How China built its ‘Manhattan Project’ to rival the West in AI chips
japantimes.co.jp328 points by artninja1988 14 hours ago
328 points by artninja1988 14 hours ago
Domestically, we often put it this way: since it wasn’t made by God, we can definitely make it ourselves. It’s only a matter of time — if not this year, then next year; if we can’t do it next year, we’ll just keep going. This is how we approach everything.
There is a small caveat, though. China was not actually that far behind in the semiconductor field in the past. The problem was that corruption and fraudulent projects were quite serious, which undermined the Chinese government’s confidence in these efforts. A few years ago, there was even a so-called “transparent computing” scam project that was awarded a national-level prize.
Corruption and fraud can slow China’s progress, but they will not affect the final outcome. This is because it is not only a government policy, but also a Chinese way of thinking. Nothing can interrupt this process.
In fact, aside from high-end chips, China already dominates the mid- and low-end chip segments.
Hopefully, this century, we can shed some of the 'dominating' mindset that has led to technological exclusionism in the first place. Not that catching up to the state-of-the-art isn't warranted, but that progress will become pocketed once more if we keep falling for the same economic traps.
I wonder how much scam there is on US side...
To be honest, maybe only Americans themselves really understand it. Our understanding of them is that they have poured vast amounts of money into areas outside of technology.
“transparent computing” -> this shit: https://www.science.org/content/article/critics-pounce-china...
The final outcome is affected by the final 10%, you can even call it 1%, for which the semi-corrupt or "communistiquesque" countries never (seemingly) have the will or sheer talent for.
> "communistiquesque" countries never (seemingly) have the will or sheer talent for.
I don't have the data to back it up, but I think that there is actually the same amount of will and talent in China as in the West
Based on the population size and school system, I'd conjecture there's more... though there is brain drain and emigration to consider.
> since it wasn’t made by God, we can definitely make it ourselves.
This implies copying what someone else did. Rather than inventing something new. I know it's not what you meant but if it wasn't made by God it's because it's already made by someone else. The sentence says to me more about copying than some relentless pursuit. The people who invented the thing to copy, those were more relentless presumably.
And then again the Chinese invented plenty over the years. These generalizations are bit meh.
You’re not wrong to think that way. But now there’s less and less left for China to “copy,” and it’s hard to argue that many things aren’t being invented by China itself.
Perhaps the real question is this: why is it that places that used to be technologically advanced no longer produce new, original inventions? Is it fear of China copying them? Did the U.S. decide not to develop a sixth-generation fighter jet because it was afraid China would copy it? Did it stop working on battery technology because it feared China would copy that too?
Can't wait to see the first Nobel prize in physics being awarded to a scientist who is actually a product of Chinese academy. Any moment now.
The Chinese are ahead at too many things at this point to think they're only good at copying
If I was in china's position and so much is at stake, how can I go towards engineering all the tech from scratch when I can reverse engineer existing tech from west?
All is copied in one way or another, progress in a vacuum is truly artificial and those who've been singularly credited for certain inventions likely have so because of the luck of the draw.
And that copying was largely enabled by our greedy western bean counters that outsourced so many things in the first place.
If I were running this show, I would have a second concurrent project as a hedge and as a chance of leapfrogging the West: trying to make free electron laser lithography work.
Free electron lasers have lots of (theoretical) advantages: no tin debris, better wavelength control, the ability to get even shorter wavelengths, higher power, higher efficiency, and it’s less Rube Goldberg-ish. Also the barrier to entry for basic research is pretty low: I visited a little FEL in a small lab that looked like it had been built for an entirely reasonable price and did not require any clean rooms.
So far it seems like Japan is working on this, but I have the impression that no one is trying all that hard.
https://iopscience.iop.org/article/10.35848/1347-4065/acc18c
Yeah I think it's likely they get an EUV machine working but with less efficiency than ASML just because of how long it takes to tune these beasts and work out all the kinks.
The big brain move is to try leap-frog the whole thing with XFEL. Smaller wavelength, way brighter source, no vaporized tin particulate, etc. It's a much bigger lift, new optics, new resists, etc. So a completely brand new supply-chain from scatch but with no competitors on that tech yet and low will for Western companies to try compete on it because they need to get money out of existing EUV tech first.
This is very similar IMO to Chinese auto manufacturing. Their ICE cars never really did meet the same standards as European or Japanese manufacturers despite JVs etc.
However EVs and green-tech are analagous to the XFEL path, they built from scratch and leapt over the competition that was happy to sit on it's existing profitable tech instead.
> However EVs and green-tech are analagous to the XFEL path, they built from scratch and leapt over the competition that was happy to sit on it's existing profitable tech instead.
I'm not convinced Chinese EVs are technologically better. They've just command economied demand and reduced costs via mass production. The technology seems pretty inline with anything available in the West but demand isn't there to take advantage of scale. China is ahead in EVs by metric of quantity for sure but I don't think they're got next gen battery tech they are keeping secret.
Making batteries for $80/kWh IS the next gen tech. I’m pretty sure China invented lipo (EDIT: I meant lfp) (at least they’re the only ones making it) and they’re currently pushing ahead on sodium ion. They are also the ones who have pushed lithium ion to the point it is today. My first EV was a Nissan Leaf that cost 40 grand and could drive 80 miles. Now you can buy 300-mile cars for about that. That was all China’s doing and nearly every EV on the road today uses their batteries.
They have done to the battery market exactly what Taiwan did to the chip market. You can buy an EV made anywhere the same way you can buy a laptop made anywhere. But guess where the chips and batteries were made.
They didn't invent LiPo (and you probably don't want those in a car), nor did they invent LFP (LiFePO4) but they did license it when no one else wanted to and turned it into probably the best EV battery tech you can buy today. They didn't innovate a ton on the chemistry but they did on the packaging side, BYD and CATLs structural pack designs exploit the low thermal runaway characteristics in a way that wouldn't be safe for NMC etc to reach near parity on density but with better longevity and cost.
They will be the first to sodium ion and solid state though.
DARPA funded a bit in this space a while ago. (Example: https://www.nextbigfuture.com/2011/01/darpa-maskless-nanowri...) I'm not sure how you get over the bandwidth limitations, even with multi-beam.
This is a totally different technology.
A free electron laser (FEL) uses free electrons (electrons not attached to a nucleus) as a lasing medium to produce light. The light would shine through a mask and expose photoresist more or less just like the light from ASML’s tin plasma contraption, minus the tin plasma. FELs, in principle, can produce light over a very wide range of wavelengths, including EUV and even shorter.
That DARPA thing is a maskless electron beam lithography system: the photoresist is exposed by hitting it directly with electrons.
Electrons have lots of advantages: they have mass, so much less kinetic energy is needed to achieve short wavelengths. They have charge, so they can be accelerated electrically and they can be steered electrically or magnetically. And there are quite a few maskless designs, which saves the enormous expense of producing a mask. (And maskless lithography would let a factory make chips that are different in different wafers, which no one currently does. And you need a maskless technique to make masks in the first place.) There were direct-write electron-beam research fabs, making actual chips, with resolution comparable to or better than the current generation of ASML gear, 20-30 years ago, built at costs that were accessible to research universities.
But electrons have a huge, enormous disadvantage: because they are charged, they repel each other. So a bright electron beam naturally spreads out, and multiple parallel beams will deflect each other. And electrons will get stuck in electrically nonconductive photoresists, causing the photoresist to (hopefully temporarily) build up a surface charge, interfering with future electron beams.
All of that causes e-beam lithography to be slow. Which is why those research fabs from the nineties weren’t mass-producing supercomputers.
What bandwidth limitations are you referencing? My understanding is that deep euv lithography is limited by chromatic aberration, so the narrow bandwidth of a single beam FEL would be an advantage. If you need more bandwidth, you can chirp it. Is the bandwidth too high?
They mean bandwidth as in rate at which one can expose a mask using an electron beam, because they’ve confused two different technologies. See my other reply.
P.S. Can you usefully chirp an FEL? I don’t know whether the electron sources that would be used for EUV FELs can be re-tuned quickly enough, nor whether the magnet arrangements are conducive to perturbing the wavelength. But relativistic electron beams are weird and maybe it works fine. Of course, I also have no idea why you would want to chirp your lithography light source.
I don't think it's strictly chirping, but there are methods to achieve that sort of time/ bandwidth trade-off with FELs. I've seen references to it pop up in high speed imaging, though the details of anything that fast and small are quite outside my expertise. Wasn't sure why you would want high bandwidth either, hence my confusion.
I'm not all that familiar with the intricacies of this industry but it seems they have at least one corporation with ambitions in this area:
https://www.scmp.com/news/china/science/article/3333641/chin...
That mention of "quantum" seems suspicious, but it's beyond me to judge whether their presentations are credible:
http://lumi-universe.com/?about_33/
If they actually produce machines that can do ~14 nm stuff on "desktop" sized equipment, perhaps we'll see a lot of it eventually. As far as I can remember a lot of decent processing and storage chips were made with ~14 nm processes over the last decade or so.
With Nvidia scaling down their consumer GPU production [0] I wonder if we will see consumer GPUs shipping from China in the future. Western companies seem to be abandoning the consumer/prosumer market which will have bad implications for hobbyists and aspiring professionals down the line.
[0] https://www.pcmag.com/news/nvidia-might-cut-rtx-50-gpu-suppl...
It’s a good thing that Chinese companies have zero expertise in leveraging consumer demand for lower-end tech to develop know-how and catch up with the state of the art from Western-aligned companies and then economies of scale to surpass them in distribution.
Exactly. That's where this is heading, and the West-- as usual-- is pursuing quarterly profits and forgetting to look up.
>pursuing quarterly profits and forgetting to look up
"Forgetting to look up" implies a desire or intent to do so. The United States - former leader of the collective West - made the choice decades ago to sacrifice everything on the altar of quarterly profits. All that remains are the consequences of that decision.
I would just be careful to discount the capitalist West. You’d have to be blind to ignore the massive overbuilding of property in China, which they are now demolishing. All of that wasted capital. Authoritarian regimes with controlled media always seem successful… until it doesn’t. Up until the USSR collapsed there were many prominent people in the West saying it was the superior system. The market test - meaning floating prices and the response to them - is a superior way of allocating capital. We need to see how all of this plays out
I swear I've been reading about overbuilding in China since, like, 2012. And I've definitely used it in arguments myself. Not only China hasn't collapsed, but it has improved massively since then, as far as I can tell.
> I would just be careful to discount the capitalist West.
I would. It's showing the weaknesses and limitations of its ideology.
> You’d have to be blind to ignore the massive overbuilding of property in China, which they are now demolishing. All of that wasted capital.
So what?
> Authoritarian regimes with controlled media always seem successful… Up until the USSR collapsed there were many prominent people in the West saying it was the superior system.
The West is literally de-industrializing and can't seem to built shit except slowly and expensively. Industry after industry gets hollowed out as China takes the lead.
Do not make the mistake of reasoning about US vs China from the experience of US vs USSR. China doesn't have a command economy, outproduces the US, and controls many key industries. The US is resting on its laurels, and its people cope by thinking of the few industries where it's still ahead, but those are dwindling.
> The market test - meaning floating prices and the response to them - is a superior way of allocating capital.
That's not truth, it's a dogmatic assumption.
China has been able to exploit a dogmatic belief in the free market to siphon the real capital out of the West and into itself (industry and know-how) in order to achieve dominance. The US elite is content to have paper. We'll see how that works out.
> We need to see how all of this plays out
If you're rooting for China. If you're rooting for the US, by then it will be too late to course correct.
> You’d have to be blind to ignore the massive overbuilding of property in China, which they are now demolishing.
Is it all being demolished, or is 95% of it being moved into?
Because all those ghost cities that China was building that the news kept bitching about... Are now all full.
Meanwhile, in the West, we have a housing shortage. Who looks the fool now...
> You’d have to be blind to ignore the massive overbuilding of property in China
Isn't the same now happening with the US with the massive overbuilding of AI capacity? Seems like a tightly centralized capitalist system is not that different from a communist one.
Or rather thinking one step forward the question arises, whether we use the right words for the right things? Does the capitalist West has any defining economics characteristics of a liberal free-market capitalism at all?
Private ownership of means of production: On an atomic, legal level of course. But if point at an NVIDIA based compute rack at a US based random datacenter, can someone tell me actually who owns it? I am interested in the actual natural person who has an ownership share of this capital asset, not the myriads of layers of corporate and financial networks of equity delegations through investment banks, but the actual owner?
Profit oriented: Of course, it is said so. But do really companies, entrepreneurs do things to maximize the profits of the actual owners, shareholders? Are the executives and boards really that keen on putting forward the interest - of the previously referenced unknown - actual owner of the capital assets?
Free market based: This has also multiple sub-characteristics, but most importantly something about competition, or rather the lack of collusion and that economic actors (including consumers, (natural person) investors) are all fully informed. How much is this true in the West?
I think we are very much lost in labels.
When did this whole quarterly profits thing start and what lead to it?
It's not real. Companies routinely lose money for years in pursuit of long term growth. But for some reason people love to use this as an explanation of everything wrong in our country.
I'd cite as a counterexample in recent memory Sears, GE, Boeing, and Intel. I think collectively they've destroyed close to a trillion dollars by focus on quarterly results over long term, and they're not alone.
I sometimes wonder what a Drucker or ishikawa would say of today's "vaunted American management". Speed roughly short term thinking is too strong of a force in our American thinking. Heck I've counted three recent HN posts this month pushing for speedy software development too.
Yes, and we all saw what happened. They've experienced serious financial consequences, some went out of business. This is exactly what is supposed to happen when you do dumb shortsighted things.
There's also risk in investing in very long-term things that may not pan out.
WAI, in other words
The exception proves the norm.
The only major example I can think of is Amazon dot com which famously reinvested all its profits into itself for well over a decade.
The fact that investors didn't punish Amazon dot com was seen as befuddling in the press.
> Companies routinely lose money for years in pursuit of long term growth.
No, I don't think this is true at all because you used the word "routinely". I would claim it is very rare.
FAANG-like stock, in general, has paid little to zero dividends for long periods of time post IPO, their rational stock values being based on hypothetical future dividends only after the initial self-investment phase is over.
Don't most tech startups lose money for years before they maybe make a profit?
I mean, I agree that such companies are over-represented in thinking about small businesses if that's what you mean. Normal companies have to be profitable quickly for sure.
It feels like tons of companies get valued based on userbase or revenue or theoretical breakthrough rather than ever having to really think about breaking even, but I know that's just because those folks get all the press.
You can round it down to Milton Friedman as the ideology and Jack Welch at GE in the 80s as the implementation and figurehead, but the original seeds were in the SEC mandating quarterly reporting as part of regulation after the great depression.
We can all agree to blame Jack Welch as shorthand though, I think.
Minority shareholder rights, you can be sued for not maximizing profits see Dodge v. Ford Motor Co. (1919).
> West-- as usual-- is pursuing quarterly profits and forgetting to look up.
The companies building out vast data centers for AI aren’t looking to make profits for several years (if ever), and are catching a lot of flak for it. The shareholders who seem to be focused on short-term profits and punish them every time they get cold feet. Oracle is a prime example of this.
I don’t know if the markets in Asia work differently, or if the investors there are just as fickle.
And they should be catching a lot of flak for it because it's not really long term strategic planning, it's overhyping a technology and running roughshod over society promoting misuses and AI slops.
To be very fair, Chinese companies are also pursing quarterly profits. They're just better at scaling things up and down very fast because of immense supply chain options.
> immense supply chain options.
so this begs the question - why isn't the west's own supply chain options as immense? My unresearched answer is that the gov't policies of the west doesn't induce it, while china's gov't does (which includes targeted subsidies, tax incentives and state driven finances).
The "hidden" cost is that the workers in this supply chain isn't as well paid and isn't as powerful as the workers from the west (there's no unions in china for example).
> why isn't the west's own supply chain options as immense?
They used to be. Since roughly the 80's, policymakers have decided it is better for the shareholders to outsource most of that industry overseas to China and India and etc, where the labor is cheaper.
Note that workers and especially union members actually have every incentive to keep that production domestic, but shareholders and CEOs profit when they can cut labor costs and the typical Western consumer values cheap products more than the health of domestic industry.
Western industries have been supported by subsidies, tax incentives, bailouts, low interest rates, and a dozen other things from the gov't but the same policies reward outsourcing and financial engineering more than actual production capacity.
> why isn't the west's own supply chain options as immense?
The US explicitly chose to be a service economy. China explicitly chose to be a mercantile economy.
The US can absolutely switch paths, it will just take a long time and will require pushing millions into poverty. But we're on track to do it.
> The US explicitly chose to be a service economy. China explicitly chose to be a mercantile economy.
In other words: the US wants its workers assembling hamburgers, China wants its workers assembling drones.
And when there's a conflict, the US will lose because you can't win a war with hamburgers.
> But we're on track to do it.
The only thing that the US is on track to is getting a taste of what real corruption feels like, enriching Trump's friends, and hollowing out its middle class.
The beneficial owners of the US economy sold our industrial manufacturing base to the Communist Party of China because the price was good. China got our hard power and US capital owners got to break the back of the US labor movement. A win-win deal for the ages.
There are lots of reasons, but also, having 1.4B people under the same government that has more-or-less aligned strategic goals help. Like supply chains within Japan, from what I've seen and experienced, are pretty strong. However, the options will always look smaller compared to a gigantic organism across the pond.
The West only represents a minority on this planet. China alone has 1 billion people.
The last 200 years has been an aberration and it is currently in the process of being corrected.
Chinese companies are simply not as beholden to shareholders - the stock market really doesn’t dominate the country’s financial landscape as it does in the US
Is anyone here calling legislators about it to inform them of this?
Does anyone here have leverage to affect strategy?
It might be a bit nihilistic but at this point I don't think the current US administration has any strategy. In past administrations, it felt like even if there was a strategy, bureaucracy and lack of caring enough to do their job led to nothing happening. In this administration, it feels like there's no care for the rules so in theory a strategy could be pushed through... except there isn't one - literally whoever is the last person to talk to the president is the person who gets to set the direction.
> I don't think the current US administration has any strategy
I 100% believe the strategy is to enlarge the Trump family's wealth, and it's been a wildly successful strategy (in the past year he's been able to create billions in wealth for his family [0]). At least this vaguely ties Trump's success to the success of the United States in a limited capacity. Completely destroying the US is not ideal for him, but it's clear all policy decisions being made are being done so based on their capacity to improve Trump's situation.
We've been headed this direction long before Trump, from both parties, increasingly American policy is about what's good for American companies and in particular the people who own them. Now that pool has just shrunk a bit.
0. https://www.wsj.com/politics/trump-family-business-visualize...
> Completely destroying the US is not ideal for him
How can you say? The ultra wealthy are not playing team sports. If the country burned tomorrow they would just sit on their yacht or buy citizenship somewhere else.
Sure the lion share of his investments are currently in the US, but that could easily change.
Most cynically? I sincerely believe there's more value he can extract from American workers if he can keep the ride going on just a bit longer. For example: it's better to keep the stock market a float a volatile so he can transfer a bit more wealth from the American people before it does eventually break down.
Yes, they'll all be on yachts when the shit hits the fan, but they're still fighting to figure out who get the biggest yacht, and right now it seems like Trump has more to milk from us before he entirely lets this thing fall apart.
> Completely destroying the US is not ideal for him,
Exactly, just as taking out structural supports when stripping copper and goods from a three story walkup is sub optimal and potentially fatal.
But make no mistake, from way out here (Australia), having watched the US for decades, it really does look like you've a grifter inside the house taking everything that isn't nailed down with zero concern for anyone else in the US.
It's a bad time for those that cannot afford shiny gold baubles.
> Completely destroying the US is not ideal for him
Nah.
He wants to be a dictator that extracts wealth from it's citizens.
He has a benefit from following the communist path to extract wealth. Make lives miserable, so they are living off the state ( eg. standing in line for bread), so they can't protest.
Putin is not Trump's friend, but Trump idolises him for extracting enormous wealth from Russia, censoring news ( propaganda) and imprisoning political opponents, ...
Just check the "firehose of falsehoods" ( a Russian propaganda method), it will explain a lot about Trump.
I think it's the same problem as the past administration and most members of congress- they're just too old to care about 50yrs from now. I don't think they're actively against the 50yr+ future, it's just that the world is changing too fast, and they're falling back to what they know- competing with their peers for power, money, and status. They only have some inkling of actual empathy for the communities their grandkids are in at a personal level, and just have the "throw money at it" mentality for the bigger issues like healthcare, since that has been their MO for the last couple decades. Instead of taking leadership positions and driving change, they seem to just want to squabble and create fiefdoms and have others do the work.
I do think the current administration is still a step down from the (not particularly great) last, though. Congress has essentially given up their authority on everything so any movement must come from the top… and the top has an extremely small attention span.
> same problem as the past administration
Did you miss the Infrastructure act that spent $500B on roads, ports, and water projects? The CHIPS act that spent $50B on decoupling and R&D?! The Climate & Energy act ("IRA") that spent $400B on clean energy subsidies??!!
I can understand the perspective of wanting more, but the forward-looking policies of the last administration were in a different galaxy compared to those of the current administration, where the big plan is to chop USAID, boost deportations, and cut capital gains tax.
This is the difference between corn and the cob and corn in the toilet. No, it is not the same.
The problem is that China could have built the same infrastructure for $100B and in 25% of the time. Pumping subsidies into our bloated bureaucratic nightmare of a system is only going to make the lawyers and bureaucrats who are its gatekeepers fatter.
I’ve met one person whose job was funded via the CHIPS act - she was a lawyer.
I bet China’s first priority when building semiconductors isn’t hiring lawyers.
Without getting overly political, it is nihilistic because one of the major US parties got so high on its own supply of lies that the people currently running it FORGOT they were lying. You're looking at a similar situation to many authoritarian or fascist political systems in which they way to get ahead in no way involves doing your job, but making sure Stalin or Mussolini is happy with you. This started in the US in the mid 1990s when GOP leadership bought in on power being its own end and is now on full display.
People have all sorts of mythologized reasons for why the USSR failed, because while it often produced immense amounts of goods and services and well educated people in certain areas (sometimes beating "the west" by a good margin for one or two years at a go), it also made long term advancement contingent on the party and not the real world and became incapable of handling major changes.
We're witnessing that now in the US with perhaps one of the most incompetent governments in history that is also burning down the non-political institutions of expertise that for all their faults and mistakes, at least had educated and motivated people that cared about their purpose.
Exactly right. Democrats have the desire to do big things, but not the capability. Republicans have the capability, but no desire.
Current US strategy is to get South America+Canada resources and call it a day. They looked at global geopolitics and it looked too complicated for them.
Oh, they know. The industry has been lobbying quite badly for exactly this to happen. Why spend a fortune on innovation when a few bucks of lobbying can get the government to ban your competition because "China bad"?
As the comment you responded to said: it's all about the next quarterly profits. The fact that we are getting leapfrogged by China doesn't matter to those CEOs: that's a long-term thing, and it doesn't impact their next bonus.
> calling legislators
Good joke. Probably a couple of tech billionaires will eventually say something and then something will happen.
Call me a cynic, but legislators own stock in these companies. Their true interest in them is also "line go up".
Legislators aren't interested in actual expertise. They only need to know what each constituency wants and how much money they have.
this is one of those 'elections matter' cases. There's no strategy. Americans made it clear they want a country run by real estate crooks, crypto bros, gambling advocates and bizarre entertainment personalities. A country sized Las Vegas maybe and the beauty is the people get always what they deserve.
Or, to play Devil's advocate, people called in the rabid dog because they felt they had been sold down the river starting in the 1970s. The hollowed out industrial towns weren't good places to grow up, and when they did manage to go to college and play by the rules the ladders were already in a fast retreat up the walls..
Damned if you. Damned if you don't.
And their mistake lies in believing that populist rhetoric is going to make their lives better…
[flagged]
The CCP has killed millions of people, and is working on a genocide of the Uyghurs through forced-sterilization. Their record in the surrounding parts of Asia isn't much better than the US's record in Latin America, and where it is better, it seems to be externally constrained rather than self-constrained.
We're not the good guys here, but if you think China's any better then I have a bridge to sell you in Beijing.
Failed empire? Probably. That doesn't mean that China's government is somehow less horrific on its human rights record. Both countries can simultaneously have issues, but I sure know which one I'd rather hang out in, even with the current dumpster fire of an administration.
Yeah no dude even if you're critical of the mass detention/internment of Uyghurs in Xinjiang (a program which ended in 2021) it's nowhere near on the level of the mass murder and depravity of the Gaza Genocide. Not even close, unless you can point to a hospital China has bombed.
GP was talking about where he'd rather hang out between the U.S. and China. Gaza wasn't in the running.
If you expect people to engage with you in a way that requires effort from them, the least you could do is not use a burner account.
They learn really fast. I will not take the past as a prediction of the future.
I took the previous comment as satire, considering that - for example - Chinese electric vehicles are now far more affordable than anything produced domestically in the US.
Just to give a hint of how we plan on dealing with this, how many of those cars can you buy in the US today?
You can expect to be able to buy exactly that many chinese GPU or neural processors.
You are 100% correct on that, I fully expect that importing cheap GPUs and NPUs will be banned in the US, or tariff'd so heavily that it doesn't matter should they become available. But that will just allow Nvidia to fall behind until they get surpassed like AMD passed Intel.
A move like that will seriously hurt our ability to train and raise new software developers and the domestic game market.
Importing might be banned until the US loses access to its main source of GPU manufacturing. And what are the US going to do about it? Defend Taiwan? With missiles made with components and materials processed where, pray tell?
The US does need to start protecting its manufacturing again, but it’d be lucky to start at a level as high as high end semiconductors. That’d be like a stroke victim trying to run before they re-learn to walk.
As others have pointed out, this means less services, more manufacturing, less consumption, a probably a lower standard of living. But with the business as usual alternative looking a lot like business as usual in the western Roman Empire circa 450 CE, taking a hit to your standard of living while investing in a future which you still have the slightest control over, maybe feels like a decent trade.
When it comes to China and its ability to quickly mass produce items while incrementally improving on them, I absolutely will view the past as an indicator of the future.
It's just what they do as a nation.
Which then goes on to be repurposed into weapons of war.
Something to think about if considering the purchase of a DJI drone.
the attack drones being used in Ukraine are not DJI anymore. both sides produce extremely cheap, light, disposable drones en masse.
also, consider that a $50 smartphone can drive an ICBM.
I remember when my Linksys was considered a controlled device because it could be used to fly missiles.
We thought it was the coolest thing.
Yeah PRC probably going to dump 1T into indigenize semi efforts by end of decade, but IMO good chance they're going to treat strategic semi as commodity utility business than make NASDAQ lines go up model. When western semi has 1st tier suppliers taking 30% margin, asml taking 50%, tsmc taking 50%, nvidia taking lol 70%, there's alot of fat on the pyramid to trim. PRC doing cost plus 10-20% will basically be able to brrrt chips stupid cheap if they have mature domestic tool scaling, enough to wipe litho+yield inefficiencies.
Western semi still "safe" since west+co aren't going to source from PRC leading edge due to national security, but pretty soon they're either going to need to compress margins to compete which means cutting costs, which means cutting R&D because shareholder going to get theirs or western semi business model going to run on permenant subsidies. Which is what will probably happen considering their performance is why stonk lines go up right now. That 1T PRC spend and choose to simply discount for utility chips is going to wipe multipel trillion of western semi market cap and all the economic implications that entails so it might not even be bad idea.
The NASDAQ line go up model is why the AI boom is happening and a major factor in why it is Western companies leading the charge. The more bigger issue is that the west refused to sell chips to China so they had to figure out how to make their own. And margin compression is what free markets do. That is one of the big motivators to putting free markets everywhere, the freer the market the more compressed the margins become. All the people working hard at crappy jobs start working hard at high paying jobs instead until the competition drives the money out of the sector.
There is a theme in the industries China does well in - western regulators ban cut-throat competition, China competes very hard and wins. The situation at scale is pretty straightforward. Usually it is environmental or labour policy, so this case of the root cause being sanctions is a bit unusual. But, once again, how Nvidia is meant to compete in China when their best products can't be sold there?
Free markets suppose to compress margins, perfect market theoretically drive profits down to zero (aka involution). But you compress margin and you lose current western semi business model that is functionally monopoly suppliers/producers who can sustain 50%+ margins to keep their monopoly. Shed those margins down to 20% because competitor enters market, harder to fund R&D to keep lead, it's still "enough" to be profitable, but then western commercial companies have to think harder how to split that compressed margin between investors and R&D. Right now we know what this leads to. Investors get paid, companies beg for subsidies. Not that PRC companies aren't concerned, look at PRC stock capitlization, not nearly to the same degree.
Free markets? No, competition does that. Competition requires free markets but free markets don't guarantee competition.
I poorly paraphrased profits converge to zero under perfect competitive market. Yes real world not perfect competitive markets. Oligor/duo/monopolies form, sometimes subsidies other shenanigans pick winner(s), winners extract huge margins/rents to build moats lock out new competition. Sometimes they collude / settle on business model with higher margins, i.e. 10-20% being normal/commodity, 50% for software and semi is top end of luxury goods. New state backed competitor enters and decides they can live on 10% margin, and then incumbants business model falls apart unless state also steps in to match.
E: And state can, but I don't know if state generally willing/able to backstop companies to 50% margin long term. I can't think of any, maybe some major state oil. Nvidia/TSMC with $$$ margins getting some CHIPs injection really meant for bailing out broke ass Intel was already anomalous, and it was basically to bribe them to onshore production.
Note "converging to zero" doesn't really mean zero because economics includes opportunity costs. It just means that outsized gains don't exist over the longer term without some kind of market power. In the long run, most industries end up making the same returns.
> margin compression is what free markets do
Except the market pretty much can't do this with Nvidia. Nobody is showing any sign of catching up: it is entirely possible we are seeing a runaway train and without the intervention of a massive state like China to create a viable competitor, there will never be one.
This situation has been going on for 5 years now. It's ridiculous to assume there will never be competition. You. could have said the same about Intel a couple of decades ago.
> There is a theme in the industries China does well in - western regulators ban cut-throat competition,
The problem is not regulation, it is the lack of it: anti-monopolist practices and deregulation of the finance industry has led us to insane bubbles, dead markets and extreme wealth concentration. Any competition gets bought, crushed, or undercut via bankrolling. This is what you get when the 0.0001% gets to pull the strings again. Must watch (3 parts): https://www.arte.tv/en/videos/103517-001-A/capitalism-in-ame...
I think you have no idea how a command economy works. Haha.
It’s literally state sponsored monopolies.
They flood an industry with funds and let them all compete with each other. Then they back the biggest winner. A command economy usually mandates X units of Y good. This isn’t quite that.
That’s only the first year after that it’s command and control for the vast majority.
Also you take is highly simplistic. Even the small players are command and control. You’re likely just not aware how it works.
Its a pretty big detriment to American thinking that they take cold-war era characterizations of the Russian economy and then apply them to 2025 China.
The current Chinese worries are about having too much competition rather than too little, Google "involution" to read about it.
Your analysis is out of date I think. This has already happened. Poor NXP just got their asses handed to them by the PRC. The fab they have in Italy looks nice but PRC has many of those.
Also Texas Instruments, STMicro, Onsemi, Microchip Tech, i.e. what PRC is doing after going big in mature nodes last few years they likely will also do in leading edge. IMO there's argument that since leading edge will definitely be strategically bifurcated PRC and western semi can pseudo collude to maintain higher margins, especially if PRC wants to claw back investment. But if western semi continues to drive economy/growth there's also incentive to weaponize margins.
You don't need CUDA for gaming but software is still just as big of a moat. Gaming GPU drivers are complex and have tons of game-specific patches.
With their new Radeon/RDNA architecture it took AMD years to overcome their reputation for having shitty drivers on their consumer GPUs (and that reputation was indeed deserved early on). And I bet if you go read GPU discussion online today you'll still find people who avoid AMD because of drivers.
That won't stop them, but it's a big barrier to entry.
Oh and that's just to get the drivers to work. Not including company-specific features that need to be integrated by the game devs into their game codebase, like DLSS / FrameGen and FSR. And in the past there was other Nvidia/AMD-specific stuff like PhysX, hair rendering, etc.
> Gaming GPU drivers are complex and have tons of game-specific patches.
I don't think the Chinese government will be too upset if cheap Chinese GPUs work best with China-made games. It will be quite the cultural coup if, in 20 years time, the most popular shooter is a Chinese version of Call of Duty or Battlefield.
They made the most popular RPG last year already - why do you think it'll take 20 years for them to make the most popular shooter? For that matter, the Singapore-HQed SEA makes Free Fire, which topped Google Play in 2019.
Im aware of Genshin Impact, and that NetEase is behind Marvel Rivals. FPS tend to have sticker fanbases, but I chose 20 years because that's what I guess is how long it may take not only for the domestic EUV to launch and get yields good enough for a cheap but competitive GPU out the door.
Yea, but less than in the past. Modern graphics APIs are much thinner layers.
This was even proven in practice with Intel’s Arc. While they had (and to some extent still have) their share of driver problems, at a low enough price that isn’t a barrier.
I believe the software will follow the hardware. Not immediately, of course, but if I want to learn to do ML and have to pick between a $2500 Nvidia GPU and a $500 Chinese GPU that's 80% as fast, I would absolutely take the cheap one and keep an eye out for patches.
When it comes to drivers, IMO all they really need is reasonable functionality on linux. That alone would probably be enough to get used in a budget steam machine or budget pc builds, with Windows 11 being a disaster and both RAM and GPU prices shooting through the roof. The choice may soon be Bazzite Linux with a janky GPU or gaming on your phone.
On the other hand, all it would take would be one successful Steam Deck/Steam Machine-style console to get all the developers of the world making sure that their games work on that hypothetical GPU.
I don't think that it will happen in the next 5 years, but who knows?
Cuda is 20 years old and it shows. Time for a new language that fixes the 20 years of rough edges. The Guy (Lattner) who made LLVM is working on this: https://www.modular.com/mojo
Good podcast on him: https://newsletter.pragmaticengineer.com/p/from-swift-to-moj...
mojo been in the works for 3+ years now.... not sure the language survives beyond the vc funding modular has.
What I gather from this comment is that you haven't written CUDA code in a while, maybe ever.
Mojo looked promising initially. The more details we got though, the more it became apparent that they weren't interested in actually competing with Nvidia. Mojo doesn't replace the majority of what CUDA does, it doesn't have any translation or interoperability with CUDA programs. It uses a proprietary compiler with a single implementation. They're not working in conjunction with any serious standardization orgs, they're reliant on C/C++ FFI for huge amounts of code and as far as I'm aware there's no SemVer of compute capability like CUDA offers. The more popular Mojo gets, the more entrenched Nvidia (and likely CUDA) will become. We need something more like OpenGL with mutual commitment from OEMs.
Lattner is an awesome dude, but Mojo is such a trend-chasing clusterfuck that I don't know what anyone sees in it. I'm worried that Apple's "fuck the dev experience" attitude rubbed off on Chris in the long run, and made him callous towards appeals to openness and industry-wide consortiums.
Most of the stuff you pointed out is addressed in a series of blog posts by Lattner : https://www.modular.com/democratizing-ai-compute
Many of those posts are opinionated and even provably wrong. The very first one about Deepseek's "recent breakthrough" was never proven or replicated in practice. He's drawing premature conclusions, ones that especially look silly now that we know Deepseek evaded US sanctions to import Nvidia Blackwell chips.
I can't claim to know more about GPU compilers than Lattner - but in this specific instance, I think Mojo fucked itself and is at the mercy of hardware vendors that don't care about it. CUDA, by comparison, is having zero expense spared in it's development at every layer of the stack. There is no comparison with Mojo, the project is doomed if they intend any real comparison with CUDA.
Its not really just that AMD drivers are not that great (they are not) but they have been stable for a long time.
Its that nvidia relentlessly works with game developers to make sure their graphics tricks work with nvidia drivers. Its so obvious you miss it. Look in the nvidia driver updates they always list games that have fixes, performance ect. AMD never (used?) to do this they just gave you the drivers and expected developers to make their game work with it. The same strategy that MS used for their OS back in the 90's.
Thats at least how things got where they are now.
AMD provides this. Example: "Fixed Issues and Improvements
Intermittent driver timeout or crash may be observed while playing Warhammer 40,000: Space Marine 2 on some AMD Graphics Products, such as the AMD Ryzen™ AI 9 HX 370.)
Lower than expected performance may be observed in Delta Force on Radeon™ RX 7000 series graphics products.
Intermittent stutter may be observed while playing Marvel Rivals when AMD FidelityFX™ Super Resolution 3 frame generation is enabled. "
https://www.amd.com/en/resources/support-articles/release-no...
The whole “improve a game’s performance on the driver side” thing: does AMD simply not do that at all? Or just far less?
They definitely do it some, like Starfield came out with FSR out of the box but they didn’t add DLSS for several months. I got Starfield for free when I bought my 7800X3D which was a nice bonus. Definitely to a lesser degree than Nvidia though.
Frankly, this always seemed like dirty hacks - either the game or the drivers don't actually comply woth the graphics API and then the drivers need to hack around that. :P
The general trend of the industry is to move computational resources from the hands of users into data centers, so that they can control what can be done and how much they'll charge for computational services. In the medium term, a lot of what we take for granted nowadays will only be accessible from cloud providers and companies will pay more and more in subscriptions for these services.
Computer freedom is dying. Everything the word "hacker" ever stood for is dying. Truly depressing...
Isn’t that mostly economics? I definitely prefer using Claude to GPT-OSS120B for a code assistant.
I don’t know about you, but I don’t have $500,000 laying around to buy myself a DGX B200 with a TB of HBM and 2TB of system ram, nor the 14.3kW of power to run the thing.
But will you still be able to afford using it once you have to pay the real price ? Once the venture capital dries up and the dumping stops ?
I'm pretty sure all of these LLMs operate in the black on inference costs.
If I were to set up a DGX200 in my garage, say the 5 year TCO is a million dollars. Split that among 500 people and we can get it done for maybe $30/mo per user in total operating cost. I would bet that these LLMs are far more oversubscribed than 500 subs per server.
> I would bet that these LLMs are far more oversubscribed than 500 subs per server.
Seems like on hn a lot of people pay for the subscriptions.
I don't personally know a single person who pays for any type of llm subscription. I am a staff sw engineer, been doing this a long time.
I acknowledge this is an anecdote. I just happen to know a lot of people at a lot of different companies from my network. Nobody pays for any of this. My company has banned llms, even if I wanted to use one, I can't.
I actually even gave one a shot tonight. I asked for a list of repos I needed to clone to build a yocto image for an nxp board. This was the result:
mkdir -p sources && cd sources
git clone -b $BRANCH git://git.yoctoproject.org/poky
git clone -b $BRANCH git://git.yoctoproject.org/meta-freescale
git clone -b $BRANCH github.com
git clone -b $BRANCH github.com
git clone -b $BRANCH git://git.openembedded.org/meta-openembedded
git clone -b $BRANCH github.com
I then pointed out that three of those lines were useless and asked it to fix those lines. The result I got was even more hilarious, and just as useless.
Disclaimer: this was the "dive deeper" button on a google search. No idea what fucking model it tried to use.
How will that work, exactly? The chip makers are going to have a list of approved "cloud providers" and they will refuse to sell to anyone else?
Cloud providers will use cheap investment capital to buy chips at increasing prices, while the public will be economically forced to get computational services from these cloud providers. After a few years, most software will work only when connected to cloud infrastructure, either for performance or for "safety" reasons. We're already seeing this with AI.
Cloud was there for many years and it's not that cheap, compared to ordinary servers you can buy. It's not clear how anything will change in the future.
Because of this I hope the current AI fad is a bubble and it bursts soon. So instead of cheap investment drying up the market for individual consumers, we'll have lots of used corporate hardware selling at scrap prices to end users.
It makes me feel so gross that these companies are leaving gamers behind. The whole idea of a GPU was from gaming and games. And the whole AI evolution was subsequently born over the fact that gamers/software engineers could toy around with incredibly powerful CUDA without having to shoehorn a weird graphics Api in the middle to do mathematics.
They did the same thing with the COVID crypto era boom. There really is no honor for these companies and I will be buying the first Chinese made silicone out of absolute spite and anger
scale down or not, we will see consumer GPU from China in the future. might be a copy or rip of existing GPUs, but it will happen. 3 of my GPU rigs are chinese MB built for the chinese market, ripoff of dual x99. They work, they are cheap, I got them for under $100 a piece. So maybe 5 years from now, we get cheap GPUs, and maybe they will be equivalent to 5090s, but who cares if the price is right?
This is nothing but good for normal people in the west. Chinese competition is needed to remove some of the rent seeking from the system.
Is AMD abandoning consumer market though? They seem to be much less in demand by AI industry.
AMD is selling close to 10% of what Radeon* used to sell 15 years ago and yet it's all nVidia's fault: https://www.tomshardware.com/tech-industry/amd-grabs-a-share...
Edit: I meant ATI but I guess AMD bought ATI in 2006! I thought that happened in the 2010s for some reason.
At this point, the only reason to be in the consumer GPU market is because that's the first rung on the ladder that leads to the AI data center market.
Would be interesting if the US decides to ban or heavily tariff these chips and if the consequence will be significanly cheaper data center access through chinese-owned sites/platforms
> scaling down their consumer GPU production [0]
>> Due to Memory Shortages
I don't think Nvidia wants to give up on consumer. They're a gateway into the overall AI ecosystem.
Having feet planted there also make sure they can play the local game when that begins to blow up. Nvidia wants a robotics play, too.
This is a pragmatic choice. And most of the money is in commercial.
> I don't think Nvidia wants to give up on consumer... This is a pragmatic choice.
You mean, NV is after the money with a heavy heart and a sad tear or two over the abandoned consumers, like "We love you so much but sorry, we must go pragmatic on you"?
> And most of the money is in commercial.
This is a serious systemic failure and it's even wilder that it's accepted without question.
They're still selling GPS. They just want people to rent them instead of buy. Its definitely shitty, but it's not like they're quitting.
That's not the right light to view this chess move.
If Nvidia had infinite supply and infinite resources, they would absolutely continue doing consumer. There are constraints that prevent them from doing so at the typical volumes.
Giving up on consumer also means giving up on a gateway to more CUDA ecosystem users.
If that happens, hardware trust becomes non-verifiable.
We will also see talent pipeline erosion.
Just further Western industrial policy failure.
> If that happens, hardware trust becomes non-verifiable.
Unfortunately I already have to run a binary blob just to play fps games from 10 years ago. I can't even load a new OS onto my phone anymore.
Ultimately I'm not sure hardware sourced from China changes the trust equation very much, at least for me individually. I have much more concern about the FBI, which has recently decided to ramp up investigations into queer people [0][1][2], than I do about foreign powers - at least as long as it's not actively destructive malware or something.
> We will also see talent pipeline erosion.
We absolutely will, and to some degree I wonder if we aren't already with how popular tablets and phones are. I've noticed many young people these days don't really know how to interact with anything on a computer that isn't an app. GPUs and RAM becoming more significantly more expensive will take a huge chunk out of the hobby market and in doing so they will intensify the pipeline erosion.
[0] https://www.advocate.com/politics/pam-bondi-trans-equality-b... [1] https://ncac.org/news/advocacy-isnt-terrorism [2] https://www.hrw.org/news/2025/09/26/us-trump-targets-opponen...
Non-verifiable by what standards?
Modern GPU's often have on device firmware, secure boot chains, microcontrollers, etc. If you don't control silicon design, firmware signing and update pipelines you can't meaningfully attest to what the advice is doing.
But for most of us, that's already the case, isn't it?
For the average paranoid person who is wasting their life on it, sure
But large organizations like defense are all about distributed trust anyway - even if you could verify the hardware, the guy you order to do it is going to be a whole command chain removed and likely a contractor with a clearance in the civilian world.
Whereas your high level political and military leadership having direct contact with managers and designers in production facilities is extremely valuable.
What if my high level political and military leadership are the threat scenario?
Remember Snowden.
You're screwed. An individual is completely powerless against the combined might of the entire country they live in. Nothing you touch and nobody you talk to can be trusted.
But realistically, they'll just bring out the wrench[0].
Or more likely you're just not that important and will be imprisoned without cause because close enough is good enough.
People always seem to imagine tyranny worries about a standard of evidence. Tyranny has arrest quotas, evidence optional.
I am kind of skeptical about the report, as there are almost no details revealed. Everyone knows that China wants to build its own semeconductor manufacturing devices. The question is how close it is to be used in real production. The report just throws out a very vague number, maybe ~2030, which I can give the same guess, too.
It's wild to me that so many skeptical westerners who want to nitpick certain unproven technicalities, when the entire world only gets bits and pieces of the on the ground reality of China's progress, like the original Reuters article which was clearly fed information by insiders.
You should be living in the world of "China has successfully developed EUV and equivalent litho supply chain" and basing your decision making off of that.
I also cant understand people being in denial about, or claiming other imagined moats or whatever. They're whipping the pants of us right now industrially, if the west has any advantages left its that we speak the truth about stuff even when it hurts, why live in denial.
Also this stuff was figured out and built once before, other than the effort and resources involved (which China has lots of), why wouldn't someone else be able to figure it out again?
The west is still underestimating China. There is a great anecdote, I think it's from the book 'Apple in China', about their engineers visiting a Chinese production plant. Some changes needed to be made to the place. The Apple people estimated that that would take two weeks.
They came back the next day. It was finished, the Chinese had done it overnight.
I recommend the HTX Studio YouTube channel. The things that they release on a regular basis would be year long engineering projects on other channels.
I believe Tim Cook himself has said Apple is manufacturing in China not because of cheap labour, but because of good engineering.
It's ironic that a lot of western domestic manufacturing takes place using machines that were engineered and manufactured in China.
I often wonder what is it that's driving the Chinese to work themselves to death to get this stuff done? Surely there must be some limit. I guess we can see it in the low birth rates, the youth unemployment, and I guess the desire to just survive because there's just so many people there. But still, I just don't get how Chinese just keep going and going. What is their end goal on a person to person level? Are they just going to keep killing themselves for the rest of their lives? What happened to the lie flat movement?
>we speak the truth about stuff even when it hurts, why live in denial
Unless you are talking about Israel :P
It's more like the genocide in Gaza is the uncommon case where western propaganda was openly rejected by the population, at least by younger people, despite a concerted top-down effort to try to convince people that genocide is actually concordant with western values. Though it did take some time.
It's the propaganda that nobody questions that is most insidious.
Oneshotted by refusing to update priors from 1990s-era 'End of History' thinking.
It is always like that. Most people just don't have the attitude of getting things done, and they can barely believe it is possible when they watch what the people who do accomplish.
A lot of things require sacrifice beyond reasonable means. I see these books on how Apple, Nvidia, or Tesla developed their innovations, its groups of people that are extremely talented and became that talented due to sacrifices from their families/communities that go and sacrifice everything themselves to achieve amazing goals. Some of that resultant wealth goes to them but most goes to the shareholders/tech bros.
Eventually less and less people want to go down this route so we get "people just not having the attitude of getting things done".
The real question is will Chinese people go down that same road or will the fact that there is so much cutthroat competition there keep people in line?
It's wild that every comment section about China these days must paint the picture of these rabid anti-Chinese Westerners who are saying that China is an eternal backwater, yet one never sees actual comments like this, and how all of Western media is pushing anti-China propaganda, when the submitted article is just a neutral bit.
A "Manhattan Project" would be building some shocking new technology that didn't previously exist.
If they're cobbling together old parts, it sounds more like something you'd to to keep things running in case a conflict erupts:
> The availability of parts from older ASML machines on secondary markets has allowed China to build a domestic prototype
A “manhattan project” can just mean a massive secret scientific war project? Seems apt.
They have high security, and obfuscating the premises is part of it, but is it really secret in any way ? I mean, we're knowing exactly what they're aiming for and could compare notes at the end of it.
Is it war ? in a "everything is a war" political sense, perhaps, but not in any other sense.
We're left with "massive project" for the analogy, that's kinda weak really.
>> is it a war?
people love to be reductionist... i wonder what aspects of a culture make everyone so black/white us/them ingroup/outgroup. Is it particular to the US, or like, is France the same way? Or Ghana? Or is it just human that everything is a war? Naqoyqattsi.
> A "Manhattan Project" would be building some shocking new technology that didn't previously exist.
You’re missing forest for the trees. ASML at the moment has the monopoly on these machines. This is not only a great tool for the West to keep China at bay, but also a way to maintain economic dominance. Even if they can’t get the machine up and running until 2030, and the machine is a generation behind, China has effectively gained leverage in world theater.
From geopolitical perspective, it’s huge. Right now Taiwan produces the world’s chips, so China plays nice. The minute they can produce their own chips, even an older generation, they can invade Taiwan anytime they want. And then the rest of the world won’t even have older chips.
Taiwan’s geopolitical position is vastly more complex than the fantasy where invasion would follow merely from fab parity.
It won’t. But again you’re missing the point. It’s one less incentive not to, a big one too.
They are acquiring parts to reverse engineer them and build their own
> setting a goal of producing working chips on the prototype by 2028
They might be, but if they plan on getting a factory running in 3 years, they're presumably planning on using what they purchased.
+ industrial espionage to be able to reverse engineer it at all.
If someone likes you, trade secrets flow like wine. That's basic humanity. It's not unique to China, though the relationships involved are a little bit different. It's not a bad thing either, we all live in the same society.
> A "Manhattan Project" would be building some shocking new technology that didn't previously exist.
Once they break even they can overshoot into shocking new technology territory.
>A "Manhattan Project" would be building some shocking new technology that didn't previously exist.
I think that both Germany and USSR were not in the least shocked ... just the USA had the resources to finish it.
Well, USSR did finish, just 4 years later.
Maybe it was because we had all those immigrants working on it (e.g. Albert Einstein, Hans Bethe, and John von Neumann)!
There's a whole regiment of immigrants who worked on the Manhattan project, as we all know. We also know that the USSR obtained much of their knowledge on how to build the bomb through espionage.
Good for them, I don't see this as a big deal other than my fear of west china invading china (taiwan! :) ).
Don't get me wrong, I want the west to succeed, but a competition from China is exactly what is needed. They're building datacenters in arizona and india for TSMC because of this competition.
I really hope we get past historical political rivalry and get along with China better. Competition is good, hostility sucks.
> I don't see this as a big deal other than my fear of west china invading china (taiwan! :) ).
Isn't that "other than" clause a big deal, though? I've read a survey and a number of articles from defense and foreign policy types, and the general feeling is there's a ~25% chance that China will invade Taiwan this decade. That's really damn big. If there's rollback in Taiwan then the first island chain could plausibly fall, or if not you will surely see Japan and maybe South Korea nuclearize. Why must we keep assuming the best with these security calculations instead of believing someone when they keep saying what they're going to do?
> get along with China better.
This will probably never happen. All countries are rivals, and the semblance of cooperation is really just the manifestation of a power imbalance.
China grew into their big boy pants and can hold their own on the international stage. They have no need to be cooperative because they are in the International Superpower Club. Their strategic ambitions do not align with those of their rivals, and they are strong enough to not need to play nice anymore.
Now that the US has also dropped their visage of being the benevolent world leader, there's even less reason for China to pretend to be cooperative. At this point, it's a matter of who is more apt to invade your country, US or China? And you buy weapons from the other one.
Maybe we see more "cooperation" between China and the EU or South America. But that will be entirely because those regions are under duress.
Yeah, all those countries China has invaded really shows how apt they are to do that.
Please spare us. China invaded Vietnam to protect Pol Pot while he was mass killing millions of innocent civilians. They have territorial disputes with over 10 countries, which they've been unable to decisively act on because those neighbors either have nukes (India) or are protected by a more powerful country (US). Not because their government is some benevolent entity. They're basically an authoritarian dictatorship that's kind of cornered at the moment (like Saddam after the Gulf War) but would kill a bunch of people and expand if the US wasn't around.
China has resolved a lot of its border disputes already. The border disputes with Kazakhstan, Krgyzstan, Laos, Mongolia, Nepal, North Korea, Russia, Vietnam, Tajikstan have all been resolved
For all the talk about how they are an equal player on the international stage PRC is still a developing country by their own assessment and WTO.
The more China advances domestically, especially in this area, the less it has to gain from invading Taiwan. China is getting to the point where the conquest is finally doable (rapidly advancing and massive military, plus a weak US president), but the potential gains are diminishing year to year.
I'd speculate that if they don't invade during Trump's term, they never will, and will pursue a different course down the road. China is nothing if not patient.
The motivation to invade Taiwan is rooted in the PRC's political and historical narrative about it's legitimacy and purpose, a narrative internalized by most Chinese, including especially the military. It's in a sense existential, not economic or realpolitik, and I don't see that motivation diminishing anytime soon. If anything it's growing stronger, as evidenced by the suppression in Hong Kong, which made zero sense without reference to how Chinese political institutions sustain themselves. The risk of an invasion sparking a conflict with the US is primarily what held them back, and at best economic and foreign strategic pain only secondarily, but all those risks diminish by the day, leaving China's raw existential motivation unchecked.
The biggest victory for CCP will be Taiwan willingly joining PRC. Nothing else will be a better testament to the CCP model
Reunification with the mainland isn’t a completely unpopular idea in Taiwan. The economic ties are already extremely deep (largest trading partner by far).
Invading Taiwan isn't about chips at all, and in fact chips are actively disincentivizing invasion. Semiconductor fabs and the oodles of atomically precise ultra clean and ultra expensive equipment inside absolutely do not mix well with bombs.
Reunification in Taiwan has nothing to do with chips, and militarily PRC was able to do so a long time ago. The political will in PRC to "kill other Chinese" is zero.
> The political will in PRC to "kill other Chinese" is zero.
Counts for nothing, these narratives are built on sand. Russians also saw Ukrainians as "brothers", as did South/North Koreans before the war, among countless other examples.
Is that's why China has started building loads of troop transport ships recently? To peacefully transport them to Taiwan?
A better title would be "New EUV light source built in Shenzhen". Light source said to be working, not fabbing chips yet. Few technical details in the Reuters article.
The light source is the “easy” bit. The mirrors, masks, and the rest of the machine are all individually as difficult if not more so.
The wafers have to be positioned to nanometer accuracy repeatedly and at high speed! It’s hard to believe that’s even possible, let alone commercially viable.
I'd argue ASML's moat isn't the machine itself but the ecosystem: Carl Zeiss optics, decades of supplier relationships, institutional knowledge.
This is clearly a significant achievement, but does anyone with semiconductor experience have a sense of how far "generates EUV light" is from "production-ready tool"?
They are nowhere close to beat ASML.
This isn't a moat ASML can keep for long though. There can be alternatove technologies to achieve the same goal. So far only China has that incentive. The real problem is process scaling is slowing down. How many more generations of lithography machines will ASML design? Probably not many. This means there will be no edge left in 5 or 10 years, as eventually brute force will work and China will achieve the same lithography resolution.
Till that point, they are just going all in with cheap coal + solar, so even if they use older machines and run longer exposure times, even if they achieve lower yields and toss away a lot of the dies, they are still economically competitive. At the end cheap enery solves a lot of the issues.
Nowhere close, but pace now seems faster than estimated, i.e. original western estimate is they won't even get EUV prototype up until 2030s.
Right now their chips are already "economically" competitive, as in SMIC is starving on 20% margins vs ASML/TSMC/NVIDIA getting gluttonous on 50-70%, at least for enterprise AI. Current scarcity pricing = litho costs borderline rounding error, 1500 Nvidia chip flips for 30000, 6000 huawei chip flips for 20000. The problem is really # of tools access and throughput. They can only bring in so many expensive ASML machines, including smuggling, which caps how much wafers they can afford to toss at low yield. They figure out domestic DUV to 2000 series and throughput is solved.
Hence IMO people sleeping on Huawei 9030 on 5nm DUV SAQP, still using ASML DUV for high overlay requirement processes, domestic DUV to fill rest. But once they figure out SAQP overlay, which will come before EUV, they're "set". For cost a 300m-400m ASML EUV, PRC can brrrt tools at BOM / cost plus margin. Think 40 domestic DUVs and associated infra for price of one ASML EUV to run 8x lines with 30% yield and still build 2x more chips normalized for compute that they can run on cheap local energy to match operating costs. Then they have export shenanigans like bundle 5nm chips with renewable energy projects and all of sudden PRC data center + energy combo deals might be globally competitive with 3/2nm. Deal with our shitter chips for now, once they deprecate we give you something better when our processes narrows gap, and you have bonus power to boot because some jurisdictions, building grid is harder than building fabs.
How does one even smuggle an ASML machine? I'd assume the machine stops working if the GPS position doesn't compute, at end of life I wouldn't expect ASML to allow these devices nor their components to end up on the second hand market, I'd expect the future transfers to require continued permission of ASML, much like weapons distribution.
These machines are not like John Deere tractors. If you own the hardware, you own it. They won't be connected to internet. Security first!
Smuggling part is happening on the old machines before EUV. There's a lot of them available on the second hand market thanks to Europe and US keep shutting down their old fabs. I don't think any DUV machine is smuggled. Even if they physically smuggled one, you need a team of ASML engineers to set it up and calibrate. You can guess what ASML will do in this case.
By the way, let's don't forget: ASML doesn't have any problems with China. They are incredibly annoyed with US and Dutch governments. This is potentially the biggest market they are missing out. Even then, they won't tolerate a summugling operation.
The machines live indoors, far from being able to see GPS signals. Sure, you could require that there be an antenna run to the roof, but you can spoof that stuff.
The thing that helps prevent smuggling of ASML machines is that a) there are few of them (i.e., people would notice), b) it requires tremendous effort to move them at all, let alone without anyone noticing.
it might contain accelerometers, which burn away cryptographic fuses ( setting them all '1' or all '0' so to speak)?
Considering that these tools are installed in seismically active areas [0], the last thing a customer would want is for the tool to zeroise itself because of an earthquake.
[0] https://www.reuters.com/technology/tsmc-says-all-its-sites-o...
earthquakes tend to be predicted a few minutes beforehand, so plenty of time for ASML to sign a temporary exception order for their machines.
> So far only China has that incentive.
The US is close to having that incentive, if the rift between the US and Europe keeps widening. The Netherlands has one lever, but damn it's a long one.
ASML develops and ships their machines at the pleasure of Uncle Sam because the USA licensed them the tech and remains a crucial part of the supply chain intentionally. It's not a lever. It's a partnership that is mutually beneficial and neither side can really ruin the other without damaging themselves.
If Uncle Sam pisses off Europa Regina enough, she won't give a damn about licenses.
I think Europe is bluffing that they can go their own way. They can't. They won't try. Europe has been whining that they're going to catch up since the 80s, but they've yet to do it.
ASML will instantly stall at that point. The EUV light sources are built in the US under US export control regulation. No EUV light source means no ASML EUV machine. I get that some European chest-beating sounds good because there's not very much tech in Europe, but this is an intentional transnational supply chain. It's no accident that the US chose ASML to develop this tech rather than Canon or Nikon. Close ally deep within the US military shield from nearby air bases.
The biggest losers from any such actual attempt by Europe will be Western Europe and the US.
I really like that Europeans are starting to be more patriotic. It's good to see. It's also fortunate that European leaders are aware of Europe's position and role in geopolitics.
Well, it sounds like an alternative supplier for EUV light sources just became available...
An alternative manufacturer, but not a supplier, no.
The US exerts sufficient control over ASML that this will not happen without NATO ending. And the end of NATO (which would be a geopolitical shift more profound than the Fall of the Berlin Wall) and a replacement with some Chinese EUV light source risks the scuttling of all ASML facilities and devices. This is vapor above a coffee cup.
The scenario I'm imagining is in fact the US further destabilizing NATO, in which case Europe wouldn't feel bound by any of the agreements we've made with Americans. Failing that, I don't think any of what was said above is relevant.
ASML owns the company that builds the light source. They acquired it, it's a US company, which is why US export controls apply, that's all. If needed, they could replicate the subsidiary in the EU.
This is too far from correct for any correction to be anything but a full restatement of the facts. Moving the tech over requires US approval. Listen, the Dutch are not going to risk it. Even if they were, ASML would not risk it because all of their customers wouldn't buy anything from a company that's on the EAR Entity List (which is where they'd end up if they tried this without the US allowing it) without US approval. I don't get why people are saying this stuff. It's like saying "Oh yeah, so you divide by zero and then multiply both sides and ta-da". Like, the whole statement is nonsensical.
To enable the whole thing to work you'd need the US to have shrunk to the equivalent of Canada in influence. I'm not saying that's impossible, but in that scenario, the Dutch might well be trying to keep Russians out of Amsterdam and the Turks out of Germany rather than trying to pull an IP heist on the Americans.
You can buy an e-book on Kindle and Amazon still controls what you do with it, right? ASML's ownership of Cymer is like that, except it's the US instead of Amazon.
Specifically control is related to the Foreign Direct Product Rule, where in which the US claims jurisdiction over any foreign product containing 25% or more of US-origins (Cymer, etc)