Did Sematech Save the American Semiconductor Industry?
Vložit
- čas přidán 17. 01. 2024
- This episode is an audio only episode
Links:
- The Asianometry Newsletter: www.asianometry.com
- Patreon: / asianometry
- Threads: www.threads.net/@asianometry
- Twitter: / asianometry
World’s hardest working CZcamsr. I’m not sure how you manage to create these incredible videos in such rapid succession. I hope the thousands who watch your videos donate generously.
As a patreon (which I assume you must be too to comment so early), I think this may be the most valuable channels to patreon to. Even if a person just watches the videos at normal releases, I think this channel is a very worthwhile channel to subscribe to if you care about anything relevant to the world today, or even a few decades ago. The breadth (of the channel) is astonishing.
His history videos are all very well done. Those are my personal favorites, though silicone wafer manufacturing processes can be interesting in it's own right.
Yes, Jon is truly the "James Brown" of telling the tale of (mostly) modern technology.🌏
@@ebx100 Agree! and thanks for referencing James Brown. I always think of that when confronted with someone who is clearly the "hardest working" (man) in (whatever).
2….months ago??? 😨
Honestly, pulls out stories which everyone forgot and tells them without any noticeable bias. A true internet treasure.
So true. I lived through this time - and am amazed at the breadth of his knowledge/recollection.
I was at Harris during this time and was at many Technical Advisor Board meetings, so I had good insight into what was going on. The loss of lithography by GCA was largely due to the lens from Carl Zeiss AG in Germany. Nikon and Canon had overtaken their capability and no amount of money could fix that. There were several US litho manufacturers, but almost all depended on the same lens manufacturer. Hence a predictable result. The exception being SVG which was adsorbed by ASML. At the time the largest companies like IBM would buy equipment and remake it to fill their needs. This was expensive to even IBM and not sustainable. All the while the large semi companies had internal investments with all the non GCA stepper manufacturers because they all realized they could not risk a GCA failure. It was a turmoil for sure. Nikon, Canon, and ASML all realized they had to supply the US market with the latest litho or they would not survive. So we have the outcome we see today. It was all very complicated for sure.
Wild idea for vid.
Old Aerospace Coot here, seem to recall a memory supply panic 20 or 30 yrs ago.
Seems the leading memory suppliers were competing for market share, and memory reached the threshold or radiation sensitivity where they all need organic epoxy that was free of carbon-14 to reduce SEU (single event upset)
Turns out all the memory suppliers were packaging their chips in epoxy that was carbon-14 free made by a single supplier. That single factory suffered a fire or something that simultaneously knocked out ALL memory suppliers. Thus a sudden spike of several X in price and lead time.
Thus a sudden shortage of memory from
All suppliers.
I don't know how you would make epoxy that's free of carbon 14 virtually all epoxy would be made from oil or coal
@@weedmanwestvancouverbc9266 There was something with radioactivity. I remember a story about lead (which is all slightly radioactive after atomic bomb tests). But epoxy also rings a bell here.
SEUs from carbon 14? That's a new one. I've heard about the issue of radioactive contamination in ceramic packaging but I would not have suspected that the tiny amount of C14 in regular plastic would be a significant issue.
Another couple of comments on the VHSIC Program. One of the most prolific work products was VHDL, VHSIC Hardware Description Language, which finally consolidated the programming language for FPGAs. The US Government was fed up with having to redo programming when moving a design to multiple companies and demanded some compatibility, hence VHDL was adopted by all the major players. Another outcome was the improvement in ebeam litho tools used then to make 500nm leading edge litho on wafers with direct ebeam writing. These improvements in hardware and software had a direct impact on the same tools used for making lithography masks used by all the stepper and scanner users. This was a subtle but important indirect outcome of the VHSIC Program.
Great episode, lot's of interesting things to discuss. Just one thing. NASA is a civilian agency independent of the DOD and although it evidently collaborates with the military and has a strategic mission it can be argued that it serves civilian and scientific purposes as a priority.
Gil Amelio deserves a little more love with regards to Apple. He’s the CEO who brought Jobs back (bought NeXT) and started some of the directional shifts that ultimately helped to right the ship. He never would have been able to do what Jobs did, but it seems to me that he was a net positive for Apple, which Apple hadn’t had at CEO in quite some time.
(But I understand, that is a tiny detail in this video.)
Anybody would have been better than Scully or Spindler. A Vic-20 would have been better than Spindler and an old shoe would have been better than Scully.
"During Amelio's tenure Apple's stock continued to slump and hit a 12-year low in Q2 1997 that was at least partially caused by a single sale of 1.5 million shares of Apple stock on June 26 by an anonymous party who was later confirmed to be Steve Jobs" -Wikipedia. If it can be said Amelio "brought Jobs back", it wasn't totally on purpose!!! Jobs undermined him to take over again as CEO (If you haven't figured it out, Jobs was not a nice guy, and I say this as someone who looked up to him as a hero once). Amelio also wanted to buy the Be Operating System, but Jean-Louis Gassée demanded $275 million. Which perhaps sounded outrageous, but Apple ended up buying NeXT for $400 million. Maybe he deserves more credit, but in the process Apple would have to admit how much of a sociopath Steve Jobs was (and just wait until people can freely tell stories about Cook!)
lol as if it was his idea? He ram apple into the dirt and handed it back broken
@@Peter_S_ 🤣
Thank you for making your videos completely listenable without requiring us to see the visuals to follow what's happening! It helps a ton for us peeps working away and listening to your work. ❤!
You should make an episode about Seymour Cray or Cray research. :)
And put it a general term : US Supercomputers (Cray,IBM,Convex) vs Japan Supercomputers (NEC,Hitachi,Fujitsu).
I've learned so much about the industry, especially the fab side, since subscribing and I've been in and out of IT for 40+ years.
Excellent video (as always); it is strange to see events from my working life as 'ancient history'. I worked at IBM Semiconductor when Semitech started up.; providing some folks with career boosts. Internally IBM was petrified by Japanese equipment companies, limiting Japanese purchases even if the tool was better. At that time I was responsible for CD/Defect Inspection SEMs, an area where Hitachi was preeminent (due to their Field Emission TEM experience). The tooling team kept bringing in North American alternatives that failed their own qualification standards. Pretty expensive to buy a tool, install it in a FAB, and rip it out again. I finally got a useful machine by picking up a surplus Hitachi SEM from the Yorktown Lab (one of the few bright spots in my time at IBM).
Around the year 2000, I visited Tool Makers around the world during the world's first practical application of GEM300 at Renesas in Japan. SEMATECH had a solid lead. The changes were so big that even they, the persons in charge of AMAT, were flustered😆
Anyhow, I admire the very detailed information you always provide. Thank you very much! !
Semiconductors are hard and the production needs to be given national/strategic importance, since relying on market forces is just not enough to maintain leadership on the market. And government budgeting plays an important role here.
Explain to someone not as well versed why pure market forces can't sustain this kind of industry?
@@AKK5I well, just look at what happened to any the semiconductor companies that disasapeared, failed or never got off the ground because they never got contracts from the defense industry, never got to be part of a consortium, never received (generous) government grants, or weren't propped by a Chaebol orwere not otherwise vertically integrated. Market forces mean there will be ups and downs. Being able to develop semiconductors fast enough to be able to put them on the market requires vast investments - and if you listen carefully to the video, Semtech, a consortium, received generous grants to be able to stay ahead of the curve. And as far as the vertical integration or support from the Chaebol in Japane and Korea are concerned - those investments come with lots of govt subsidies and tax breaks. Or look at Taiwan, semiconductors are key to their security and sovereignty and the government provides generous grants, subsidies and tax breaks. Let's not even mention the US chips and sciences act with $40bn available and the EU Chips strategy is trying to catch up with over €40 billion up for grabs as well. Chips can't be made without public subsidies and grants. Period.
Cost sensitive, zero sum game industry@@AKK5I
he's probably suggesting that the risk component in terms of volume fluctuations guaranteeing a return to scale.
an interesting theory which may have some grounding but I'd need to better understand when the disreturns to scale occur in a typical fab. and also why speculation/normal financing couldn't be a way around this.
@@classiclibertarian Shareholders want stability and reliance. It's just a natural facet of finance in general.
Pure market forces can handle this, but the only way to do it is if the ones manufacturing them are also producing the products they go in.
The real problem here is that we use semiconductors way more than we really need to which causes a general lack of competition and over use of semiconductors in worthless consumer goods that don't need them. There's no reason you need semiconductors in your kitchen appliances or really even cars, we just mae it that way.
I'm surprised Sematech still exists
My graduate research was funded by Sematech. Never knew the details of the consortium. Thank you!
I don’t think most understand the prolific impact Sematech had on the entire fab tool industry. Many leading edge tools had horrible uptime and costly maintenance. Their productivity and cost were a major hit to profitability. Sematech changed all that by introducing methodologies and tech transfers of tool technology directly related to productivity. The bar was raised and equipment manufacturers had to adopt or were left behind. Before Sematech it was not uncommon to see 50% uptime. The expectation was raised and achieved to high 90% over a very short timeframe. This matured the worldwide fab tool industry and was a profound impact as most of the fab tool OEMs sold internationally.
That is an insanely large building, at 33:33. I didn't realize how large it was until I saw the trailers next to it, to give me a sense of the scale. I can't tell if it's an actual photograph, or just a conceptual rendering of an upcoming build. Cool art on the side, regardless.
Well done as always!
Awesome work, as always!
7:58 great video as always,, also good looking out making sure your car is locked
Another fantastic video. I'm always excited to watch and read your content. Keep up the great work!
The Japanese dominated DRAM and SRAM for many years. Tj_max was key to reliability of military electronics. It was typically 105C, allowing margin for 60C chassis temperature.
Wow again. Great episode. Valuable to the world. Shorter than a book, a movie - this is how lessons should be and rewatchable to really get it
thank you for your work, you excel in everything you do!
The best thing that happened to Sematech and the worst thing that happened to DEC was that Palmer stayed at DEC as CEO.
You can definitely tell that Taiwan learned a lot of lessons about how not to piss off the Americans from looking at how Japan fumbled in this period.
It helped that Morris Chang was a Texan long before he was Taiwanese. (I think his family fled directly from the mainland to the US after the CCP won the Chinese Civil War.) Though the disaggregation of the industry into pure-play foundries and fabless chip designers seems to reflect a broader Taiwanese insight into Americans’ #1 paranoia: that cheap suppliers will one day become competitors. But the US will happily let you climb up the value chain, so long as you have a slightly different view of what the top is and stay in your lane-e.g. Nvidia vs TSMC. See also: the Apple/Foxconn relationship.
Loving the videos. Thanks A
Another fascinating video.
‘Or… listening I guess’ - busted
This is better than TV
But I do miss futzing with the fine tune to get the best possible picture without too much sound buzz!
I love his delivery. A pleasure to listen to. The subjects of his videos are very absorbing.
Great thumbnail photo
another excellent video
Would've been interesting had Noyce lived to see how, or if, his influence continued.
I still find it funny the Reagan admins didn't like the Sematech plan. Old people and tech, especially administrators simply does not work. This too is true of the EV mandate in 2034
THANKS
An illustration of the value of open standards, even to fierce competitors.
It is just as well they didn' ttry to hang on to a declining commodity market like DRAMs. Markets like that inevitably degenerate into an arms race whose survivors can hope at best for a modest rent on their capital, if they're smart and lucky. Major profits depend on "moats", whether those are location, cultural advantages, or difficult skills.
Light speed uploads
Thanks
What a terrific overview of an important time in semiconductor history. Imagine a world where Reagan and his objections to "industrial policy" never came to power. The US might well still be a world leading semiconductor manufacturing country instead of that leadership moving to Taiwan and Korea.
'NASA a military organization' ? Huh ?
I wondered how much industrial espionage and forced technology transfer was involved. Japan couldn’t say no to the US back then like Samsung and TSMC couldn’t say no now
i like the show and how the channel grows, the way you address topics is appreciated, cheers for the free content, here's a comment
Just FYI, 3:15 NASA is not a military organization.
@@notanymore9471 It was specifically formed to be a civilian organization. Google is your friend.
Great video - I have to thank Asianometry for turning me on to Charlie Sporck's book "Spinoff". I loved that book and the videos this channel has done on it.
Wow, this video must have been a LOT of work! As another commented, I'm absolutely amazed by the amount of content you're able to output, but even more impressive, is the quality of said content. While I know it is kind of a personal question, is the channel and your supporters garnering a decent income or are you essentially still in the growth stages of creating your channel and trying to establish a viable income? I would like to support the channel but I'm not really in a position to donate on a monthly basis. However, I can make a one-time moderate donation if the channel is really in need. I can't stand the thought of losing the content you provide to the community. You really are truly a unique resource.
I still think that there's just not enough focus on MFG'ing as there should be.
Audio only? But there is a visual component?
Best documentary content on this site
"military organizations like NASA" some 45yo kids in the back are crying now Mr. Nometry
Some older kids like me are just saying, "well, that's an error."
It is pronounced "D-RAM", as in two syllables and not one. Great Video!
You found a nit and chose, nay, picked it.
Jon gets this comment each time he uses the Dram…
Since he is the expert… I’m going with the way he says it…
Go Asianometry!
😃
@@AC-jk8wqhis way is wrong, though.
@@peterfireflylund it’s like faulting a guy for having an accent…
The message that Jon delivers each time is tremendous…
If that was the only error… Yay!
Maybe… he is trying to improve the world’s efficiency in typing and saying D-RAM…
Wonder how he says CD-ROM…. 😃
His father probably is known for being first saying CDs and ignoring the rest…
No matter how he says it…. We all know what he means….
Fortunately, it’s not in Cantonese, mandarin, or Japanese…. 😃
It would be nice to learn about TEL, Renesas and Rohm, alongside the semiconductor business of Canon and Nikon. You've done a lot of historical videos on American firms, but not too many about semiconductor firms outside the US.
Might be easier to get info about US firms. Potentially language barriers etc. with others.
How is an audio only episode different from the other videos? I can't notice any change, so why the title?
In other videos there's:
Transcript
Follow along using the transcript.
and then a button labelled:
Show transcript
I mean, apparently you don't notice very much!
NASA was a military organization? That is new to me! Time Code at 316
I have to agree with the vast majority of comments here. Your channel is pretty amazing. Thank you
Danke!
Would you consider doing a video on the Japanese atomic weapons program during World War II? I was surprised to find out that at least had a fledgling nuclear program during the war:
NOYCE!
Bailout and propup!
The corporate welfare system!
.but hey lets just bitch about individuals!
Superior AGI AI wins the AI WAR or WWWIII or Technological Singularity #singularity #AGI #AI #ML #DL #ChatBot4 #OPENAI #WWWIII
Watch "Princes of the Yen" documentary and read the book by R A Werner
I am curious how do you do the research for this, how much is personal knowledge? One problem with education is that you are not explained the history of the field, it is considered useless, such as the history of the semiconductors. Did you ever contemplate writing a textbook for colleges with thiese videos?
One insurmountable problem with Sematech is the USA's mindset to maximize investor returns a quickly a possible. This just isn't practical in an industry which has immense complexity that must be developed over years, if not decades of work that may not produce any profit at all. In my career of data storage technology with both American and Japanese tech companies, the difference between the mentality of the American vs. Asian company decision makers was truly night vs. day. With obvious results
Law of headlines strikes yet again.
Yuotubers never learn...
3:15 “… military organizations like NASA …”
Huh? NASA would be a bad example of a military organization. The USAF might have been a better example to use, unless that statement was supposed to be facetious
a lot of NASA's funding had strings that had to do with military. The size of the Space Shuttle's bay for example. It was made totally too large and unwieldy for the Shuttle's stated purposes, but NASA was ordered to make the bay bigger for the purposes of bringing specific secret spy satellites into orbit; a military program. I suppose it might have been a facetious comment in the video, but NASA was always beholden to military interests. This is why we haven't been back to the moon in over 50 years -there haven't been secret Cold War strategies that would benefit, even very indirectly, from returning.
NASA is not a military organization, it's a civil agency. Nothing NASA does is classified or secret, everything their instruments uncovered is available publicly online for all to see. It would be the USAF that's got secrets or the NRO (National Reconnaissance Office) that runs spy satellites.
ultra nationalism also helped japan.
Since we have no industry I’m gonna say no it didn’t.
As Obama declared those jobs aren’t coming back.
Can’t compete with slave labor
a
Europeans are very individualist they want name attached to self-made genius Edison, Tesla, Westinghouse, Ford, Elizabeth Holmes even had apparatus named Edison. Asian make consortiums and care less what Elon Musk say but benefit of all organization, so yeah Sematch used Japanese way to success. I don't say that people must be left-out from nerdy things that is suicide for nerd, but westerners is to much focus on leaders. In book about Steve Jobs he cared only about industrial design of his products 97% work was done by other people additionally 7 inventions was take from Xerox R&D department for free. Huawei also same principle more focus on consortium success that on personals achievements, even with Asians intrinsic despotism for individual rights (that is A. Huxley words not mine) they achieved more than USA.
Lord Jesus Christ is the only solution, He is the Truth.