we went ahead and made them, and they did pass spec (too late for commercial success). I stayed in the precision optics industry, and have worked with practically all the Japanese, European, and US companies mentioned in the video. While each company has a distinct culture, there is much in common among people that make things on the hairy edge of possibility.
Perkin Elmer was originally funded by DOD . IBM funded PE thru the years for certain upgrades , the step and scan was a original PE project and heavily supported by IBM but when the last infusion of money did not yield a working prototype is when behind the scenes and other issues forced PE to go out of business in the semiconductor business and sell to SVG , ASML purchased SVG , got rid of most of SVG product line kept the step and scan that is now the base of their e uv machine . PE was the first contactless scanning tool. ASML originally was a Philips Electronics internal stepper that went out on it own , it also developed a its own step and scan 4x reduction lens based machine before acquiring SVG .
Fun fact: Perkin Elmer was also the company that fucked up the Hubble primary mirror. I remember a coworker of mine recalling he would bring it up constantly to his Perkin-Elmer FSE at Motorola just to get a rise out of the guy. It’s not often the US government has to launch a space shuttle just to fix someone’s incompetence. 😆
It is ironic to think that the breakthrough of Extreme Ultraviolet Lithography was the result of American financing and research done in labs in the United States and yet the beneficiaries turned out to be ASML and Carl Zeiss.
@@souvikrc4499 actually no. The power to control who ASML sells to through the EUV license is proving, at least so far, to benefit US national security interests, which was the plan all along. There wasn’t and still isn’t any company in the US with anything like the skills or capital to build out EUV infrastructure and produce tools.
Cymer and Applied Materials are critical suppliers to ASML. ASML relies on suppliers from Japan, USA, Germany for most of its important parts. @@Grak70
the US has a poisoned business culture, people who do good work clash too much with MBA monkies who cannot and are not legally allowed to think beyond the next fiscal year. It's a doomed project to bring any technology that requires long term vision to the US as a public company. Private companies can do it but the culture is still pervasive. MBAs have no tolerance for building businesses around important employees. In the US workers "must" be a commodity and the short sighted consequence, which the brilliant MBAs failed to see, is that work places must also become a commodity for employees. That means you can never have employees who are uniquely important to the company, which it turns out crushes long term plans. If you want everyone to be replaceable at any moment they have negative incentive to invest any of their efforts into long term plans. That is why any industry that relies on long term vision is doomed in the US, including lithography. The one exception is if it gets placed well within the military umbrella, where the government does think longer term.
Having just read: Chip Wars from Chris Miller and seeing the video which ended around 2000 it seems now ASML is the only company who has EUV lithography. The machines are sold or around 100 mio USD. Knowing that Zeiss Oberkochen is a very important vendor I was not aware that Trumpf Ditzingen with Laser is also a key supplier. It seems the next generation for improved EUV will be available (if it works) around 2025 and then cost most realistic 200 mio USD/machine. There was also a comment that ASML has to buy vendors if they are not able to manage improving and delivery. Seeing that Philipps has more or less gone but left some successors as ASML in so called Brainport Eindhoven the seeds they planted are still there. ASML is the key to keep Chinese vendors behind. Huawei was the 2nd largest customer of TSML. Not sure whats the role. It seems additionally to the allegations of IP etc risk, they were on. the way to become a technological leader. As Fredmund Malik is saying: 'More from the same' the current chip industry maybe can go ahead. But it's also possible that new technologies like quantum computing come up what. would fit better on the comment from Malik.
Regardless of how important you think your company or product is to consumers and industry, if you get too comfortable and fail to innovate and improve you will become obsolete and unnecessary overnight
So ASML gave the US control of EUV tech export to buy a company for over 1 billion so they could shut down the company within 2 years? A company that people thought wouldnt be competition in the future due to falling behind. Is it just me or am I confused why the deal was deemed useful to ASML and the Dutch?
At the time, it was seen as the only way ASML was going to force Intel’s hand to buy from them. I think if Intel had understood how unstable and hollowed out SVG was, they would have switched earlier.
they r buying geopolitics interest, very adorable of you think ASML can be as big as it is today with out the blessing of the american captial/tech/market. in a way u can say the same thing about TSMC which in the same position, all these r backed by US military and wall street. dont kid ur self think its just free market, thats for children.
Buy what we build and don't bother us 😂 yup ive troubleshooted a lot of systems and got the same response... then fixed it of my own volition. When the CVD machine got a couple extra shots of Et3Al i discovered it was that a nearby machine made some EMI that got into the valve controller and triggered it. The fix was simple enough, adding a few ferrite beads to the control cable. ❤
=I GUESS THAT'S BC OF LIKE .............. *_HOLLAND IS THE US COLONY_* ,SO WHY NOT ...........IN ACTUAL,THERE'S LIKE *_90%_* OF DECISIONS ABOUT THINGS LIKE THAT ARE POLITICAL
You don't talk about VC firms role in this. The financial "industry" in the US basically abandoned investment in hardware because they could make so much money in software.
I like the thumbnail. It's like cryptozoology for equipment. Is that a Emcore d180? ASM Episilon 2000? Bootleg wet bench or something completely unknown?
@@root42 I was commenting on the aesthetics of the thumbnail rather than the technical content. There are many better pictures of the tool, but instead a grainy, lo-fi one was chosen. This was a choice and one that I liked. It reminded me of the Patterson Bigfoot film because there is the potential to see what you want. Oh well, back to counting rivets.
is it really a failure to have allies lead an industry? the US leads the smartphone and OS market. no country needs to lead in everything. just focus on what you like.
When high-tech equipment breaks down frequently, that says something very important about the design and/or manufacturing of that equipment. They either cut corners, or they're simply not very good at design/manufacturing.
I don't think this is true. The US regime would NEVER allow crucial technology to be OUTSIDE of its political power structure. Hence, wherever the companies are located, the US regime has FULL CONTROL over that company/organization and can directly influence its commercial policies, i.e. which countries it is allowed to sell this technology to.
It is interesting how successive waves of industrialisation favours the newly industrial giants as 1. Government subsidies cover large costs and 2. Newer plant and equipment has advantage as do younger more recently educated and energised technical personnel. 3. Industrialisation started in the UK then Germany then US then a recovered Germany then Japan , then Taiwan and South Korea and finally China leap ahead . China has twice as many industrial process engineers than the rest of the world put together. This means Americas plans to stop Chinas technical advancement is futile . Companies like Apple have recent experience just how difficult it is to bring complex industrialisation back home , I think that Apple said that for a particular industrial process it needed there was 50 such highly skilled engineers in the USA and enough to fill a football stadium in China . Tesla said something similar . China's industrial technologies should continue to advance ahead of its competitors.
I dunno about better, but the scale was different - things were still accessible by craftsmen - nowadays it takes a bazillion dollars of equipment and a small army of craftsfolk of a dozen different disciplines even to design products, never mind manufacture them..
There was a lot of competition and variety back then. It was also an age of discovery of all kinds of things. The Internet was interesting back then, especially toward the end of the 1990s. Really miss that.
Curious when is it time make a video on the failure of American fab business. Maybe in another 5 years I guess. Intel, Samsung, both large brand name owners, cant give customers such as nVidia AMD peace in mind that their latest, cutting edge design secrets wont be leaked.
America basically owns ASML, a Dutch corporation and NATO alliance. But I do agree, the US needs to develop it's own. "Never put all eggs into a single basket" - Aristotle. Reply
The term "lithography" refers to a LOT more than computer chip manufacturing. In fact, the term originally had nothing to do with computer chips, and is still in use today to refer to offset lithography printing. Someone not knowing this would think a bunch of geeks invented the term, when in fact, it was invented over 200 years ago, or by the Greeks if you go back that far.
Can you please use Nvidia Broadcast to record? I don't know if I'm in the minority or the majority but sometimes the background noise is cosy and nice, and sometimes like in this video (especially in the beginning) is loud and distracting
It’s a PAS5500 body DUV scanner. The Illuminator optics train in the back is behind the projection optics, in the direction of the excimer laser unit peeking out behind it.
Intriguing. I did some research and found higher res images in a couple of ASML PDFs. "History & Future of TFH lithography" and "Chapter 4 1994 | 1998". In that image, the branding on the machine calls it a PAS 5500/700. You can also see the beam path more clearly in that higher res image.
@@Gameboygenius correct. 5500 refers to the body type. /xxxx is the model #. So for example a /100 is an i-line stepper with fixed pupil illumination and lower lamp power. A /275 is the same but with variable illuminator/pupil, better lens, and higher lamp power. Etc. A 5500/700 would be a single wafer stage body (limited to 200mm wafers max) containing a mid-range KrF exposure system.
I would like to see the US with a more creative industrial policy towards semiconductors, but if they won't listen to Gordon Chang they definitely won't listen to me.
i like how you try claim that ASML's success is a result of TSMCs choosing them. The fact is TSMC did not have a choice, ASML is the only capable machine supplier at cutting edge. If there was a better lithography manufacturer they would have chosen them. Remember the only reason why the US created TSMC was to try lower manufacturing costs and bring chips nearer to the low cost factories in china, with a secondary goal of trying to save the environment of the West at the expense of Asia's environment. They kept the high skill Chips manufacturing in NATO friendly countries (Taiwan mostly) while outsourcing the low skill manufacturing and assembly to china. This was first done in japan. and then moved to china. since china's manufacturing sector is now mature, they are increasing costs. the next manufacturing move will be either india/middle east/africa/south america. thinking that TSMC won out as a result of being superior is shortsighted. all the chips made at TSMC could easily have been made in the US. even more TSMC would not even exist if it wasn't for American companies who are their clients. Apple/Nvidia/Intel/AMD/Qualcomm (All American companies) are the real driving force and source of funding behind everything done at TSMC. If those American companies chose to, they could move everything stateside if they wanted to, but it would increase costs. I can prove all this easily: Look who dictates policy. Its the US.
At the time Nikon was pretty much as capable or even better but they decided to ride on the back of the success of Intel and pretty much like SVG they payed a price when Intel kick the off to the streets for ASML. ASML ride they way to top in the backs of a diversified clientele, Samsung, TSMC and SMIC that allowed their lithography machines to evolve faster than Nikon ones because client feedback.
Right now is a good time to buy litho tools as orders are cancelled and the loss of the Chinese market hits ASML, Nikon & Canon. US domestic politics is really hurting the IC Industry, particularly tool suppliers, and the loss of Chinese market share is likely to be permanent is things don't change soon. And I'm not talking about the pathetic waivers for Samsung and SK Hynix to save their memory fabs in China from bankruptcy (Yangtze is already ahead of them technically) but forcing Chinese fabs to roll their own tools will pretty-much kill the Japanese in the middle of the market. Thank Joe Biden, he seems to be screwing up just about everything now.
I was really wondering why and how the USA government managed to gain so much control over what a certain company in the Netherlands does, this clears it up.
Because US oligarchy controls IBM, Intel, US, ASML, Netherlands.
The EUV was originally developed by US, the early efforts on EUV were done by NIST and DARPA and then a consortium of US DOE labs ( Sandia, Lawrence Livermore, and Lawrence Berkeley National Laboratories) and the private sector was formed (Intel, AMD, IBM, Micron, Motorola, SVGL, Tinsley,....), and more importantly the EUV light source which is the main and the most complicated part of the EUV system is designed and built by an American company, so US gov still have cotnrol over both the technology and ASML.
If so why can't those American companies make their own lithography machines
@@yaoliang1580 Because it would be a financially foolish idea.
@@JigilJigil thks for info
The truth is Intel became the top customer of many US equipment companies in the late 1990s. They then used their PICOS (Program for the Improvement and Cost Optimisation of Suppliers) strategy to continually beat them on price. I remember stories of SVG engineers essentially reporting to Intel reps on their Wilton campus. In fact, Intel would cherry pick every lens and then demand more qualifications than any other customer. As a result, Intel killed many of its suppliers. SVG, Gasonix, PRI Automation come to mind.
That's why Texas Instruments never went with SVGL. I was a Field Service Engineer Starting with Canon in the 1970's, moved to Ultratech Stepper and then to SVG, I saw the many changes and was always concerned about my job. Retired now and enjoy not having to travel.
This kind of fits with my pet theory that high-end lithography is something the semi-conductor industry simply can only afford to develop once (i.e. at one vendor). If there are multiple vendors and only a handful of sizeable customers, and each customer chooses one or at most two vendors to do business with, then the vendors end up each serving a very small number of customers. The customers then have a large amount of power, and short-term gain will drive them to squeeze the vendor for all their worth.
I think here maybe US tendency to keep technology to itself by restricting export actually hurts it, because it shrinks the set of potential customers for vendors, thus making the vendors more vulnerable to over-exploitation by customers.
@@SylphDSThe US doesn't really keep technology to itself. There's lots of tech export bans in the news but that's only part of the story. Another side of it is everything is handled in treaties and technology sharing agreements. And yet another side is how companies everywhere now days use patents to prevent competition instead of the original intent of being able to collect a fee from licensing the patent. Bans are diplomatic statements that say someone made us unhappy. No more, no less. We've too many open ended technology sharing agreements for import and export for them to mean what they sound like. Half the stuff we ban wasn't even invented in the US, sometimes US companies don't even make the thing we banned. It's all diplomatic posturing, except for the times the EU and US act in lockstep. If that happens then it's actually important. If that happens it means both got together to figure out how to punish someone. Otherwise the bans don't mean much. Chips are about the only thing the US has a true monopoly on, and the US has that not because of bans but because of patents. Intel, AMD, Nvidia, and Qualcomm don't share their patents with anyone but each other. No one else can break into the market, they're looking at decades of patents by four companies. And I'm aware Apple designs chips now and for reasons too arcane to go into I'm not counting them as a high end designer. They're mid tier that gets to pull a very very old trick that no one else gets to do anymore.
Which nicely highlights one of the problems of SVG: inconsistency. They could never build exactly the same thing twice.
I remember working on old GCA wafer tracks in the 1980s. They used paper tape to load their programs. Really, paper tape! 5:13 I remember also working on the Perkin Elmer Micralign, though I'm sure they looked a bit different to the one shown here, probably earlier models. I had a lamp explode in one once, that ruined my day. Later worked on the Canon steppers, we had one pre-production early model and it was possible to get the wiring loom wrapped around the precision mirror assembly beneath the stage, guess how I found that out. The software was ghastly but generally the machines performed well.
What do i need to study to have experience like yours? 😊
Who ARE you? We used Micralign tools at my first job as a process engineer at Harris Semiconductor... in 1979. I thought I was the last of a vanished race...
This is a great channel. I've been working in the electronics industry for decades, as an end user for finished semiconductors. When reading the industry "rags" (EE Times, EDN, ...), the semiconductor industry, is, and has always been, the poster child of volatility. Even compared to nonsense, such as crypto. But its amazes me that our modern technological world is dependent on the products of such a volatile business !
Video on China's real estate bubble failure - WHEN?
this channel is a CCP soft propaganda. No surprise this dude is hiding his face
The best part is that Nikon and Canon were denied by the US gov to be a part of EUV development program, not only ASML was able to convince the US gov to join the efforts but to buy SVGL and a bunch of other companies to become the sole player in game, (the initial semiconductor equipment manufacturers were; SVGL, ASML, and USAL).
Free market is only when you have the advantage.
If you do not like this empire, you are free to choose the Polonium Empire. Or the Total Videosurveillance Empire. Seriously, this world is very complex and fragile. It is much more than simple slogans. Think about preserving international security and stability.
Don't despise what you have, considering that it could be much worse. Be humble, see the good things in this world.
Smarter people than you and me are working hard to save what we have. Just destroying the old order does not make a better world.
Dear Asianonmetry, Thank you once again for another well developed and informative piece. Of course the irony is that the very same voices in the American political sphere that moaned about national security and loss of technology issues, were the exact same voices that promoted low taxes, small government, and the idea that business needs to support itself without State and Federal Tax Breaks or outright Funding to support that very same technology. Add to this the ever present ( and primary ) American focus for quarterly earnings, and we have the reasons for this story.
Exactly. This attitude is how giants like GE and Boeing were brought to their knees, and how we lost so much of our industrial base over the past several decades.
Free trade agreements.
@@souvikrc4499, I agree. When the CEO starts to genuflect to Wall Street's demand for quarterly earnings rather than keeping the employees on top and focusing on customers, you get the destruction of once mighty companies like Motorola and Sun.
@fluffybunniesandkittens8818, if my memory serves me correctly, the driving motivation for SVG's executive group during this period was quarterly earnings. Rather than invest in their people or their technical roadmap, they sought to please Wall Street. Everything else you mention follows as a result of this failure to keep their eyes on the ball and do the things that their engineering staff indicated needed to be done, but to chase immediate reported earnings. Execution depends on what the execution is aimed at, and the American semiconductor equipment manufacturers all aimed at pleasing the bankers rather than running their companies and organic growth through innovation. See also IBM as the captain of the Good Ship Financial Engineering.
It's crazy how volatile the industry is. How did Cannon and Nikon not learn from the past to not lose out to ASML?
ASML had access to all U.S patents and technologies, Canon and Nikon did not as it was seen as a global competitor. Add in significant funding from several key major fabs and chip designers, ASML becomes well armed to lead the R&D to fully commercialise and scale the production of would-be EUV machines
In addition to the patent issues already mentioned, at least one of the two Japanese makers (I can’t remember which) bet the farm on DPP source tech rather than LPP. That was the wrong choice and they paid the price.
Canon actually is making some strides in the field atm. A great alternative for amsl for Asian countries
I think every bad decision catches up with the organization, and you can't win em all.
Its easy, the Japs only really iterated. AMSL Evolved
We always hear how free market America has ostensibly been since Reagan, yet there the federal government was trying to pick winners in the lithography business during and after him.
Not just in the lithography business. If Reagan followed his own free market ideology, the American automotive manufacturers wouldn't exist today.
America only likes "free market" when America is dominating the market place. As soon as it no longer dominates the market, it turns to government control. Example: Huawei sanctions.
Just like it only likes "democracy" when the democratically elected leader of a country supports American interest.
If he doesn't, he gets toppled by CIA 😂
Don’t know if America can be completely free market when its major competitor countries are not.
Reagan was a hypocrite in some cases
No way politicians would be hypocritical, don't be ridiculous
In 1999, 193 nano meter was cutting edge, just in 20 years we went to less than 8nm is astonishing!! 😮
And the machines to do it have changed from fridge size to container size :-)
Wonderful as always, I really look forward to you releasing new videos.
Very interesting on the history of how ASML became the monopoly they are today. Thank you Jon.
The term you were looking for is "Sole Survivor"
Video on China's real estate bubble failure - WHEN?
this channel is a CCP soft propaganda. No surprise this dude is hiding his face
"meanwhile in 1988"
This channel is amazing.
So much Facts 🤔. This can't be USA based.
I love the dry humor on this channel.
Jai Hinduja. Please produce a story to clear up Huawei chip foundry business. Canadian rip artists had declared that 9000s is 7nm finfet but Japanese reverse engineering experts think that it is a 14 nm chip.
SUSS aligners are way easier to make the programs and use but require more carefully maintenance specially on the production floor since been a 1to1 aligner the masks have to be constantly cleaned to avoid exposure defects and that's a huge time drain.
Not to mention you cannot make field term corrections if previous layers were exposed by stepper/scanner.
I don't know man, seems kinda sus
For the record, Torossian is an armenian surname. According to armenian sources I found, the "Der Torossian" family was "famously one of the larger and more influential Armenian families who fled to Syria during the Armenian Genocide of 1915-1917". However, I could not find definitive confirmation whether he is armenian or if he considers himself armenian; or if he merely has the surname as heritage from his father's line.
Based on his looks, I'd wager he is at least mostly of armenian heritage, though of course take my guess/"feelings" with a giant grain of salt
The DerTorrosians most definitely are Armenian !!!!
Papkin did flee to Syria and later got a scholarship to MIT.
The DerTorrosians most definitely are Armenian !!!!
Papkin did flee to Syria and later got a scholarship to MIT.
Raytheon, Santa Anna, CA, had a fab right next to and visible (glass) from the main design room. This was in '67 when fabs were very small and clean rooms were only used for CMOS (ESD).
Imagine how much has changed. On that level this video is correct. On an industrial intellectual property level this video is amateur.
Please do a video on the history of Ultratech Stepper! My dad used to work there and it's cool hearing some mentions of it in this video.
Utratech Stepper is why I'm living in Texas. National Semiconductor was opening a large fab in Arlington Texas and they couldn't convince any of their California reps to move. I took the job and left Canon. Moved from New York to Texas. BTW, the National fab has since disappeared. All traces gone. Old fabs don't die, they just disappear...
And don’t forget the tabloid worthy stories about Art Z!
@@chetk4624 ...especially National Semi fabs! LOL Go try to find any remainders of the old Bell Labs fabs in PA and you will not find much either. You can't repurpose a fab easily.
I did see one once in Boca Raton where they had turned the Motorola facility into public streets and someone bought an old fab. We were building a Postal Facility where the old loading docks were.
Mine too! First time I've heard the name mentioned anywhere!
There is, actually, one remaining American niche lithography tool manufacturer: Onto Innovation litho division in Boston, formerly Rudolph Technology, which is focused (ha-ha) on Package and RDL related steppers that compete with (mainly) Ushio in h/I-line tools. Of course, the technology is pretty low and not really competitive with Asian tools.More of their revenue comes from metrology & inspection tool division in Milpitas, where they compete with nearby KLA.
Their AOI tools are pretty good, especially if you can’t or don’t need to pay KLA prices.
Heck, I remember when transistors were the big thing in the '50s and how cool it was to have little transistor radio you could carry with you...
Thank you for this very informative tale of how lithography went to ASML! Does not bode well for the US semiconductor industry onshoring efforts. Too many failures indicate that its time has passed.
Another nice summary video.
Thank you for your very fine videos!
A small correction: The name is DerTorrosian , who by the way, later was (is ?) on the board of Zeiss.
The DerTerrossians are very much Armenian though Papkin grew up in Syria before moving to the US on an MIT scholarship.
I worked for GCA's optical subsidiary, Tropel, for 10 years, until the end of the General Signal era. What an amazing roller coaster.
vibrolax
That sounds _fascinating!_
I don't want to pry, but anything about the video to add? If not, that's understandable.
@@ivoryas1696 I could go on forever...One thing that impressed me as a young engineer was that some optical designs barely met spec in simulation and had zero or negative margins for manufacturing. Yet we went ahead and made them, and they did pass spec (too late for commercial success). I stayed in the precision optics industry, and have worked with practically all the Japanese, European, and US companies mentioned in the video. While each company has a distinct culture, there is much in common among people that make things on the hairy edge of possibility.
Interested to know if you'll be talking about Cannon's news soon. If their promises pan out, it could be really cool for the chip market.
Perkin Elmer was originally funded by DOD . IBM funded PE thru the years for certain upgrades , the step and scan was a original PE project and heavily supported by IBM but when the last infusion of money did not yield a working prototype is when behind the scenes and other issues forced PE to go out of business in the semiconductor business and sell to SVG , ASML purchased SVG , got rid of most of SVG product line kept the step and scan that is now the base of their e uv machine . PE was the first contactless scanning tool. ASML originally was a Philips Electronics internal stepper that went out on it own , it also developed a its own step and scan 4x reduction lens based machine before acquiring SVG .
Fun fact: Perkin Elmer was also the company that fucked up the Hubble primary mirror. I remember a coworker of mine recalling he would bring it up constantly to his Perkin-Elmer FSE at Motorola just to get a rise out of the guy. It’s not often the US government has to launch a space shuttle just to fix someone’s incompetence. 😆
It is ironic to think that the breakthrough of Extreme Ultraviolet Lithography was the result of American financing and research done in labs in the United States and yet the beneficiaries turned out to be ASML and Carl Zeiss.
We basically gave away our innovations in the name of short-term profit.
It isn't when you consider the people who did this research.
@@souvikrc4499 actually no. The power to control who ASML sells to through the EUV license is proving, at least so far, to benefit US national security interests, which was the plan all along. There wasn’t and still isn’t any company in the US with anything like the skills or capital to build out EUV infrastructure and produce tools.
Cymer and Applied Materials are critical suppliers to ASML. ASML relies on suppliers from Japan, USA, Germany for most of its important parts. @@Grak70
Great god bless you brother for always giving such massively great information.
No homo. I love you man!
the US has a poisoned business culture, people who do good work clash too much with MBA monkies who cannot and are not legally allowed to think beyond the next fiscal year.
It's a doomed project to bring any technology that requires long term vision to the US as a public company. Private companies can do it but the culture is still pervasive. MBAs have no tolerance for building businesses around important employees. In the US workers "must" be a commodity and the short sighted consequence, which the brilliant MBAs failed to see, is that work places must also become a commodity for employees. That means you can never have employees who are uniquely important to the company, which it turns out crushes long term plans. If you want everyone to be replaceable at any moment they have negative incentive to invest any of their efforts into long term plans.
That is why any industry that relies on long term vision is doomed in the US, including lithography. The one exception is if it gets placed well within the military umbrella, where the government does think longer term.
Nonsense. Just search for a better employer. And try to see the good things. Dont always look at the a$$hole of the elefant.
They are allowed to think beyond the next fiscal deadline. They just don’t.
Having just read: Chip Wars from Chris Miller and seeing the video which ended around 2000 it seems now ASML is the only company who has EUV lithography. The machines are sold or around 100 mio USD. Knowing that Zeiss Oberkochen is a very important vendor I was not aware that Trumpf Ditzingen with Laser is also a key supplier.
It seems the next generation for improved EUV will be available (if it works) around 2025 and then cost most realistic 200 mio USD/machine. There was also a comment that ASML has to buy vendors if they are not able to manage improving and delivery.
Seeing that Philipps has more or less gone but left some successors as ASML in so called Brainport Eindhoven the seeds they planted are still there. ASML is the key to keep Chinese vendors behind. Huawei was the 2nd largest customer of TSML. Not sure whats the role. It seems additionally to the allegations of IP etc risk, they were on. the way to become a technological leader.
As Fredmund Malik is saying: 'More from the same' the current chip industry maybe can go ahead. But it's also possible that new technologies like quantum computing come up what. would fit better on the comment from Malik.
Thanks again!
Great update, right on point with current talks. Its risky business for the west today but may be worth trying, even if we fail...again, and again.
Regardless of how important you think your company or product is to consumers and industry, if you get too comfortable and fail to innovate and improve you will become obsolete and unnecessary overnight
Very interesting ! So much happened in the 1990's.
So ASML gave the US control of EUV tech export to buy a company for over 1 billion so they could shut down the company within 2 years? A company that people thought wouldnt be competition in the future due to falling behind.
Is it just me or am I confused why the deal was deemed useful to ASML and the Dutch?
They bought some customers.
At the time, it was seen as the only way ASML was going to force Intel’s hand to buy from them. I think if Intel had understood how unstable and hollowed out SVG was, they would have switched earlier.
they r buying geopolitics interest, very adorable of you think ASML can be as big as it is today with out the blessing of the american captial/tech/market. in a way u can say the same thing about TSMC which in the same position, all these r backed by US military and wall street. dont kid ur self think its just free market, thats for children.
I'm in the mood for a bit of the old ultraviolet.
PE micraalign are.still used on non high end products like rf filters, they are very temperamental equipment.
Can you do a video on Ultratech Stepper ?
ASML depends on so many companies in the USA to make all the pieces of the machines so it’s relationship with the USA is very important.
The netherlands depend on NATO cooperation, in many ways.
@@frankgerlach4467 very true
Do you have a link to Craig Addison? I can't find him anywhere.
Can you please make a video about LAM research and LKA?
"Lam Research"- it was named after its founder, David Lam, so the name isn't an acronym.
You shudda given brief overview of Lithography
Paying tribute
Can you do a dive on ultra tech?
Which one? The Australian lighting company, or the Indian cement company, or...
Buy what we build and don't bother us 😂 yup ive troubleshooted a lot of systems and got the same response... then fixed it of my own volition. When the CVD machine got a couple extra shots of Et3Al i discovered it was that a nearby machine made some EMI that got into the valve controller and triggered it. The fix was simple enough, adding a few ferrite beads to the control cable. ❤
My go to was almost always MORE C. Or LC filters. Beads sometimes.
Didn't know that Elmer perkin was in that business, have seen the name on big pcb on a concurrent computer.
Do you have any interest in doing a video about CNC machines or company like Hans
They were talking about such small amount of money for the rescue! What on earth!
=I GUESS THAT'S BC OF LIKE .............. *_HOLLAND IS THE US COLONY_* ,SO WHY NOT
...........IN ACTUAL,THERE'S LIKE *_90%_* OF DECISIONS ABOUT THINGS LIKE THAT ARE POLITICAL
Not only is it a US Colony but we make them pay tribute for the privilege.
i can't find craig addions video is there a link?
You don't talk about VC firms role in this. The financial "industry" in the US basically abandoned investment in hardware because they could make so much money in software.
I like the thumbnail. It's like cryptozoology for equipment. Is that a Emcore d180? ASM Episilon 2000? Bootleg wet bench or something completely unknown?
Says in the video: Micrascan V 193nm
@@root42 I was commenting on the aesthetics of the thumbnail rather than the technical content. There are many better pictures of the tool, but instead a grainy, lo-fi one was chosen. This was a choice and one that I liked. It reminded me of the Patterson Bigfoot film because there is the potential to see what you want. Oh well, back to counting rivets.
Sounds like a a similar situation would be for nano graphene circuit production today.
I have literally never heard it pronounced "nee-kon"
That's how it's pronounced - ニコン. 'ニ' is Romanized as 'ni' but its sound is more like the word 'knee'
謝謝！
Hopefully not the end, just a hiatus of a few decades
{]*;😊;*[}
In those days, Wall street hated companies that made things.
Nothing earth shattering, just like the automotive industry, we are seeing American dropping out of a specific industry and let others pick up.
1:44 more better?
Look who this is coming from
canon ?
Do you ever talk about Lam Research? Some of the things they have done over the last 20 years has been revolutionary. STI. 3D nand. Etc.
Capitalism drives innovation but it pays no attention to national borders. Another great video Jon!
But it does depend on government investment.
capitalisms fatal flaw is greed
Beggining of Bring back jobs to America
its not the end, its coming back real soon
is it really a failure to have allies lead an industry? the US leads the smartphone and OS market. no country needs to lead in everything. just focus on what you like.
Bingo. NATO+JP+SK+ANZAC is actually the greatest empire that ever existed. It does have its challenges, but definitely also strengths.
So who won lithography, the world?
A very heterogenous society has always less competitiveness over homogenous societies.
#belllabs #seimens #iam How can you save these learning technologies for the next generation?
#fieldMuseum
Unfortunately, it's complicated.
When high-tech equipment breaks down frequently, that says something very important about the design and/or manufacturing of that equipment. They either cut corners, or they're simply not very good at design/manufacturing.
Knee con? LoL
I don't think this is true. The US regime would NEVER allow crucial technology to be OUTSIDE of its political power structure. Hence, wherever the companies are located, the US regime has FULL CONTROL over that company/organization and can directly influence its commercial policies, i.e. which countries it is allowed to sell this technology to.
you can always go back to Chinesium, if you like. Enjoy the abacus !
I just so doubt the conclusions of this video.
Super!!!😊😊😊
The end? Try again.... There are NUMEROUS fabs in the US today and more being built.
It is interesting how successive waves of industrialisation favours the newly industrial giants as 1. Government subsidies cover large costs and 2. Newer plant and equipment has advantage as do younger more recently educated and energised technical personnel. 3. Industrialisation started in the UK then Germany then US then a recovered Germany then Japan , then Taiwan and South Korea and finally China leap ahead . China has twice as many industrial process engineers than the rest of the world put together. This means Americas plans to stop Chinas technical advancement is futile . Companies like Apple have recent experience just how difficult it is to bring complex industrialisation back home , I think that Apple said that for a particular industrial process it needed there was 50 such highly skilled engineers in the USA and enough to fill a football stadium in China . Tesla said something similar . China's industrial technologies should continue to advance ahead of its competitors.
I guess we should all bow down to communism, eh ?
Communism invests in education. It pays off. No school shootings. Each talent is considered an investment opportunity for the government.
Weird, how everything has become boring nowadays, 1980s and 1990s were interesting times, quite a few things were better back then.
I dunno about better, but the scale was different - things were still accessible by craftsmen - nowadays it takes a bazillion dollars of equipment and a small army of craftsfolk of a dozen different disciplines even to design products, never mind manufacture them..
There was a lot of competition and variety back then. It was also an age of discovery of all kinds of things. The Internet was interesting back then, especially toward the end of the 1990s. Really miss that.
Curious when is it time make a video on the failure of American fab business. Maybe in another 5 years I guess.
Intel, Samsung, both large brand name owners, cant give customers such as nVidia AMD peace in mind that their latest, cutting edge design secrets wont be leaked.
America basically owns ASML, a Dutch corporation and NATO alliance.
But I do agree, the US needs to develop it's own.
"Never put all eggs into a single basket" - Aristotle.
Reply
Biggest flaw in your video is you do not explain what lithography is in the beginning of the video making assumptions with the audience
Wake up people rip USA
Capitalism is brutal
The term "lithography" refers to a LOT more than computer chip manufacturing. In fact, the term originally had nothing to do with computer chips, and is still in use today to refer to offset lithography printing. Someone not knowing this would think a bunch of geeks invented the term, when in fact, it was invented over 200 years ago, or by the Greeks if you go back that far.
Can you please use Nvidia Broadcast to record? I don't know if I'm in the minority or the majority but sometimes the background noise is cosy and nice, and sometimes like in this video (especially in the beginning) is loud and distracting
Good, mickey mouse are loooozer
:(
@3:24 WTF is that machine? It looks like an ASML PAS step-and-scan system, but the configuration gives the impression of a 2-perspective illusion
It’s a PAS5500 body DUV scanner. The Illuminator optics train in the back is behind the projection optics, in the direction of the excimer laser unit peeking out behind it.
Intriguing. I did some research and found higher res images in a couple of ASML PDFs. "History & Future of TFH lithography" and "Chapter 4 1994 | 1998". In that image, the branding on the machine calls it a PAS 5500/700. You can also see the beam path more clearly in that higher res image.
@@Gameboygenius correct. 5500 refers to the body type. /xxxx is the model #. So for example a /100 is an i-line stepper with fixed pupil illumination and lower lamp power. A /275 is the same but with variable illuminator/pupil, better lens, and higher lamp power. Etc. A 5500/700 would be a single wafer stage body (limited to 200mm wafers max) containing a mid-range KrF exposure system.
I would like to see the US with a more creative industrial policy towards semiconductors, but if they won't listen to Gordon Chang they definitely won't listen to me.
How is a creative policy going to help if the companies can't make their machines work?
Rip the USA
i like how you try claim that ASML's success is a result of TSMCs choosing them. The fact is TSMC did not have a choice, ASML is the only capable machine supplier at cutting edge. If there was a better lithography manufacturer they would have chosen them. Remember the only reason why the US created TSMC was to try lower manufacturing costs and bring chips nearer to the low cost factories in china, with a secondary goal of trying to save the environment of the West at the expense of Asia's environment. They kept the high skill Chips manufacturing in NATO friendly countries (Taiwan mostly) while outsourcing the low skill manufacturing and assembly to china. This was first done in japan. and then moved to china. since china's manufacturing sector is now mature, they are increasing costs. the next manufacturing move will be either india/middle east/africa/south america. thinking that TSMC won out as a result of being superior is shortsighted. all the chips made at TSMC could easily have been made in the US. even more TSMC would not even exist if it wasn't for American companies who are their clients. Apple/Nvidia/Intel/AMD/Qualcomm (All American companies) are the real driving force and source of funding behind everything done at TSMC. If those American companies chose to, they could move everything stateside if they wanted to, but it would increase costs. I can prove all this easily: Look who dictates policy. Its the US.
At the time Nikon was pretty much as capable or even better but they decided to ride on the back of the success of Intel and pretty much like SVG they payed a price when Intel kick the off to the streets for ASML.
ASML ride they way to top in the backs of a diversified clientele, Samsung, TSMC and SMIC that allowed their lithography machines to evolve faster than Nikon ones because client feedback.
In a way, cost and existing infrastructure is more important than merely the technology or patents.
lithography doesn't matter.
calligraphy is what matters!
america has many excellent calligraphers.
NATO invented technology to manipulate single atoms. How do you like this ?
"More better"? Really? C'mon man, do better
Jai Hinduja. The failure has not ended US chip dominance. But Canon will be rising up with nano printing lithography.
Hello snitches
🤣😂😅🤐
Yes, hello. Are you here for your stitches?
First
😏
ｐｃｂ需要移出中國
The 1.6 billion is just enough for the expenditure of American workers getting their daily fixs. A day Americans stay sober is a day wasted.🥳🥳🥳🥳🥳
Addiction is a symptom of PTSD.
@@msimon6808 Why are so many Americans obsessed with getting stoned and getting laid?
Right now is a good time to buy litho tools as orders are cancelled and the loss of the Chinese market hits ASML, Nikon & Canon. US domestic politics is really hurting the IC Industry, particularly tool suppliers, and the loss of Chinese market share is likely to be permanent is things don't change soon. And I'm not talking about the pathetic waivers for Samsung and SK Hynix to save their memory fabs in China from bankruptcy (Yangtze is already ahead of them technically) but forcing Chinese fabs to roll their own tools will pretty-much kill the Japanese in the middle of the market. Thank Joe Biden, he seems to be screwing up just about everything now.
Chinese Rah Rah ?
And the Republicans won’t go further in this direction?😂
"Nee con"? Baffling. Neither the American pronunciation ("nigh con") nor the Japanese ("nee cone").
I'm worried about China
No worries!
CCP China is corrupt and incompetent!