So if they're making machines then all those things will follow.
We do not have any details about these machines other than they are DUV. All that has been posted is they are making 25 machines in the next 2 years. Meanwhile ASML is shipping 130 every year. Until these machines are actually making product, I am skeptical.
but they sure as hell can make the majority of chips in production today.
Not today or the near future. Manufacturing 25 machines in the next 2 years, it will be a problem to actually make the majority of chips. By comparison, ASML is shipping around 130 DUV machines every year. That does not factor what Nikon and Canon ships.
1) "smaller" = "Bigger" (in terms of sales, and modern utility. This is exactly what I meant, UV Light has a threshold over which the structures are smaller than the wavelengths you are trying to use to make them. this means you have to use some kind of mask technology,
It is not a "mask technology." You are talking all the trade secrets of ASML to make feature smaller than 7nm.
Most of the technology used in a fab is either precision electronics (needed to control the energizing source's true positioning to efficiently control its beam), which you can do with less sophisticated chips than the ones you are making.
You said " But homegrowing that tech expands what they can do *now*, builds local knowhow on building fab machines, and primes them to not be as stymied when they move on to building more recent forms of lithography. ". More recent forms is 2nm. Future forms are less than 2nm. Not sure how you can argue that China does not need more sophisticated tech while arguing they can make sophisticated tech.
You dont need 2nm tech to make 2nm tech. You can make 2nm chips with 14nm ones though.
Citation Needed. Please show me one example where a 14nm DUV machine can make 2nm feature chips. I'll wait.
The knowledge of how to make and run fab machines is not to be discredited-- Japan's big LED rise came from a single guy repeatedly rebuilding and tinkering with a vapor deposition machine at a university in the US, where he was treated like ass, because he didn;t have a pHD. DO NOT MAKE THAT MISTAKE TOO.
You seem to ignore the fact that Japan with their decades of lithography experience could not make EUV work.
t is not a breakthrough. It's a milestone on independent progress that is occurring at a very quick pace. Again, you are proudly crowing about how big your old growth forest is, while discounting how quickly your competition is doing forestry work.
I never said any of that. Never. What I said specifically China has been able to buy from the East for decades. Your attempt to make it some sort of East/West battle is factually untrue. What I also said is DUV is 30 years old. These are facts.
If AI goes away, will nVidia fail? Even before the huge demand from AI, they were the most successful graphics company. They use their GPGPUs in supercomputers. They are the go-to for gaming.,
They might if they produced a lot products they cannot sell. In favor of AI, Nvidia stopped making consumer GPUs. That means they used their allocations from TSMC to make data center computers. If the AI companies renege or do not pay NVidia for all of those chips, it is not like NVidia can sell them to consumers.
A lot of startups will go under. Most established companies will take a licking, but will be fine (Oracle is in a LOT of verticals) AI will still be around.
Oracle's credit rating was downgraded to BBB-/A-3 because they borrowed a lot of money to fund their AI expansion. It does not matter if Oracle is "in a lot of verticals" if the company as a whole owe more money than they can pay.
That is very likely to be the main asset that people buy up.
Who will buy this up? The CPU hardware for AI was specialized; they do not make good webservers using AI hardware.
Memory prices however I expect to crater. Once that compute supply is shifted away from making AI slop to more productive uses, we will have a glut of memory chips combined with excess production.
The RAM was specialized for AI called HBM; they cannot be used in place of DDR5 in consumer or servers. Those RAM chips can't even be melted down to be recycled as silicon as they have too many impurities. HBM might be used by HPC like the top 500 supercomputers.
The main fall will be the idiots that 'invested' in AI companies. The AI founders will still come away in the 1%, but most of that money will have been striped away from the investors.
The main fall will be any one that indirectly invested in AI. Retirement accounts that had stock in Intel, Nvidia, Oracle etc. Basically many people will be hurt by the fall not just the 1% you think./P.
UV has a top limit, sure.
DUV has a bottom limit. It is about 7nm. Beyond that, circuits can be printed however it requires multiple passes. Multiple pass means more defects and less yield. Realistically any fab could have used DUV for features less than 7nm; the problem was that the yields were so low due to defects that they could not make a reasonable profit. EUV has been able to get past the limits of DUV and.
But homegrowing that tech expands what they can do *now*,
Homegrown only means they are making older technology in China. That does not mean they somehow can advance the technology. That's like saying if I build a 2015 Ford Mustang in my garage, I can design the next generation of 2027 Ford Mustangs.
builds local knowhow on building fab machines, and primes them to not be as stymied when they move on to building more recent forms of lithography.
Nikon and Canon who have decades of lithography experience with no trade restrictions were unable to make EUV practical. Yet you think China with many trade restrictions can make EUV. That also does not underscore the problem that China does not have the suppliers that are required for EUV. For example, China would have to replicate all the know how and experience of Zeiss who have been making lenses for over 150 years and are the only company who makes the lenses for ASML's EUV machines.
You kids cant see the forest for the trees, complaining about the saplings china is planting being small, while croaking pridefully about how big your own trees are.
I've worked in the industry telling you why this announcement is not the breakthrough you think it is. Have you considered that you simply do not know enough about the industry?
In time, china will dominate chipfab.
I'll believe it when I see it.
. But you witness them taking milestone after milestone every year, that perspective becomes challenged.
The problem is you think copying a machine and technology from 30 years ago is a "milestone". Most people do not think that.
Now you can piss on every achievement and say that they are nowhere near yet, but with every step they are getting closer, and they are taking the steps at speeds no one thought possible even a few years ago.
I am presenting facts. I am not trumpeting every announcement as a major breakthrough when it is not.
China has a homegrown AI industry running on a homegrown tech stack, and they are leaving the US in the dust by now, all the while being about 100x cheaper.
There are claims by China about their AI. Given how AI claims from US companies are dog shit, I am reluctant to believe China's claims either.
As to them buying lito from other countries, the US is working hard at putting a stop to it. Whether they succeed or not, China can not afford to get caught by surprise.
Your entire point was about how China doesn't have to buy from "the West." They haven't had to buy from the West all this time. Decades now, China has been able to buy from Japan if they wanted.
and the 3D stuff we've seen from the chinese sidesteps most of the issues
What 3D stuff are you talking about and what issues are you talking about? When it comes to "3D stuff", the first FinFET transistors were demonstrated in the late 1990s with TSMC making the first ones in 2002. For the most part, anything below 22nm uses FinFET.
China is learning that there's a long road ahead just to replicate the equipment that's been available for a decade.
More like 3 decades if you count the earliest DUV machines.
We don't really understand it, so we'll give it to the programmers.
