hckrnws

The "Manhattan Project" part is that the research lab was confidential...which doesn't seem that unusual for a high profile research lab, but that aside.

Comparing China's public efforts to build a computer chips industry to the US effort to nuke Japan is kinda wild. Outside of the bait part, the piece coming from Japan Times makes it that much spicier.

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.

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).

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 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.

> 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,

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.

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.

Call me a cynic, but legislators own stock in these companies. Their true interest in them is also "line go up".

> calling legislators

Good joke. Probably a couple of tech billionaires will eventually say something and then something will happen.

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…

[dead]

[flagged]

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.

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.

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.

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.

They learn really fast. I will not take the past as a prediction of the future.

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.

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.

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.

> 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.

> 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.

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.

"your margin is my opportunity" - bezos

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.

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.

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.

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?

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

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.

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.

aka "return of the Minitel"

[dead]

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?

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

Is AMD abandoning consumer market though? They seem to be much less in demand by AI industry.

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.

There's still AMD and Intel doing dGPUs

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

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.

> 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.

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].

[0]: https://xkcd.com/538/

Is it verifiable now?

Fair question. The technocrats are less than trustworthy.

This is undoubtedly a good news story, and the most wonderful part is that the article mentions that 14 organizations declined to comment on the matter.

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.

and a critical nation scale ambition

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.

>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.

Hopefully this doesn’t stoke a third and final world war.

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.

> 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.

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.

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.

They built the project, the bomb hasn't gone boom yet though.

There's a lot of machinery for moving the wafers around precisely in vacuum. But that's ordinary engineering, although the speeds at which ASML moves wafers are impressive.

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.

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)?

> 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.

> even if they use older machines and run longer exposure times

How do longer exposure times and older machines enable 2nm process nodes?

If you didn't care about exposure time, you could build 2nm chips with brute-force electron beam lithography. But the limited throughput confines EBL to research and very low-volume applications. ASML's EUV-based processes are what permit industrial-level scaling, ultimately because parallel beams of electrons repel each other while parallel beams of photons don't.

I don't personally understand why suitable EUV light sources are so hard to build, but evidently, they are. It sounds like a big deal if China is catching up in that area.

No, they won't beat ASML but they'll be good enough and most importantly cheap. And they'll catch up eventually.

They are "extracting" optical devices from other machines, imagine how desperate they are for this "machine".

As I ironically said in another comment, all you need is a retired Chinese ex employee at Zeiss.

Nothing can stay private or secret forever, and they have the money and people to achieve that. Even if it takes them another 5 years to reach what we have today.

I bet the ex employee doesn't even have to be Chinese. I'm not, but get me FAANG-level salaries and decent working hours I'll 你好 all you want.

我们不会大声说这种话。

Given the current high prices for chips and memory due to "AI" artificial resource scarcity, the world will welcome the additional chip production from China.

I agree. They have a long way to go. There is also something happening in Shanghai but I don’t know the progression.

This article is more skeptical:

https://www.tomshardware.com/tech-industry/semiconductors/ch...

I've read both articles and they say basically the exact same information. Only the tone of this article is a little more skeptical. It also just includes less context/information in general than the featured article

The "problem" with China is that they move from "the amazing thing about a dog playing the piano is not that it plays it badly, but that it plays it all" to Franz Liszt very fast.

I can't wait for China to put its full heft in manufacturing advanced graphics cards, fast storage and much more. We need competition.

The knowledge came from former ASML employees. I wonder if countries will sanction these individuals given the geopolitical implications of their assistance.

> The team includes recently retired, Chinese-born former ASML engineers and scientists — prime recruitment targets because they possess sensitive technical knowledge but face fewer professional constraints after leaving the company, the people said.

> Their recruitment was part of an aggressive drive China launched in 2019 for semiconductor experts working abroad, offering signing bonuses that started at 3 million yuan to 5 million yuan ($420,000 to $700,000) and home-purchase subsidies, according to a review of government policy documents.

I guess they won't leave China anyways. So what's to sanction...

Sanctioning won’t do anything. These former ASML employees know that their professional careers in the western world are finished. I bet they know when they are signing that they are going to stay in China or countries friendly with China for the rest of their lives.

>for the rest of their lives

You overestimate length of the western outrage.

Anyway what's to sanction? Almost no country recognizes Taiwan. Diplomatically they changed one job in China to another

A good benchmark would be how long the west remains mad at Russia for invading Ukraine.

Who is "the west" in your eyes? I personally know plenty of people in the USA who are openly cheering for Russia in the conflict. So I don't think we could say that "the west" is mad at Russia. Certain people in western countries are, but plenty are quite happy with Russia and wish them well in their endeavors.

Why would Holland sanction people who switch jobs? Don't get me wrong, I can absolutely see how it might have happened if ASML had been an US company. I'm just not sure how you figure that it would happen in Europe.

They were under NDA, probably

But the science was probably not

I don't think this is classified technology, although asml would like it if they were punished.

And even if it's patented, China has been stealing everything with little consequence

Patent system is an abomination anyway.

  The team includes recently retired, Chinese-born former ASML engineers and scientists — prime recruitment targets because they possess sensitive technical knowledge but face fewer professional constraints after leaving the company, the people said.

  Once inside, he recognized other former ASML colleagues who were also working under aliases and was instructed to use their fake names at work to maintain secrecy, the person said.

Well real question is how much would that limit PRC talent from working abroad. PRC will be producing plurality of STEM / high skilled talent for decades. They're going to be the only country with project intergrated circuit talent glut in next 10 years, every other semi power projected to have 100,000s shortage. No PRC talent, and you cap western semi talent pool.

Ultimately a lot western innovation run on brain drained PRC talent. There is bamboo ceiling in western tech for east asians, specifically to restrict reverse knowledge transfer. Side effect is once PRC talent hits this ceiling they know big title and fat paychecks and upward mobility is back home, where frankly QoL is off the charts. Ultimately PRC wealthy enough to reverse brain drain aka brain recirculation and PRC talent aren't retarded enough to limit their career aspirations because west decides to cap their career trajectory and try to lock their future behind noncompetes, especially in cold war vs their birth country. Worse, PRC wealthy enough even if there's no bamboo ceiling they can afford to reverse brain drain top 1%, hence current equilibirum. West needs PRC talent, west cannot afford PRC talent to climb too high, PRC can afford to take them off west's hands.

Until west figures out another source of talent, they're stuck in this talent trap. And IMO India ain't it, they don't have the integrated industrial chains and academic structure to produce same kind industrial ready workers yet.

Controversial and possibly politically incorrect take, but the People's Republic of China sends many, many, of its citizens to study at top universities and work at top companies all over the world. I'm sure even at sensitive defense related orgs too.

While I am sure that the vast majority of them are just regular people, I'm also pretty sure there are True Believers amongst them whose mission is to go out into the world and enrich themselves with the skills and knowledge to bring back to China and further the CCP's goals. Some of them might even attain citizenship in the country they go to while inwardly retaining full allegiance to the PRC.

Heck, I know people from other, friendly/allied countries who obtain US citizenship who, if you pose the hypothetical question "If your former country and the US got into a shooting war, who would you fight for?", they would pick their former country without hestitation.

And despite public policy and rhetoric sometimes stating how the PRC is becoming a rival or even existential threat to the Liberal Democratic World Order (TM), the Western democracies don't do anything to secure things. And quite frankly, I don't know if there is anything that could be done, short of getting into... highly controversial territory. Which if the situation were reversed, the CCP would probably not bat an eye to do.

> knowledge to bring back to China and further the CCP's goals

You're forgetting to mention that they're also getting paid a lot of money. Quite a lot of people will sell out, given the right conditions, for that amount of money especially in lower CoL areas. To be honest, I'm sure Western governments and companies could do the same if they wanted to bring in the expertise from China.

In late 2022 our telco soft eng team got purged and everyone who was even friends with people who might be Chinese were removed from the project. That included the original architect and product owner, both Americans but with Chinese roots. So there that!

Is there any other way to see it than just we are too divided and 50% of our own people just think we are the bad guys? What you describe is so obvious but one political side in the US at least pretends this isn't happening and actively does anything they can to hamper any response to it. I would love to be convinced otherwise because I am also part of the division, I truly don't understand the other side at all.

I think there was a time when the other side truly believed globalization and economic progress would turn the CCP into a democratic ally. Maybe both sides believed that for a while. What you see now is just the fragmented and incoherent remains of a failed philosophy that hasn’t yet come up with a coherent replacement, so we’re left adrift with no rational foreign policy from either side (in my opinion).

Controversial take: Democracy and the US are awful at keeping secrets, and are incapable of winning by an information delta, if we followed your strategy we would surely be doomed. Our greatest advantages come when we work in the open, and share knowledge and empower ordinary people and the world with technology. As things stand, we are funneling our brightest minds into creating proprietary (secret) technologies... And it turns out the only people for whom the technology is uncopyable or secret are... American citizens. The "proprietary" technology is trivial to steal, and legal protections don't matter outside of our borders, the legal protections and subsidies afforded to those building proprietary (secret) technologies only deprives Americans of the ability to innovate, while in peer nations like China, individuals and startups are totally free to use and enjoy American technology without any restrictions.

But that only works if China reciprocates, which they show no sign of doing.

I’d imagine a Chinese citizen living, studying, or working in the US has access to a lot more advanced knowledge than a US citizen trying to do so in China.

Up to this point, the US has been the one with the advanced knowledge. We now face a world where the opposite might become true.

But using the previous example, I’d imagine a future hypothetical American going to China to study or work would face a lot more roadblocks to obtaining and extracting any advanced knowledge, especially anything with strategic importance.

We haven't always been awful at keeping secrets, see the actual Manhattan Project. I like the optimism of your proposal, but how would those US companies continue the same level of R&D investment without those extra profits? If the government just directly invests, then you've just become the enemy.

Didn’t the actual Manhattan Project leak to the USSR?

It seems most of those ASML employees were already Chinese engineers. I doubt they would care if they got caught and had their careers restricted to China

Handing out sanctions without at least a plausible legal cover, sounds like a recipe for disaster that would come back to bite.

I wonder what could be used here, non-compete? IP infringement? Or doing it "for all mankind"?

As for knowledge, the YouTube channel Branch Education explained EUV lithography in great detail, sponsored by ASML itself.

My impression is that the knowledge is not that secretive, the precision required at every step is the key.

Yeah it reminds me of the Smyth report, published in August 1945 about atomic bombs, commissioned by the director of the real Manhattan Project. It’s fine to reveal knowledge in detail, if it doesn’t reveal anything related to constructing the apparatuses (the chemistry and the metallurgy) needed.

It seems extremely dishonest to frame the project of improving computer chip manufacturing to the development of weapons of mass destruction— weapons that went on to be used against civilians. Sensationalist and propagandistic framing for what is otherwise an interesting article.

The term 'Manhattan Project' is a common and widely used metaphor for R&D programs with effectively unlimited resources applied to them. The actual Manhattan Project is simply a very famous exemplar of such a program.

Use of that term is not propaganda, it's normal English.

It's not just about the use of the term "Manhattan Project". It's about the framing and wording of the article. There is literally an image of a PRC soldier in front of a rocket in the article.

When referring to the efforts nation-states, I'd be very interested to hear how often such metaphorical usage is used to describe the work of adversarial vs. friendly countries. I would be shocked if it's as often (in the Anglophone press) used to describe the work of US-aligned countries as it is that of US-adversarial countries.

This is literally by the country that suffered the most from it.

The interesting part here isn’t “can China copy ASML’s machines,” it’s whether they can copy ASML’s ecosystem. EUV is a stack of insanely tight supplier relationships, Zeiss optics, service networks, and years of painful yield tuning, not just a light source in a lab. China can absolutely brute force its way to “good enough” over time, but what still holds them back is everything you can’t buy on the secondary market: trusted optics at scale, field-proven reliability, and the boring industrial plumbing that lets a tool run 24/7 in a fab without drama

Most of that entire network is easy to replicate, as in it's not technically hard, the hard part is validation, no one in PRC wants to use unproven PRC inputs if risk is 100m wafer run goes to trash. Hence PRC now basically insuring domestic fabs on risk runs using domestic inputs, which are being validated on full scale production, instead of taking 5 years to verify, it'll take 2. Export controls help this, i.e. domestic resist basically required now after JP export controls.

The hard part, i.e. optics, light source. Zeiss had like 3k engineerings, Cymer 1k, ASML 13k during EUV commercialization process. PRC can (and is) just throwing bodies at problem, lots of parallel execution with clear second mover road map. That and as this article suggest, they're literally poaching people with the tacit knowledge which will help speed run. I'd wadger they get there sooner than later.

China seems to be doing well on supply chain integration (with the exception of the trust part).

Being how strategic this is, I imagine that the investment won't be entirely laissez faire and there will be lower tolerance for cheating in this endeavor. I think that ultimately they'll do quite well with their efforts.

It's kind of nasty that a fresh society of capable people has the drive to achieve technological excellence and the incumbents do whatever they can to delay this, even though it's inevitable and there's a lot to gain by empowering them. All in the name of "they are not us".

World has gained so much from modern Chinese industrial revolution. Why suddenly everyone got cold feet? Nobody was stopping Germany or Japan on their way up even though they were literal former enemies with history of brutal warfare. China never done anything even comparable to others.

I mean, if we’re going to make that comparison, China today looks much more like pre-war Germany and Japan— set on expansion. That’s pretty clearly what the anti-China crowd is worried about. Tibet, Taiwan, Philippine islands, ever expanding naval bases, aggressive displays of power around Australia, prison camps and sterilization programs, and so forth.

How does it compare to the last 50-70 years of US actions? Even if all narratives are true and significant, China could still massively benefit the world same way US did despite its vast shortcomings.

Also interesting huge project: China is building a $116 billion dam which, according to Bloomberg, is expected to generate 70 GW, just to compare: UK whole capacity (de-rated) is around 70 GW.

https://www.youtube.com/watch?v=zAwJESmfy10

Chinese dam projects have not always gone well:

https://en.wikipedia.org/wiki/1975_Banqiao_Dam_failure

intensive vs extensive quantities

Seems like demographics, AI, and tech parity are converging on a Taiwan takeover attempt in the 2027-2030 timeframe.

If China actually catches up and surpasses the West/TSMC in fab technology and production, I think they'd have a better option, which is simply flooding the world market with high-end chips and obliterating the Taiwanese economy. Eventually, joining an economically dominant China might become more palatable, or a necessity.

At this point I'm willing to wave around the little red book for a 1TB of ram.

I don't have that many kidneys left to buy gpus, ram and ssd at the prices they are now, let alone the prices next year.

This would be more in-line with their strategy in other areas. Quietly massively improve technical capability and then utterly out-compete international competitors. They did this with solar, multicopters, are in progress with doing this with TVs, nuclear power, etc. War is expensive and destructive, it's easier and nicer to just negate the economic relevance of your opponents if you have the time and resources to do it (which they do).

China is also doing this with weapons. It's just a little more difficult normal people to see the results because people can't get a Dongfeng 2x series rocket from Ali Express.

Realistically, the general public doesn't have access to an honest appraise of their capabilities. So we are left to infer from their accomplishments in other high-tech areas what their military industry is capable of producing.

How much money would Taiwan have to be offerred to voluntarily place their heads under the boot of China

The leadership will have a price in mind and they won’t be the ones under the boot. Everyone has a price to look the other way even if they think they are principled now.

China is building all kinds of suggestive tech including invasion barge piers to land heavy stuff quickly once a beachhead has been established http://www.hisutton.com/Chinese-Invasion-Barge-OSINT.html

I know mass media keeps pounding this "eventual scenario" (manufacturing consent and all). Maybe it will happen, but the Chinese think on longer timelines than the ADHD West and are probably banking on A). Out-attriting, B). Out-innovating. If both happen, we might find ourselves with a situation where Taiwan voluntarily wants to align closer with China as the West flails.

China know that the one child policy has fucked their demographics and that their future isn't as rosy as it might appear now

Why would the nation that implemented the one-child policy be unable to implement a three child policy?

Reversing demographic momentum isn’t so easy. You have a cascade effect which happens, and high costs of a massive elderly population can’t be borne by a sudden baby boom. Also, through the medical system, etc, you can force people to abort their children, but it’s harder to force them to make children.

They will just tax condoms to sponsor the elderly homes.

And that's more of the "propaganda narrative" (google that term) getting pounded at us from all channels over the past several years. It's so 1984 - remember when we were told they were growing too much and getting overpopulated and the planet was going extent any moment? So which one is it? Anyways, their population decrease will be offset by AI and automation all while they still pump out more honor students than the entire West combined. They'll be fine is what I'm saying.

> manufacturing consent

I think it's more like smearing/projection, like Republican conspiracy theories about Democrats being pedophiles. Guess where the real pedophiles were hanging out the whole time.

The track record says China will probably just buy Taiwan.

If you hate invasions so much, you should probably focus your energies on Venezuela. Looks like Trump might start a war for Christmas.

Wars are old fashioned. This is a “special military operation”

is 1950 old-fashioned? The Korean War was originally called a "police operation"

In 1945 as World War 2 wrapped up and the Cold War started, many in the US believed that it would take the Soviet Union 20+ years to build the atomic bomb. It took 4 years. There were several reasons for this. It became a national security interest, there were leaks to the USSR by people who thought the US shouldn't have a monopoly on the bomb and Americans in general viewed the Soviets as backward farmers.

I see the same thing with China. It's not so much espionage now (although there might be that) but China instead will just hire people with the right knowledge, so former employees of ASML, Nvidia, TSMC, etc.

I've been saying for awhile that China won't tolerate the export ban on ASML's best lithography machines and NVidia's best chips. It's a national security issue. And China is the one country on Earth I have faith can dedicate itself to a long term goal.

And yet I got the same reaction. "The Chinese will never catch up", etc. Reports have been comiung out that Huawei has started developing and using their own 7nm chips.

Weirdly, the US created this problem. By restricting exports of chips to China, Chinese manufacturers had no choice but to develop their own chips. Had China been flooded with NVidia chips, there would be far less market opportunity.

The American economy is essentially a bet on an AI future now. Were it not for like 7 tech companies, we'd be in a technical recession. I also believe that bubble is going to burst. But the economy as a whole pretty much now requires US dominance of an AI future and I think a lot of people are in for a rude shock as China completely disrupts that.

China hasn't caught up yet. There are still many steps in the supply chain and chip design as a whole but making their own chips at sub-7nm is a massive step in that direction.

> It was built by a team of former engineers from Dutch semiconductor giant ASML who reverse-engineered the company’s extreme ultraviolet lithography machines (EUVs)

This seems like the obvious conclusion of an ethnic bloc against a mercenary creedel nation?

Any westerner reading this right now wouldn’t die for their country, it’s almost absurd. It’s like asking them to die for Walmart.

archive link??

It's quite easy to do it yourself - just open archive.is and paste the original URL in.

https://archive.is/tKZmn

FWIW, this seems to be a Reuters report reprinted in Japan Times. Previous HN discussions got just a couple comments:

https://news.ycombinator.com/item?id=46301877 https://news.ycombinator.com/item?id=46307819

> China’s prototype lags behind ASML’s machines largely because researchers have struggled to obtain optical systems such as those from Germany’s Carl Zeiss, one of ASML’s key suppliers, the two people said.

So, now they just need an old retired Chinese that worked for Zeiss and build a prototype for the optical devices they need.

They use armies of graduates just to literally copy, when they could build something new or different.

We learn that before 2023 EUV lithography was worthless. "AI" is the only reason why China would want this technology!

EDIT: Given the dramatic downvotes, I repent: China will use these EUV machines to build AI sharks with lasers that will swim towards Taiwan! Is this better?

How did "we" learn that?

By reading the headline? Ask your favorite clanker, it will understand.