Rather, the only thing that will truly motivate TSMC to take on more risk is competition.
If big tech wants TSMC to increase capacity drastically without TSMC having to take all the risk of CapEx, then they can pre-pay for wafers from TSMC.
They can each give TSMC $10b now in cash and guarantee themselves wafers in 2-3 years that it takes to bring a new fab online.
TSMC is rightfully conservative. If they commit to spending an extra $30b on a fab now that won't make a single wafer until 2029, without any guarantees from big tech, they're stupid. Who knows if the demand will still be there (my guess is yes, but who knows?).
In my opinion, I think it's getting close to this. Nvidia will surpass Apple as TSMC's biggest customer this year. This will start a war for TSMC wafers in 2026 in my opinion. When you have that much demand, customers will be forced to pay well in advance.
There is already a war for memory, silver, copper, energy. No reason why chip production won't be next.
I do wish Intel an Samsung would cooperate on open source EDA (etc) software to make switching to other fabs less risky and capital intensive.
This has always been the case. And the article is not well written, but generally speaking everything written about TSMC or any sort of manufacturing has been wrong most of the time.
Just ask Intel on what happen if they overbuilt capacity and their node didn't work. The most expensive Fab isn't the most advance one, but the empty one.
As a matter of fact, if everyone wants it so bad TSMC could bring up a whole new foundry within 18 months. With lots of pre-planning and assuming Taiwan Government in full support. The trouble is not a single big tech company are willing to bet it all by themselves either. Not until now.
However, I'm talking about booking wafers from a fab that hasn't started and won't make a single wafer 3 years from now. The scale is different. Imagine Nvidia, Apple, Google, AMD, Microsoft, Meta, Amazon telling Wall Street in their earnings report that they sent TSMC $5 billion each this quarter and won't receive a single wafer for another 3 years from the investment. I fully expect this to happen soon. I'm almost certain that you'll hear in the upcoming earnings reports that big tech sent Samsung, Micron, and SK Hynix billions in advance payment to secure memory supply years from now. I think this is likely the same for chip fab capacity soon.
And this would be no different than investing billions in R&D (Ex. Meta and AR) for future payouts.
Or, Apple buying 10000 advance CNC machines for their manufacturer. In this case timeline for future payout is perhaps much shorter but the pertinent point will be Apple invested in Capex upfront.
My perspective on the China risk differs some, though. China wouldn't benefit much from attacking TSMC. This is the first time I've heard anyone suggest that they might. At best they'd like to have it in-tact if they do take Taiwan, but there have been talks about machines being rigged to explode to deny them from China, or the US striking them in that scenario.
If neither we nor China get to work with TSMC, then we're still ahead in relative terms. If China did attack TSMC, they set the norm that the fabs (including their own) are now a fair target which would be a larger disadvantage for them than it would be for us since China's physical power projection remains pretty regional outside of Chinese nationals abroad engaging in sabotage.
That is one of their biggest weaknesses. Yes they have a lot of manufacturing capacity and a large population with many talented people, but in a way we have lent them the power to scale up to see what they'll do. We are already putting some pressure on that scale now that they've shown who they are, but if it came to war it would be very doable to start reversing their scale and their capacity to do the same to us would degrade as ours increases.
Even if all the AI in the world was destroyed, that's how it would play out. The problem is that Taiwan remains in close proximity to China so similar to Ukraine it would likely come down to how long they're willing to throw everything at it.
If Russia and China wanted to be powerful, it's just idiocy to show the existing superpower that you cannot be trusted with the power you have. If they fancy a merit based society, they forgot that merit isn't omnipotence and you still need the right ideas to be at the top to accompany the merit. For China maybe they need AI for that alone, but western societies at least have ways for the right ideas to make it to the top without the strict need of AI.
Big tech manufacturers are treating consumers like crap by selling chips and memory to over-invested start-up companies that will go bankrupt, as these products will not be profitable due to the high costs (the technology that would make them profitable does not exist and has not even been conceived), in addition to the low long-term quality of what they offer.
The thing is, right now, billions of consumers around the world see how those big tech manufacturers are not serving them the pieces they need, and that such techs will not do in the near future with fair prices (prices abuse escalation, the consumers lost their strength).
Right now, if China takes Taiwan, 2026-2027, even if they lose the fabs, the billions of consumers in this planet will see this as a real f** you big techs, f** you overinvested startups hoarding, go go China, as we realize that we are third category citizens in this "first the riches" spiral, and will be no much difference of what is going on now.
If China takes his media news cards right, and makes know the consumers this is a revolution, and combine it with one of the numerous Taiwan's corruption scandals, I bet they will not find the opposition from citizens around the world that they would find in a different period of time.
Besides, if you want to optimize for pure consumerism, you can just look at how roman slavery worked out for roman freemen in the late republic.
Good try. We are talking about a window of X years were the average normal consumer, average people, will not be able to buy tech, and should not buy tech, due the prices will put, has already put, average low-end technology at prices of extra mega top tier technology, anything with a memory or disk or CPU ( too expensive, this is an abuse, it is better to wait given the current terms).
This pause were the tech supply is not destined to the average consumer (who is being abused due this) is the perfect window for China to take Taiwan if they plan the strategy well.
And it is not needed even to predict, Chinese government just need to follow/observe the prices and discomfort, and if the pattern follows, it is the moment. The average consumer, the people, even will aim them.
None of this is "unaffordable" as you say as it is just a month's worth of savings.
Average low-end technology at prices of extra mega top tier technology.
You are saying people will buy, and smile without resistance saving months to pay those abusive prices for low-end tech, even knowing this is because four or five companies are hoarding the market. Take for sure people is really tired/sick of this, check the forums.
China, are you reading this? You should study this along the months. Then you can make your own conclusions.
[0]: https://www.youtube.com/watch?v=Kcae6aiV1vs
Imagine if China has a foothold in every industry. Sure the US can tariff itself but the rest of the world is not really competing in most of those industries and so consumers will be able to see that they dont have to settle for overpriced junk anymore. What will American/European or even other sian companies do? In America most companies have financialized so much that the underlying product that made the company famous have rotten in quality.
I recently was blown away Laifen's P3 Pro electric razor. I always thought I would be a loyal Panasonic customer for life (since I had family work for the company) but here comes this Chinese company from nowhere and they produce such an amazing device at an amazing price. I never thought having a CNC milled pocket razor using some sort of tiny linear motor would be something I would want but now I can't see life without it.
They are doing it to every industry. I always accepted things like 3D printers were gone thanks to Bambu but I now have to consider every industry at risk.
It will tank the markets because people will assume a depression-level event and WW3. But Trump isn't like other presidents. He'll make a deal with China. And finally, the China/Taiwan cloud over the markets will go away for good and countries can start trading freely with China again. Markets will severely over react initially.
I can see TSMC benefiting hugely from this long term, as long as the reunification is peaceful no damage to any TSMC fabs or people. The reason is because TSMC will most likely be forced to open up to both Chinese and US customers. Right now, they can't serve the world's second largest market. Nearly half of their customers can't use them.
I'm making these assumptions:
1. China won't use force (or very very little) to take Taiwan.
2. There won't be WW3 that will come out of this. You'd have to be an idiot to think that Americans will die defending Taiwan or that Europe will send troops when China is quickly becoming their biggest trading partners and US has shown they're susceptible to annexing Greenland.
3. China will operate 1 country 2 system long term with Taiwan.
2. If Russia, North Korea, South Korea and Japan join in there is a lot of potential for it to scale up. Whether it would become an all out horrific war like a World War or stay a little bottled up, it does risk becoming a huge conflict. Many Americans love South Korea and Japan, though they're less informed about Taiwanese. If South Koreans and Japanese are dying, we will be involved in one way or another.
3. No it won't. Look at what happened with Hong Kong, it broke its promise, just like the CCP breaks many of its promises. Not sure how bad they are compared to Russia in that regard, but it's pretty bad. Besides, if China wants to expand the way it seems like they want to, they need to take Taiwan so I doubt they would slow roll it.
3. Hong Kong is still a different system last I visited (2024).
Let's set X = oil. Oil is a critical resource that is huge for transportation. If you are the one that controls it and people need it from you, that gives you leverage to encourage them to do what you want. An authoritarian country that engages in mass killings and torture of its own people could get away with a lot and even abuse other countries if it has enough oil.
You can apply this to nukes and many other things. Like maybe you defend an ally. Why defend an ally? Having them makes you stronger. Why do you need to be stronger? To defend the way of life you believe in.
If the US truly does something horrifically wrong, we're free to report on it in our own country. Even with the Vietnam war, we reported that we were explicitly killing women, children and babies as a result of some realities on the ground. Just a few years ago the Biden administration drone striked a vehicle of school children on accident and admitted it. We waterboarded known terrorists at Guantanamo Bay. For a country the size and influence of the US operating all over the world, you can find many examples of problematic activity. Sometimes the "why" we did it helps it make sense, and sometimes it doesn't.
One case where the why made sense, is when we nuked Japan. Nukes weren't the most deadly option and it was done during horrific times that even seeing film of the war you cannot fathom. Other methods were already killing far more people. Nukes were devastating in a different way, but the benefit of using them was to try to stop the war to save millions more. Back then, precision strikes were barely a thing so collateral damage was an accepted and understood aspect of war.
Now we have enough sensors in space, on ships, in the air and so on that we can be more precise. We're much better at avoiding casualties like that now, but our adversaries also know that we try to avoid them so they use innocent people as human shields. Now, many of our strikes are so precise that they're just labeled as assassinations now. https://en.wikipedia.org/wiki/List_of_assassinations_by_the_...
Russia and the CCP can't even admit the things they've done which have killed millions in their own countries not to mention other countries, so they can't improve and their populations can't press them to be better. Instead they shame, attack, arrest, assassinate and execute their own people to try to justify to their countries that their tough authoritarian control is necessary.
In Ukraine, Russia has been attacking apartment buildings, hospitals, schools, power plants, police stations and so on intentionally and repeatedly. Iran just mowed down thousands of protestors. I don't doubt that some were armed or that there was violence, but it looks very much like they went far beyond any justified level of force. Venezuela was arresting and torturing people with opposing political opinions, sometimes gunning them down openly in the streets.
Much of our popular media still encourages kindness and the preservation of life.
People talk about inequality in the US and injustice, but poor people in the US are better off than most other countries. Most of the poor people I know still manage to have large TVs, hot showers and eat Taco Bell, which is not the kind of absolute poverty you see in countries where people are eating out of the trash and washing their butts with water trapped in potholes in the road. Many people DO get free healthcare and actual real healthcare costs are often much lower than the numbers people publicize since those are pre-negotiated. There are also free food pantries all over the country. There are places that offer free clothing and much more. People donate stuff all the time. Our philanthropists have invested all around the world to stop diseases and improve access to education.
Our military members regularly risk their lives in dangerous situations to fight back against criminal pirates, actual terrorists (not just mislabeled ones), violent dictatorships and so on.
Whenever we are in a country and doing business, people's lives are generally better. Venezuelans were better off when the US was there. Under Chavez and Maduro, who are supposed to be directing resources to the people, they have been worse off. This is just one interview: https://www.youtube.com/watch?v=cylfhA8pNgY
There is a basic logic usually at play. Are we killing someone who has better morals than we do? How would we compare the morals? Most of the people the US goes after have clearly worse morals and a history of evidence that backs it up. We're not out in the world killing nice Thai people or rounding up Canadians to put them in mass graves. It's generally people who already don't respect the lives of others. Look at Singapore. It's a dictatorship, yet we're not over there wrecking them.
You should start asking yourself if the US is so bad and a giant evil empire, why are we not in the places that we're not in? What are the people inside the US actually like? You know, it is the people who end up becoming the government. So who are the people? Almost universally in my entire life in the US, my encounters with people have been kind and positive. Not just surface level, but giving and considerate. Someone who didn't even see me and doesn't know me bought me breakfast recently just to be nice.
Social media and emotional entertainment comedy news programs are not reality. There is an information war happening against the US, because it is a vector they have a numbers advantage on and you can make a lot of noise there without needing as much money. There are bots ALL over youtube now spreading lies in comments. It's tricky to figure out exactly how to handle it without censoring real organic opinions. The CCP seems to identify inconvenient news and spread the link to a network of people that bombard a story. There's no way the algorithm would be recommending these stories to an overwhelming majority of people who have these negative opinions and you wouldn't get a giant pile of these comments within the first 50 minutes of it being posted either.
Just have to make up your own mind, but first you have to get good at making up your own mind. If you don't, someone else will make up your mind for you.
Um, based on the HK experience I would revise this to "2 countries 1 system".
If people generally understand that we're intervening in a China/Taiwan conflict for the right reasons, and China attacks us at home it would only accelerate the west. If you look at Vietnam, the logic around that war fell apart and it no longer made sense to continue it. The people were right to push back. Iraq and Afghanistan went on for a long time without much fuss.
Some kind of conflict in South East Asia would likely largely be a naval and resource war, with many casualties being naval rather than mainland. Most losses on both sides will probably be drones, AI or not.
If it came down to attrition, it would maybe be AI machine attrition or drone/missile attrition which is in a way a resource war which the US could win even without TSMC, but from where we're standing today it would take more ramping up which is a process that has already started.
If OpenAI, Google, Microsoft, Amazon, etc were attacked by China in a more critical way, it would have to be for some major short term advantage that is capitalized on immediately, because long-term it would be a losing strategy by itself. China has systems to disrupt as well, so if they let loose on cyber then the US has options too.
Either way, don't conflate general economic preference around an election for whether people would tolerate being unable to access Gmail or order from Amazon like they would all rush to riot in the streets. I think that misreads the situation.
>If it came down to attrition, it would maybe be AI machine attrition or drone/missile attrition which is in a way a resource war which the US could win even without TSMC, but from where we're standing today it would take more ramping up which is a process that has already started.
I'd argue China adding 1/3 of the entire US electricity capacity in a single year and increasing along with their extreme battery overcapacity makes AI and drone production something that China will win. Like I said it remains to be seen how will the US military will hold up because they do have the Arizona TSMC facility running and that could be a buffer to help the US hold on but people will still feel the pain in massive inflation in all areas and thats where peoples selfishness will rear its ugly head. Why care about Taiwan when the population could just elect someone that will negotiate a short term win for the US (at the expense of a long term loss).
>If OpenAI, Google, Microsoft, Amazon, etc were attacked by China in a more critical way, it would have to be for some major short term advantage that is capitalized on immediately, because long-term it would be a losing strategy by itself. China has systems to disrupt as well, so if they let loose on cyber then the US has options too.
So I have been in meetings for things like Franklin Project (https://defconfranklin.com/) which are among many great initiatives the people are doing to help the US prepare for an initial attack where China disrupts all the little mom and pop orgs scattered across the US in a first strike to disorient the homeland. Is it enough? I dont know, we will have to wait and see. I dont know what China is doing to prepare for a response. It seems like their AI initiatives are a pragmatic move (use low cost Ai implemented at various layers across the stack) I do worry that in the last few years its been revealed how little so many Americans take education and critical thinking seriously and that will directly translate into sloppy IT infrastructure around the country.
>Either way, don't conflate general economic preference around an election for whether people would tolerate being unable to access Gmail or order from Amazon like they would all rush to riot in the streets. I think that misreads the situation.
I think you misread my point: Americans would make themselves heard by electing someone who will deliver a quick relief for them at the cost of long term loss.
Fresh off WW2, with a titanic arsenal and industrial base, America and all of its allies couldn't end the war on their terms after China intervened.
That's why the US only did search-and-destroy missions, targeting Vietcong cells in the south and bombing supply lines in Laos. Which didn't matter much.
Once the Americans left, the North marched down a proper army and wrapped it up.
Another note most of the world understands US's strategy for this type of disruptions used as a weapon, if they can cause civil unrest, the US can use this against the government and it will into the advantage of the US. I would like to note this is not a country of easy convincing, The majority of the seating members of the CCP are deeply interested in their society's interests, then globally second. You have 1.4b people this is not an easy management to handle.
In my eyes best to not compete and work together. There are hard obstacles ahead that are going to need all of our efforts collectively, in Hines-sight this is childish and a waste of time, literally. We made it this far, all of the great achievements and innovations that with out a doubt all have collectively contributed as humans. Climate changes, Populations management, Biological threats, viral threats, pandemic management, genome advancements etc...
Some of the moves that leaders make are simply moving us back in time and when one does something unfavorable to the other it sets the tone for how future engagements and decisions are made, whether silly or not. Indeed no leader is innocent in their decisions, but my point still stands.
If they could not be insane and trade, that would be great. Unfortunately that's not the world we live in and the US has to push back against it or the world can fall into ruin. So we're cleaning up Venezuela. Maybe Cuba and Iran.
Can you give some examples of why you think this? I really can't imagine how this would be true.
Best examples would be in last 25 years when they've had mass affluence.
Pain tolerance might be the wrong term. Pain tolerance implies speaks to something intrinsic about a population, while really what we're looking at is how much discomfort a population can endure before it really agitates for policy/political change, and so it's much about how a population feels, as the tools available to the government to control, manage, deflect and address the pain/discomfort.
I do think that the US population is able to bear incredible levels of pain if it's packaged a certain way. Examples:
-20 year Global War on Terror which cost $6T+
-Healthcare costs which far outstrip other western nations, mostly paid for out of pocket, and which increase every year
-Opioid Crisis which killed more people than all our 20th century wars combined
-Lack of workplace protections, time off, etc which our peer nations enjoy
The Chinese have not dealt with any of these things, so yeah, they have more available capacity to manage new social disruptions. That said, Americans love war, so we could probably add another war without disrupting things too badly.
The availability to good healthcare for many conditions in China is quite subpar compared to the US. They don't have many physicians. Their healthcare outcomes in most things are worse than those for Americans. Good example is their lung cancer morality rates.
Of course both these things are expected for a much poorer country.
This was enabled by the US population not even feeling any of the effects of the war: https://www.cbsnews.com/news/were-at-war-americas-at-the-mal...
"Healthcare costs which far outstrip other western nations, mostly paid for out of pocket, and which increase every year"
This is a reason Trump was elected. I already mentioned this in costs. Imagine what will happen if China takes Taiwan, inflation WILL creep into this with increased medicine costs (many pmade in China) and throughout the stack.
"Opioid Crisis which killed more people than all our 20th century wars combined"
Regional pain does not equal national pain. Ask West/East coasters and they dont know this problem as deeply as the deindustrialized Trump voting rustbelt.
"Lack of workplace protections, time off, etc which our peer nations enjoy"
Why do you think AI is being pushed so hard. This is a coming pain that I hope will finally push the US into a true leftist position.
"The Chinese have not dealt with any of these things, so yeah, they have more available capacity to manage new social disruptions. That said, Americans love war, so we could probably add another war without disrupting things too badly."
Do you have any evidence to support this assertion? Do you live in the US?
Do you really think the Chinese have not dealt with bad workplace conditions? The US population may be jealous of some of their neighbors but they have significantly better working conditions than most Chinese.
From 996 from Alibaba to the swaths of cheap, manual labor used for outsourcing by, among others, the US.
The high population with oversupply of STEP means everything is more cutthroat there. They have 25% youth unemployment. 996 produces lots of problems but also higher pain threshold. When a new idea comes along they move at lightning speed.
Oversupply of companies in every industry. For example: China mandated companies start producing EVs. Next thing you know there are 100+ car companies in China. All creating jobs that local areas depends on. Now the EV incentives are gone and it has led to extreme price war and everyone fighting for survival in the jungle. What do you think those surviving companies will be like once the weak have been killed off? Imagine them coming after American companies who are fat and bloated and may not know whats coming for them (the employees certainly don't).
This ability to push harder than anyone else is already paying dividends. In 2024, China added 429 GW of new power capacity, more than one-third of the entire US grid. Thats many more factories producing weapons, data centers crunching AI, and lower energy costs enabling more new opportunities.
Now take these people and put them on a war footing? They will out manufacture, out speed and with the amount of STEM they have possibly out wit the US.
The ability to mass produce a Pascal or Volta comparable GPU or Apple A11 comparable SoC is all you need for more cutting edge systems.
Power Electronics and Compound Semiconductors (GaN, SiC) have historically been the biggest bottleneck.
The bigger risk for the TSMC-China aspect is TSMC's planned exit of GaN foundry production by 2027. Most Chinese manufacturers still depend on TSMC-produced GaNs wafers instead of domestically produced GaN vendors due to reliability concerns. China will probably end up matching TSMC's specs for compound semiconductors in 4-7 years, but that implies that the Sullivan Doctrine still holds and is a loss for China.
Every other country with compound semiconductor production capabilities at scale (US, Germany, France, South Korea, Japan, Israel, Russia, India) either limits their exports or cannot export them to China without facing sanctions from other buyers (primarily Russia as India does not allowing commingling of SKUs for defense vendors who sell to Pakistan/China as well, and Russian vendors are members of India's EW and DEW program).
If a US-China War was to arise, worst case we would probably see a mass reversion back to 2018-22 level technology, which isn't the end of the world.
Additonally, the China-Taiwan situation is orthogonal to semiconductor dependency.
The argument is that existing weapons systems are essentially old tech, left over from previous wars in previous decades. Many of them are less useful in the modern battlefield, e.g. defense systems built to shoot down missiles are easily overwhelmed by drones.
If there's a real war, it will be fought with next-gen weapon systems - probably autonomous drones that will require high-end AI chips.
It wouldn't be the end of the world if those 2018 chips came at 2018 prices and only impacted commodity stuff like phones, datacenters and laptops, but they'll be at 10x the 2018 price and impact critical stuff like automotive, the cars and trucks that gets your food delivered to the supermarket, and if those become impossible to buy or fix anymore, then your groceries will also get more expensive, triggering an end-of-the-world riot from taxpayers who can't afford food anymore.
People need to view this issue as not just being stuck with 2018 laptops and phones which isn't that bad, but has much wider societal implications.
Most capacity for legacy nodes is already ExTaiwan and ExChina. TSMC leapfrogged American and Korean fabs in the late 2010s/early 2020s with sub-14nm process nodes, but Samsung and Intel have caught up for 5nm and 7nm capacity. And Taiwanese firms have largely diversified OSAT and ATMP away from Taiwan and China to ASEAN, US, and India.
This is something everyone adjacent to this space has been thinking about and acting on since 2017.
> only impacted commodity stuff like phones, datacenters and laptops
Data Center and enterprise applications are prioritized in most ExTaiwan sub-7nm fabs such as Intel 18A.
---
A US-China War over Taiwan would be devastating, but not getting an M1 Macbook will be the least of your worries.
Isn't this what I said?
What I worry more about is the full lock-in of TSMC production capacity by nvidia/apple/amd/etc for their chips on their latest and greatest silicon process (aka the best in the world). There is 'no space' for performant large RISC-V implementations or other alternative (and it will require several iterations and mistakes will be made)
I know I can already replace my rpi3 with a linux supported out of the box RISC-V SOC board (aka, the enabling of assembly written software = no planned obsolescence from computer languages anymore, near 0-SDK).
Maybe I missed something, but if Google, Microsoft, Apple, OpenAI, etc want more capex on fabs, they can front that money themselves?
It's going to start a wafer war in my opinion. Best way to secure wafer supply is to pre-pay.
The second reason is the opposite though. The industry is cyclical. If you invest a tonne now, and that investment comes to fruition just as you hit the bottom of a cycle you're just going to go bankrupt. You'll have a tonne of capacity, no demand and businesses are designed to go bankrupt in exactly that way. You only need a small surplus of demand to skyrocket prices, and you only need a small deficit of demand to turn into a fire sale.
Let's put it this way - Ben points at a graph and says "Look, here's where ChatGPT got released - and the YoY CapEx went down!". Yes, that is exactly the cyclical nature. It takes a long term to sort the logistics for this industry, so if you see demand in early 2023, that's going to turn into CapEx in 2025 - and it did, a 35% increase in CapEx YoY is yuuuge.
The problem is that if you say competition is the answer what you're really saying is you want one player in the industry to chase a pro-cyclical demand cycle and that will blow them up. The result of this strategy is just to sacrifice the leading players each cycle.
The logical answer is actually, if Google wants TSMC to push CapEx into a bubble, Google should under-write it. Judge where you expect demand to be and hand over cash to buy the chips now. TSMC will build it. If you guarantee revenue in 2029, TSMC will absolutely guarantee you capacity. But Google doesn't beleive it, and they certainly won't pay up front for it. No, what you'll get is special purpose vehicles that look a lot like the big tech companies guaranteeing projects - in a way that doesn't show up on their balance sheets now - but turn out not to be guarantees if it turns sideways. And TSMC can't afford that counter party risk.
This is all a smoke screen. He knows very well that China can and will develop their own hardware to train AI models (and in fact, they are successfully doing just that; e.g. the recently released GLM-Image was trained on their own silicon). His only objective here is to slow them down enough so that they don't eat Anthropic/Claude's lunch releasing open-weight models that are increasingly competitive. But he can't just openly say "hey, we don't like that they release open weight models for free", so he's engaging in the AI version of the "think of the children" argument.
Anthropic's whole modus operandi was always pretty much "we should control this technology, no one else". It's not a coincidence they're the only major lab which has not released any open weight models, they don't publish any useful research (for training models) and they actively lobby the lawmakers to restrict people's access to open weight models. It's incredibly ironic that Dario is worried about (I quote) "1984 scenarios" while that's exactly what his company is aiming towards (e.g. giving Palantir access to those models is not "unsafe", but an average Joe having unrestricted local access is an immediate 1984-style dystopia).
Basically the current stack is primitive, so we waste 80% of the juice and then worry about how to power our inefficient mess. That’s the wrong set of questions.
No one will listen to me until I build it myself because everyone thinks they know everything.
If it also diverts workloads away from LLMs that we all know really belong elsewhere, more’s the better.
It seems wholly illogical that Apple would get refused wafer volume by TSMC and still refuse to give volume to intel foundry services. When you layer on geopolitical factors and national security implications + the fact that Apple is a US company - what reason could they possibly have to turn the shoulder to intel's foundries?
If Taiwan ends up imploding in any of the numerous ways we are aware of today - and which this article adds to - I think there are exactly zero reasons to feel like this couldn't have been avoided.
There are risks with TSMC, but "TSMC just decides it's not interested in making chips for other people, and cancels the whole business" isn't one of them. The same cannot be said for Intel.
Intel doesn't have any spare capacity.
A nice comparison table linked below. The best comparison imho is transistor density which is 313MTr/mm2 for TSMC's latest in-production process vs 238 for intel (higher is better).
https://en.wikipedia.org/wiki/2_nm_process
Also make sure to read the drama around https://en.wikipedia.org/wiki/Arrow_Lake_(microprocessor). Intel literally gave up making their own CPU on their own 20A process and instead utilised TSMCs fab.
intel's 18A nodes are significantly ahead of what TSMC is offering.
Further more, Intel's 18A yields are not where Intel needs them to be to be competitive.
Arguably false. Why do you think the US has encouraged TSMC foundries, now inside Arizona ? It's obviously to protect against the scenario that China takes Taiwan. In that case, give it 6 months or less for US TSMC foundries to produce the finest. China taking Taiwan will likely not result in the CCP getting any technology, certainly Taiwanese have "contingency plans" to vaporize all tech in the event they are invaded.
> Why do you think the US has encouraged TSMC foundries, now inside Arizona ? It's obviously to protect against the scenario that China takes Taiwan. In that case, give it 6 months or less for US TSMC foundries to produce the finest.
American business culture works pretty strongly against "give it 6 months or less for US TSMC foundries to produce the finest."
I think it's most likely those fabs will stagnate and the American MBAs running them will just milk them for short-term profits. Why invest in technology when you can buy back shares? After all, your only goal is number-go-up.
> China taking Taiwan will likely not result in the CCP getting any technology, certainly Taiwanese have "contingency plans" to vaporize all tech in the event they are invaded.
Not necessarily. Technology isn't so much the machines, it's the know-how. TMSC employees will still need jobs, post invasion, and I'm sure China will pay them very well. Some fraction will go to work for Chinese fabs, and teach them TMSC's tricks and knowledge.
>I think it's most likely those fabs will stagnate and the American MBAs running them will just milk them for short-term profits. Why invest in technology when you can buy back shares? After all, your only goal is number-go-up.
The same American business culture that produced Apple/Amazon/Alphabet/Microsoft/Meta/Tesla/etc? They seem to invest quite a bit in "technology".
Yeah. There can be sparks of innovation, but the overall trend seems to be squandering advantages for short term financial gains. And honestly, some of your examples aren't great: Telsa's probably going to lose to the likes of BYD (but for unusual leadership reasons); Microsoft seems to be losing former capability, which is reflected in many of its products (https://www.windowscentral.com/microsoft/windows-11/windows-...); and I don't think Meta has done much besides sell ads.
IIRC, Chinese companies, on the other hand, tend not actually make very much money for their investors.
You can even look to pharmaceutical companies for American business that have development timelines measured in the decades and investments measured in the billions. These aren't plays that executives profit from in 6 months, or maybe even 60 months. Real estate development is another business where investment timelines are in the decades.
I don't know enough about Chinese businesses, but I assume I cannot compare the motivations of Chinese business leaders to American business leaders, as Chinese business leaders seem less free than American businesses to capture profits for their investors.
One of the reasons Intel failed and TSMC succeeded is because Intel was unwilling to pay what the other big tech companies were paying (and part of that pay is with RSUs which stock buybacks help offset). However, for workers in Taiwan, TSMC was the best option, so TSMC could pay less.
A war would not wipe out chip production, but the squeeze would be immense for many years.
It probably wouldn't but it definitely could.
Also TW politicians have objected to notion they'd vaporize their fabs / golden goose. That meme started by US Army War College + Colby who said US should blow up the fabs, as in it shouldn't be a TW decision. Which TW have rebuked said they will defend against US attacks. Also other shenanigans like when US suggest they would paperclip TW semi engineers, and TW basically said there's no way they'd send semi engineers to safety before children. AKA TW not retarded, they know not to toast their golden goose, because golden goose for PRC still gives TW leverage even if TW forced to capitulate. They'd still rather be wealthy semi producers than pine apple farmers under PRC.
It's really a blind belief in american exceptionalism that makes you think this is even possible.
No, the chip factory that has had dozens of years of experience and local talent scaling up to make the most complicated products in human existence doesn't magically get up to par in 6 months. At best in 6 months they've figured out how to be less sensitive to vibrations and reach a low yield. The US doesn't have the trained workforce for this job, nor the infrastructures _around_ the fab (specialized hardware, electronics and engineering schools, various bits and bobs).
US TSMC doesn't get properly running in less than 5 years, and even that would be a miracle. You're also assuming that US TSMC has the current N2P or even N3E processes, and that agent orange doesn't burn bridges with europe hard enough that ASML stops selling to anyone related to the US.
> if we got to a situation where only the U.S. had the sort of AI that would give us an unassailable advantage militarily, then the optimal strategy for China would change to taking TSMC off of the board.
Lmao it's not. The author doesn't know what they're talking about at all. Let's be realistic: the current TSMC technology will be accessible to China, likely via espionage. The question is just how soon. It has already happened before. China's 7nm process was developed with the help from one of the highest level ex-TSMC researcher[0].
Realistically, to train a frontier model you’d need quite a lot of compute. GPT4, which is old news, was supposedly trained on 25,000 A100s.
There’s just no reasonable way of catching modern hardware with old hardware+time/electricity.
Also keep in mind we aren't talking about a small company wanting to do competitive R&D on a frontier model. We're talking about a world superpower that operates nuclear reactors and built something the size of the three gorges dam deciding that a thing is strategically necessary. If they were willing to spend the money I am absolutely certain that they could pull it off.
My hypothetical is probably somewhat over the top given that isn't China somewhere in the vicinity of 7 nm at present?
Taiwan might go scorched Earth and destroy them but that sounds more like a threat to foster US' support.
For the US the threat is either destruction of the fabs or China leverage against them if they get to control the fabs.
The USA has an act for it: Trading with the enemy etc etc
For different reasons you can't also just hire Iranian citizens straight into Los Alamos or wherever you are making plutonium nowadays.
National security > profits
2028 is not in six months, but GP is not an order of magnitude off either.
He didn't suggest anything like that, did he?
"certainly Taiwanese have "contingency plans" to vaporize all tech in the event they are invaded"
Given the way things are going, a rationalist would surmise that Taiwan is likely in talks with China for a peaceful reunification, Hong Kong style. The old way is very much over, the US is a worthless if not negative-value ally, and it's pretty clear to every living human with a functioning brain that this is going to be China's century.
Indeed, the article casually says "Taiwan is claimed by China, which has not and will not take reunification-by-force off of the table", which is technically true it isn't contextually informative. For those not fully up to date on the history of this conflict, for decades Taiwan claimed all of China (and most Western countries treated Taiwan as the singular government of all of China), and held out for reunification-by-force. This isn't as simple and straightforward like Trump's "we have a military ergo we get to steal better countries because they make us look bad".
I strongly suspect some sort of fab built by Elon associated companies will be announced soon. Almost all supply can be bought by Tesla and xAI.
It makes sense, IF he can get the tech to work at the bleeding edge. But he seems to be quite good at this.
I strongly suspect some sort of fab built by Elon associated companies will be announced soon. Almost all supply can be bought by Tesla and xAI.
Right now, it costs TSMC roughly $25-$40b for each N2 fab. It's going to cost $50b - $80b for an A14 fab 4 years from now. And then $100b - $160b again.
You need many customers in order to justify the cost. Tesla and xAI themselves won't be enough.
Frankly it is hopeless, I would be dumbfounded if Elon ever walked through the doors of ASML.
> Becoming a meaningful customer of Samsung or Intel is very risky: it takes years to get a chip working on a new process, which hardly seems worth it if that process might not be as good, and if the company offering the process definitely isn’t as customer service-centric as TSMC.
TSMC is a reliable supplier and there are no doubts about conflicts of interest. The same cannot be said for Intel and Samsung. If Intel's AI chip business faces chip shortages (like what may already be happening), can their foundry be depended on to ship your chips?
No one wants to be the idiot who staked their future on Intel and then gets wiped out when Intel doesn't deliver.
Ahem, ASML, who makes the manufacturing equipment for TSMC is a Dutch company, not a US one.
US origins of the key technology here give the US a veto here regardless of ASML's Dutch headquarters.
Is this an actual thing? Genuinely asking. Not that rules and laws seem to matter anymore to the current US administration anyway
That’s not guaranteed at this point, but it’s also not off the table since all the old rules are being shredded as fast as possible.
I'm glad the article dismissed this as a threat because it isn't one. The official policy of the US is the One China policy. You'll see this described as "strategic ambiguity". That's another way of saying that the official policy is simply to lie about supporting Taiwan's independence.
China can only hurt their position by taking military action against Taiwan. Also, it's highly debatable if they even have the military capability to invade Taiwan. Naval blockade? Sure. But to what end?
China is going to make their own chips. They'll just hire the right people to replicate EUV lithography. The article brought up nuclear weapons. It's a good analogy. At the end of WW2 the thinking of the US military was that the USSR would take 20+ years to get the bomb if they ever got it. It took 4 years. The gap with the hydrogen bomb was even less.
Western chauvinism in policy circles completely underestimates China's capacity to catch up in lithography. Not selling the best chips to China created a captive market for Chinese chipmakers.
I also think TSMC is being understandably cautious in not expanding their CapEx. AI companies really should focus on an economic use case for AI more than worry that foundry capacity will somehow limit a theoretical future AI use case.
"Putin is in the wrong here but there are no good guys. US rhetoric on this has predicted a full-scale invasion that hasn't come to fruition multiple times and the media just laps it up. It's reminiscent of the WMD justification for invading Iraq. It's straight up Manufacturing Consent [1]. However, Putin has a point: extending NATO membership to Ukraine is an overtly hostile act by the US and NATO member states. Putin no more wants NATO bases in the Ukraine than the US would want Chinese or Russian military bases in Canada or Mexico.
But Russia is not and never was going to launch a full-scale invasion of Ukraine. It would destroy Russia. Trying to do this in Afghanistan, a substantially smaller and less developed country, played a significant factor in the collapse of the Soviet Union.
Russia wants a buffer between it and NATO and access to the Black Sea. That's it."
it doesn't sound that Xi is as dumb as Putin but who knows.
Taiwan has missiles with the range and warheads to strike the three gorges dam.
An attack by China would end very poorly for everybody. There are millions of people living in the inundation zone.
So unless Taiwan has a method to deliver something the size of a bunker buster to the underwater base of the upstream side of the dam I don't think it's going to happen. And if they did manage to pull it off they'd presumably be condemned as war criminals more or less universally.
China obviously has a lot more resources than Taiwan, but then you have a concentration effect where an attacker can focus their resources on a single target, but a much more resourced defender can't necessarily afford to defend that target. We saw that play out with the UK's nuclear deterrent strategy in the cold war, where they focused on overwhelming Moscow's defenses, and were (probably) able to do it despite the USSR being so much bigger.