Apple mobile processors are now made in America by TSMC
(timculpan.substack.com)1592 points by colinprince 2 days ago
1592 points by colinprince 2 days ago
Morris Chang, the founder of TSMC was reported [1] by Nikkei Asia in March 2023 as saying this about the work culture:
> "Design is the U.S.'s competitiveness. On the other hand, Taiwan, Japan and South Korea have competitiveness in manufacturing...It's also about the work culture and the people."
> The TSMC founder cited chip production equipment as an example. Because these machines are so expensive, they need to be running 24 hours a day to justify their cost. "If it breaks down at 1 in the morning, in the U.S. it will be fixed in the next morning, but in Taiwan, it will be fixed at 2 a.m."
> "If an engineer [in Taiwan] gets a call when he is asleep, he will wake up and start dressing. His wife will ask: 'What's the matter?' He would say: 'I need to go to the factory.' The wife will go back to sleep without saying another word," Chang said. "This is the work culture."
This isn't all so unusual if its written into the job description. SREs in tech companies are expected to respond within a few minutes if they're paged in the middle of the night. They are usually compensated for their oncall time, however.
Expecting a worker to come to the factory out of fear or good will is not the way. Just write it into the contract/expectations/evaluations.
Out of the major players, Google is the only one I know of that compensates SREs for oncall time (and they do so fairly generously).
As a person who runs manufacturing in the US, this is our work culture too. Also the same at the other 3 previous places I have been at. The company culture asks for something, if it is not provided, you find someone who wants to be a part of that type of culture. There is an expectation to pay for it though.
Having to wake at 2 am on call is just bad for cardiovascular health, it's really just paying for one's life at retirement age and there's no real salary that can level that. Young people have an invisibility bias in psychology, they underestimate the physical toll of late nights and workplace stressors, which is cumulative over time.
As far as I know, they do. I think the bigger problem of the US manufacturing industry is that the most talented and motivated people have gravitated towards Wall Street and the "ads" companies. They not only pay significantly higher (due to cost/revenue structure) but also have a comfy working environment compared to factories.
The west has the same work culture when the industry and the pay demand it. The difference is that it may well be the woman who tells the man she's on the way to the factory. Or the wife who tells the wife. Thank goodness for liberalism.
I am not sure framing it as work ethic is right. It is simply the cost of labor. Some people might argue American's are more or less productive the hours they are working, which means just because someone from Taiwan is willing to be oncall 24/7, doesn't mean you'll have to hire exactly 3 American workers at 8 hours each to match productivity. You might need 5 because Americans truly are that lazy, or you might only need 2 because the on-call isn't that demanding since Americans are more productive.
Not saying any of these specifics are true, but framing it as work-ethic is not accurately capturing why it is more expensive to run factories in one country vs another.
To be clear, you are saying work ethic does exist and it is a factor but it's not the only factor, with productivity being another one?
I was commenting on a story about how americans had problems adapting to taiwanese work culture, and saying how they could benefit from this cultural exchange to optimize their productivity. Like you, I also do not thing this really captures why it's more expensive to run factories here, because I've never even suggested this.
Many of those articles came out before TSMC received CHIPS Act grants. As soon as the CHIPS Act money was committed to TSMC, the factory was suddenly ahead of schedule. Noah Smith called it out here:
> Three months after TSMC announced further delays at its $40 billion Arizona fabs, the chip manufacturer has now said the plant is expected to be operating at full capacity by the end of [2024].
> The announcement comes several weeks after it was first reported that TSMC is set to be awarded more than $5 billion in federal grants under the US CHIPS and Science Act…
https://www.noahpinion.blog/p/at-least-five-interesting-thin...
Some of the processors used in the iPhone 14 Pro, 14 Pro Max, 15 and 15 Plus are being made in America by TSMC.
Not even. The only device still in production using the A16 is the iPhone 15 (and plus if you consider that a different model).
It seems likely the new iPhone SE will be released in the next 12 months, and if so, and it follows past patterns, it'll roughly use the iPhone 14 hardware and thus, the A16.
The SE has always used the latest chip.
No chance in hell Apple releases any new phones — even SEs — that can’t do Apple Intelligence.
Rumors say an update is in the works; my guess is a release early next year. https://www.macrumors.com/roundup/iphone-se/
Yeah why are these chips still produced at all? The iPhone 16 just came out and the 14/15 stockpiles will be sold off for cheaper just to get rid of them. What am I missing?
The iPhone is not the only product Apple makes.
The Apple Watch, TV, iPad, Studio Display, etc. all use variants of older A series SOCs.
I totally forgot about the Apple TV devices and the screens. Yeah they also need SOCs .
Hats off to TSMC. Spinning up a new factory with processes this complex is very difficult, as anyone with manufacturing experience can confirm.
My understanding is it isn't a new factory. Wasn't the equipment moved from another operational factory so they could get up and running quicker?
Also, doesn't chip manufacturing require a lot of water? Water is not the first thing that comes to mind when I hear Arizona. I think I'm about to learn a lot with this.
Yes, ~10 million gallons per day (equivalent to 33,000 households). But the plant's water recycling and re-use is very efficient, so it's mostly a one-time hit up front.
From what I read the overriding factor was geological stability. Apparently these factories are very sensitive to vibrations. I guess when you do precision work at nanometer scale these things matter.
Arizona isn't water rich, but it manages to keep the 4 million people around Phoenix hydrated, so there is water.
> From what I read the overriding factor was geological stability.
Guess they are tired of dealing with all of Taiwan's earthquakes.
I mean, yeah, they have become an agricultural region... but it's not good for the people who live in Arizona right now. [0]
[0]: https://www.cnn.com/2022/11/05/us/arizona-water-foreign-owne...
Elon Musk did it with his Giga Factory somewhere in the range of 11 to 14 months in Texas. It’s 5 times the size of the pentagon and they build more than just top of the line Tesla Chips there.
Some states have less regulations than others. He said if he tried that in California, it would take an unlimited and indefinite time to complete (100 years+) due to so many permits and regulations.
to be honest, this is far better news than I was expected, and sooner, too.
is anyone else besides Intel making ~4nm* node wafers on US soil?
*yes I know I know I know about the misnomer with using nm measurements nowadays
I thought Intel 4nm was outsourced to TSMC. Or it's a rebranding of an earlier node. Am I mistaken? Do they actually produce that?
Here we are:
https://www.guru3d.com/story/intel-to-strategically-use-tsmc....
One of the chiplets of Intel's Meteor Lake laptop processors launched at the end of last year is made on "Intel 4"; the rest of the chiplets are TSMC N5 and N6. It was not a meaningful improvement over Intel's preceding generation that was made on "Intel 7" aka. the iteration of 10nm where the process was finally good enough for their whole product line.
Intel's Lunar Lake low-power laptop processors shipping in a week will be the first all-TSMC x86 processor from Intel. Their desktop/high-power laptop processors (Arrow Lake) will also be all-TSMC, and should be launching this fall. After that, Intel intends to resume using their own fabs for consumer processors with their 18A process. There are some datacenter processors using "Intel 3" and the 20A process was cancelled in favor of the more fully-featured 18A.
(In case of nitpicks: Intel is also manufacturing the silicon interposers that the chiplets are mounted on, but since these dies are completely passive and have no transistors, I'm not giving them credit.)
With Intel 4, Intel has not succeeded to obtain clock frequencies as high as with Intel 7, which is why the older Raptor Lake laptop CPUs still beat the Meteor Lake CPUs in single-threaded benchmarks.
Moreover, the new Intel 4 process had low fabrication yields, so Intel has produced less Meteor Lake CPUs than it could have sold.
Nevertheless, the Intel 4 process has demonstrated a much greater energy efficiency than the previous Intel 7 process, which is why the Meteor Lake CPUs beat easily the older Intel CPUs in multithreaded benchmarks, where the CPU performance is limited by the power consumption.
Intel 4 was the initial internal-only version of that process. The version available to external fab customers is Intel 3, which is in production in Oregon and Ireland. See https://www.trendforce.com/news/2024/06/20/news-intel-claims...
I mean, there are really only three bleeding-edge foundries: TSMC, Samsung, and maybe Intel if they've gotten their yields back on track.
Samsung has a fab near Austin, TX that was slated to make 4nm but it's been postponed to 2026 along with a shift to 2nm: https://www.tomshardware.com/tech-industry/samsungs-yield-is...
But their yields on 2nm are apparently... not great, so even that's in question. https://www.businesskorea.co.kr/news/articleView.html?idxno=...
SMIC is apparently making low-yield 7nm and is supposedly working on even lower-yield 5nm, but absolutely not in the US.
Yeah I was surprised to hear this is already at a point where they can produce chips. From what I've heard it takes a really serious amount of effort and expertise to calibrate the machines, and get the water filtration and other chemistry working in a new location.
Genuine question: what upside does it have against supply chain attacks?
Is it possible that an adversary to implement a backdoor into a chip design, without Apple noticing it?
I'm not a chip designer so perhaps the answer is obvious to some of you guys, but I'd expect some verification mechanism at Apple's side of the manufactured chips to match their original design to verify that they aren't tampered with?
Very much possible. Talking more generally about microelectronics - You can imagine the DoD is very interested in making sure they're not putting 'bad' chips in their military hardware, whether 'bad' means backdoored or merely counterfeit.
Manufacturing chips in the US means the DoD can investigate the acutal fabs and put cleared personnel on the manufacturing line to make sure nothing untoward is going on. Another strategy is to investigate the chip after it's been manufactured somewhere else and prove that it's the same chip you designed, but that's quite difficult.
If you're interested you can read up on the Trusted & Assured Microelectronics (T&AM) program.
Zero upside, probably a downside.
Apple has a top notch logistics and security processes which had mitigated the issue of supply chain attack in China which his willing and capable of producing such attacks.
Moving some production to the USA might induce some sloppiness in this due a perceived inferior risk.
Also, some security measures requested by Apple to manufacturers in other countries are probably illegal in the USA.
The article doesn't mention supply chain attacks. What context are you referring to?
I'll take a guess and agree with TOMDM. It's about China invading or blockading Taiwan (remember the US attack/blockade against Cuba? Exactly that.) and thus denying America physical chip shipments.
Isn’t Taiwan’s success in creating a TSMC correlated to the pyramid of their workforce which supplied skill at every level commensurate for high tech manufacturing’s demands? They have a high number of post-grads in their population AND also a large number of what we in the US call vocational/technical-educated working class. How are we doing as a country over time by the same measure?
At what cost meme, but literally, what's the cost of a chip made in TSMC US vs TSMC TW.
It's hard to not cost more than employees who think it's literally a patriotic duty to sacrifice yourself for a profitable company. American execs Dream of this kind of non-monetary influence on the work force
Presumably competitive otherwise Apple wouldn't be buying them?
IIRC morris chang indicated US operated fabs would cost ~50% higher, which is not cost competitive, well not something buyers would sign for without something happening behind the scenes (i.e. US gov pushes Nvidia and Apple to use Intel foundries). If A16 is $100 from TSMC TW, it's $150 from TMSC US, presumably $50 to BOM is something Apple can afford, but most others might not. List of companies who are willing to source at 50% limited (unless incentives).
>shipping, importing fees
Don't these chips still need to be sent to China for assembly by Foxconn? If anything this will increase costs even more and seems like import fees could potentially be even higher due to the current US-China trade war. Unless there is a plan to assemble everything in the US/Mexico as well. But then the costs would be way higher to assemble outside of China: certainly in US but probably also in Mexico, and I don't think they have any factories there.
Nah
$100/$150 would be the "shelf price" of the Apple chips if they were in a box for sale like an Intel/AMD one
I believe you that the cost of the delivered (roll of) plastic chips is 50% bigger in the US. Probably less but it might be (also need to include the logistic cost to send it back to assembly on iPhone, etc)
Apple (pre-)pays for stuff and probably doesn't have any orders where Qty is under 7 digits with these big vendors.
This is good news for the US and bad news for Taiwan, geopolitically.
The idea that the US only cares about Taiwan because of chips is popular on HN but just dead wrong. Taiwan has been part of the China containment strategy before TSMC was founded.
I think it cares about Taiwan as a democratic country but I think the chip fabs are becoming a geo-political factor as much as oil fields or other resources.
i.e. we don't want [Russia/China/Whoever] to invade Country X as Country X is an ally and a democracy, but as Country X has [Oil fields/Chip Fab/Lithium Mine] we REALLY don't want them to invade.
The difference between an oil field and a chip fab is that the equipment is more easily destroyed in a chip fab, vs a hole for an oil well. Not to mention that expertise in human capital required for chip fab is way higher than that of an oil field.
Even a successful invasion of taiwan guarantees either the people important to the fab will leave, and the equipment evacuated, or destroyed if unable to evacuate.
Well, yeah… The US doesn’t really have allies (the one exception some would say is Israel) - why would Taiwan be any different from the others? Interestingly, TSMC only became a stand-out player in the last 15 years, before then there were basically zero reasons for the US to care about Taiwan except to contain China. Now they have one reason apart from containing China, but it’s still mostly just about China.
Depends. China's obsession with Taiwan is a mix of domestic signaling and posturing internationally and the latter is mostly aimed at the US. China could choose to be more aggressive over Taiwan, as the US should care less. But, since the US care less about Taiwan, perhaps China will turn its saber rattling to other strategic interests of the US, giving the Taiwanese some reprieve.
You're overestimating the importance of that specific chip.
No, TSMC said that the cutting edge fabs will always be in Taiwan
This isnt Nvidia and this isnt some high end CPU.
This is a mobile phone CPU, and its Apple. You are getting insignificant technology.
Not so. Apple's new mobile processors are routinely the fastest processors in the world (single threaded) when they come out. The A17 pro is currently the 17th fastest CPU, and the M3 (which is in MacBook airs and iPads) is number 2.
Sure these don't have the scope or number of transistors of like an NVIDIA Blackwell or something but in terms of performance/watt these are ultra high-end ICs.
No one is competing on CPU though. Its like having the highest RPM lawnmower, no one cares, its not useful.
To make it worse, they arent even the best. Its getting mid tier, and in 2024, its nearly unreasonable to buy mid-tier when low-tier is good enough for everything.
Not sure about that, the consumer CPU market is probably more competitive right now than it's been in a decade, primarily on efficiency.
Also not sure what that has to do with the original point, which is that the A15 is not an impressive chip to be manufacturing in the US because it's designed by apple and meant for mobile devices, neither of which are reasons to discount the complexity of the chip.
There's a lot of chip packaging in Taiwan, Malaysia, and maybe Singapore so these A16s are probably racking up frequent flier miles. Probably not China though.
In the future they will probably be packaged in the US: https://www.apple.com/newsroom/2023/11/apple-announces-expan...
Most likely frequent boating miles given their push to use water transport over air when possible for the environmental benefits.
I think "for assembly" here means iPhone assembly, ie. the final SoC will be sent to China to assemble the iPhone. I don't think GP is referring to packaging.
Please follow the site guidelines when commenting here: https://news.ycombinator.com/newsguidelines.html.
We've had to ask you this before (https://news.ycombinator.com/item?id=35998957), so if you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
Two points to counter the snark:
1. The output of a "chip manufacturing" process is a wafer. There is absolutely further assembly (bonding, packaging) done on this output.
2. The chips themselves are not for the end user's consumption. They are assembled into a product, a "consumer electronic".
The dies themselves are "assembled" - cut from the wafer, bonded to the wires (or solder bumps) that carry signals to the rest of the system, and packaged for physical protection and thermal management.
In recent times, multi-chiplet architecture has added its own layer of complexity to that process.
See also: OSAT.
This should go a long way to ensuring our national security does not suffer. We don’t need TSMC level volume production, plenty of non Taiwan entities exist to balance the risk.
We do need latest edge tech to be within our borders and TSMC and Samsung will deliver that in 2-3 years.
Any de novo chip plant operation in the US seems like a big win. Machinery can change any day. Operational workforce is significant.
The article says that the A16 processors being produced here are using the N4P process, and are referred to as both 5nm and 4nm, confusingly. But they are used in the iPhone 15 and 15 Plus which are still available.
Looking back on some of the comments from this thread^, especially those claiming that TSMC building a factory in the US was infeasible/impossible, was entertaining in light of the current thread.
With fairness to those commentators, Taiwan will still have a process lead on US fabs until 2030 at least. It's entirely feasible that an American TSMC fab will get beaten on both density and price by Intel and Samsung in the immediate future - these fabs won't be manufacturing flagship nodes.
A little-known fact is that Donna Dubinsky, the CEO of Palm/Handspring in its glory days, is on the staff of the Dept. of Commerce specifically to help with the CHIPS Act.
Or at least was:
https://www.commerce.gov/news/press-releases/2022/09/biden-h...
I don't know what role, if any, she had with TSMC.
Wow that was fast, is this a regular timeframe to get a new fab working?
From the conversations about China catching up on smaller chips I got the impression that it takes loads of iterations around how to calibrate the machines but it seems TSMC nailed it not only on Taiwan but also overseas very fast.
This happening at the same time as germanys intel project freezes makes the fall seem even harder
https://www.ft.com/content/3fa44901-33e4-4ab4-9f7b-efe1575a6... & https://archive.ph/FDmwq
> US and Japan are close to a deal to curb tech exports to China’s chip industry.. export controls are designed to close loopholes in existing rules.. make it harder for China to obtain critical chipmaking tools — restrictions that would have the biggest impact on ASML in the Netherlands and Tokyo Electron in Japan.. to restrict servicing, including software updates, and maintenance of the tools..
Unless this is election propaganda, which very well it might be, this is huge news. I know there were a lot of problems for this facility and wasn't aware they were this far advanced in production.
Tsmc will never allow the Arizona plant to be a viable replacement. They are extremely incentived to prevent this happening.
That's OK. It's on US soil with US employees and can be nationalized if and when need be. I'm sure ASML will be happy to comply or else risk their US operations being nationalized too. Like their DUV/EUV light sources office https://www.asml.com/en/company/about-asml/locations/san-die...
How so? They also are extremely incentivised to make this happen. A war on your front door is not good for business.
You don't have to be a rabid nationalist to not wish for your country to be invaded and annexed by others. You don't even have to live there. I'm sure a large percentage of Taiwanese living in countries outside of Taiwan would not wish for it to be invaded.
I'm not even Taiwanese, don't know anyone of Taiwanese descent well, and I don't want Taiwan invaded.
The suggestion that there's some kind of weird oligarchy class of TSMC-controlling Taiwanese who couldn't give a toss if Taiwan was invaded is a mustache-twirling level of caricature.
TSMC is governed by the Taiwanese government, which is a puppet government controlled by the US government and military. TSMC answers to the US directly, as without US support, Taiwan falls to China almost instantly. Nobody besides the US can prevent a blockade of Taiwan
A "puppet government".
This claim is based on... Them wanting not to be invaded?
If anything they had the foresight and took advantage of US companies not wanting to fab their chips at higher prices domestically. This has led to cooperation between the two.
I see this argument in some fashion every day, claiming US allies are puppets. When in fact, they just find commonalities and cooperate.
the strength of the US defense commitment is likely proportional to the strategic value of the economic assets they still hold. the taiwanese have every incentive to do just well enough at the AZ plant for the $39 Billion checks to clear and no better
While true, TSMC has a stronger incentive for its own survival than the survival of Taiwan. If it's easier for them to shift operations to the US and continue to make $$$, I suspect they'd do that over retaining operations in Taiwan and hoping it will convince the US to protect the country.
This factory is not for economic independence or economic strategy. It is for geopolitical strategy. This factory is meant to build smarter munitions if war breaks out, not the latest cellphone. The US gov does not give a fuck about Apple's stock price and product plans if war breaks out with China, since, you know, there's real adult problems going on.
That's great to hear. I hope other countries, like Canada and the EU, also do this. I think it's important for all major nations to have this sort of critical capability in house.
Covid showed this well, despite being allies, countries tended to get vaccines to their own people first, even breaking agreements with allies. That's likely normal, and a bit of mutual distrust is healthy.
Knowing all the efforts that the US government has had to devote in order to push Apple to bring those jobs home, for other countries that do not have as much muscle as in financial and industrial leverage, their industrial future must look quite bleak.
Everyone seems to be celebrating this as a victory for the US, but I can’t help but think of David Ricardo’s Law of Comparative advantage. National security concerns aside for a second, what high tech sectors will the US necessarily be investing in less now that we are putting those valuable resources into chip-making? Are these sectors more or less valuable/profitable than chip-making? I don’t have an answer, but this is the framework that needs to be used to address the question. The US can’t do everything, especially with current immigration restrictions on high tech workers.
Less adtech and crypto? Fewer gamified dating apps?
I can think of a lot of negative/zero sum things that have next to no return or longer term advantage than monopoly seeking or greater foolism. They already got plenty of investment when interest rates were near zero.
If there hadn't already been a significant semiconductor industry, or if there was some similar employment for those employees/grads to go maybe it would be different. If there wasn't large local demand for the product (and I'm including the packaging which is another issue) it would be different. Given what the US has it makes long term sense to put some 4nm and even 2nm Fabs in the US. Creating geopolitical risk by outsourcing ALL supply is sort of silly, quarterly profits be damned. (even $50B is <0.2% of annual GDP).
We could probably find some ways to encourage the Wall Street types to go get real jobs also.
> Everyone seems to be celebrating this as a victory for the US, but I can’t help but think of David Ricardo’s Law of Comparative advantage
This theory has always been an overly simplistic model designed to promote the ideology of free trade. The most obvious problem with it is that it only works in a static world where everything stays the same and as such specializing makes sense. But the world isn't like that, and if everybody invests only in the places where they have a comparative advantage, then you have set up a trade network that is very vulnerable to asymmetric shock: if one good becomes irrelevant or too desired, then the system starts failing.
Germans are learning it the hard way now that ICE cars are getting out of fashion.
As always, there's a yield/resilience trade off, and at nation scale, favoring yield is a recipe for disaster.
>Germans are learning it the hard way now that ICE cars are getting out of fashion.
Sadly, "the hard way" is the only way Germany learns lessons. All that national pride on German ICEs is coming home to roost. I remember when I was working for a large German auto company a while back, a division manager laughed at a Chinese auto company in a presentation that "they have tradition since 1995 lol". The arrogance aged like milk.
It's not a nation that values proactive thinking and adapting to change but stubborn pride and conservativism.
Fun fact: If you read Ricardo you will find the modern form of "Comparative Advantage" isn't really there.
Taiwan doesn't have a "natural climate for chipmaking". In a modern industrial economy, endowments are not natural/fixed by the result of previous rounds of investment.
> what high tech sectors will the US necessarily be investing in less now that we are putting those valuable resources into chip-making?
There is no evidence it is actually zero-sum
> especially with current immigration restrictions on high tech workers.
Yes, more immigration would be greatly appreciated. Probably won't happen until we unfuck housing, however.
ASML for those who don't know is a Dutch company, and supplies the EUV machines for both TSMC and Intel (It is not clear to me if Samsung uses EUV in its current process nodes). I believe they are the only EUV supplier in the world. There are certainly other suppliers other that ASML, since there is a lot of other equipment other than lithography, but that's a critical one for modern process nodes.
ASML has important ops in US offices fwiw https://www.asml.com/en/company/about-asml/locations/san-die...
Wonder how apple was able to curb the (assumable) higher cogs of producing this domestically.
Its a mobile CPU chip, and its Apple. This isn't going to move any markets, its insignificant. Maybe its politically useful for Apple and the US government for PR purposes, but there is no rush for CPU chips, or Apple hardware.
ASML make fancy printers.
TSMC and other ASML customers build the designs that let those fancy printers create transistors and then logic gates, as well as a basic library of arrangements for those logic gates (PDK). They also provide all of the raw materials and processes and physics that go into said printers.
Apple and other design customers then compile RTL using that PDK to produce a design that can be manufactured using the fab’s process steps.
The printers are A hard part but far from The hard part. If you have an ASML machine it is useless to you unless you have also figured out how to build a 3D transistor in layers. Good luck!
Apple doesn't design on ASML equipment. Apple (and other fabless companies) designs to a PDK (process design kit, basically rules about how to layout transistors and passives on the die), which is given to them by their foundry (TSMC in this case).
There's a lot of steps between circuit design on the PDK to a working high volume process; and ASML machines are only part of that.
There are a lot of steps involved in making the chips - lithography is only one of them. You have to have the supply chains set up for massive amounts of silicon, you have to have a process for doping the silicon properly, you need quality control, you need to actually build a fab to house the lithography machines, I could go on.
ASML makes tools for only a small part of the semiconductor production process. It's true that EUV lithography is the big limiting factor right now, and that it is a field dominated by one manufacturer. So it's reasonable to credit ASML "as much as" TSMC for the current dominance of their high end nodes.
Nonetheless if it was as simple as buying ASML boxes there would be more than one fab at the top of the heap, and there isn't. TSMC absolutely "contributes" to their own dominance, arguing otherwise is silly.
A couple trillion dollar valuation?
You can start here: https://www.youtube.com/results?search_query=asianometry+tsm...
So this is the N4P node... From way back in 2021.
And these are 2 year old chips for a phone that is about to stop being sold...
Seems this news might be more political than strategic... The US still relies on Taiwan for every modern chip.
And then sent for packaging to Taiwan and assembly in China?!
Exactly what does “manufacturing in America” mean? It could be as little as final assembly with most of the work still being done in Taiwan. Like Cook said Mac Pros were “being made in America”.
Since US manufactured products are traditionally reputed to be low quality, should we expect to have to look for serial numbers to get iphones with non buggy A16 chips?
And GlobalFoundries (ex-AMD, ex-IBM). There's also less cutting edge process stuff at ONSemi, TI, Micron, Analog Devices, Diodes Inc and I'm sure I'm missing a few.
Even Apple has their own fab.
Build a workforce with the world's leading company. Pull those people into senior positions at new companies nearby. An industry is born and competition can grow. All with export limits so the jobs hopefully can't be outsourced.
Are the inventors of the wheel still in the driver's seat?
The past is another country.
Find me a modern US exec willing to actually invest in a risky hardware prospect, rather than throwing a billion dollars into real estate or "content" that can be filled with ads.
Curious to see when the US will force the TSMC to sell because it presents a national security threat.
The new Apple chips are second generation 3nm. This 5nm stuff is old tech. Why are people celebrating?
I wonder if the US plant ASML equipment also have a destruct mechanism like the Taiwanese plants have.
This is title seems to be quite the overstatement of the facts.
Hats off to TSMC. They had big culture clash and US has a lot of red tape and high labor costs. They did it!
Also kudos to CHIPS act.
I'd rather have Boeing and Intel wither off, for them to be replaced by new players who bring highly efficient manufacturing to the table.
All this because Donald Trump claims (contrary to nearly all economists) that forcing companies to manufacture products on US soil is beneficial in some way that he (Trump) feels confident will make America great again. It is so embarrassing that these outdated ideas are entertained for even a second by HN readers.
only because of the outsized political importance of a few states that happen to specialize in outdated manufacturing technologies and happen to have enough electoral votes that politicians have an incentive to subsidize them.
It's a massive tax on the economy all to provide a tiny bit of welfare to a small number of workers. Better to just pay them a welfare check!
The world is the most unstable it has been in decades. If/when a war kicks off, you have to have your supply chain local because the oceans will be instantly impassable until we can work out how to counter submarines.
This has little to do with welfare and everything to do with national security.
Empirical studies show that governments typically do not introduce policies that result in benefits overall, and the costs of those polices are typically higher than if everyone had just paid a tax that was given as welfare to the small number of workers in the effected industry.
What position is a homeowner in when they decide to hire someone else to mow their lawn? Economic specialization generally a good thing.
US politicians get enamored by industrial policy when they see what happened to the "asian tiger" economies over the past decades. They forget that those nations were so destroyed by war that the "growth" was less due to the policies than to the people's motivation to live in a free and peaceful society.
China is also now the poster child for industrial policy. China had many years of intentional economic suppression in the name of societal harmony (preventing chaos resulting from some regions being poor and isolated and others being rich). In recent years China has managed to use some of the wealth to undertake a social policy (plus industrial policy) of bringing wealth from the coastal manufacturing regions into the agricultural regions, training workers, etc.
Even in spite of all this, China's GDP is still significantly lower than it would be without all the policies, but the societal order is preserved and there is likely greater social stability.
China faces unique challenges in these areas relative to other countries (largely due to geography) which is why it had suppressed its economy so much for so long.
We are getting a glimpse at what a modern approach to Chinese capitalism will look like and it has already left the US in the dust in terms of productivity. It's ironic that the US mis-attributes the success to the industrial policy rather than to the repeal of it.
Well actually, Alexander Hamilton is the father of mercantilism. And it's been followed and promulgated by pretty much every country at some point in time.
It's not wrong, it's one strategy given the political goals of a nation. There are other strategies and other goals, like economic liberalism.
Saying a behavior or approach is wrong and/or outdated shows a particular misunderstanding of what policy is for.
What "most" economists believe in the West (and "believe in" is a perfect way to put it, because it's a belief) is economic liberalism. Underlying/embedded in that belief are a number of assumptions, policy goals, and desired outcomes.
For a limited subset of countries on earth that worldview has been incredibly successful. However, for the vast majority of countries on earth economic liberalism has been a failure, and a costly one.
Unfortunately, there aren't many new alternatives out there, and the current system is heavily biased towards economic liberalism.
But it's important to remember that all this is relatively new. The era of modern states is relatively new, and the current postwar order is well, 80 years old or so. The Wealth of Nations was only published in 1776, Report on Manufctures was in 1791, and Das Kapital was in 1867.
These are good points. In my view the idea that "making America great" entails illiberal economic policies which benefit a small fraction at the expense of the rest of the population is a non-starter because in my view "greatness" does not come from propping up outdated industries (coal extraction, steel production) and taxing everyone else to do it.
I don't think economists are ideologically opposed to central planning. There are simply enough empirical studies that show how badly it fails. In fact most of the "economic liberalization" failure stories you refer to are actually centrally planned thefts that benefit specific firms but were sold as liberalization.
China is an example of a state that does very smart central planning. Everything from its central bank to its subsidization of small businesses doing embedded systems (hence all the super cheap gear on Amazon sent via subsidized shipping to customers around the world) is intended to enhance the capability of the workforce and guide the workforce toward a future of technological change and rapid (but not too rapid) advancement.
In other words, China's industrial policy is forward-looking, America's is backward-looking. The very phrase "Make America Great Again" is backward-looking.
China's policy is essentially an education policy disguised as trade policy. Corporate espionage leads to more knowledge, subsidized shipping leads to more low-end consumer devices and engineers who need to learn to build them, etc. There thousands and thousands of low-end consumer electronics, test equipment, etc., manufactured in China that are built upon the many low-end DSP chips and microcontrollers. This is not an Apple-esque 2nm process, it's much lower tech, lower cost but it offers far, far better experience to so many more workers than all but the best educational background can offer. What percentage of first or second year US EE grads could build and ship a $50 spectrum analyzer?
In my view, China has already overtaken the US in key areas of technological innovation and the US is "copying" by deploying industrial policy that has the opposite effect and entrenches and protects top US firms while having minimal educational impact on US workers and minimal impact on educational and early career choices for US workers.
It's funny how (supposedly liberal) opponents of Trump will strategically whine about economic theory when generally otherwise if you were to make appeals on the basis of economics you're labelled heartless, etc.
You know what? I like when jobs are based in America because people need domestic careers that can sustain communities. There are non-monetary costs to outsourcing that are not mere quantities for an economist to decide for us.
Ross Perot was wildly successful running on a platform like this, at least relative to any other third party in American politics, and as soon as he appeared to be a threat to the establishment, strange stuff started happening to him much like the assassination attempts in this election.
Heartless? In my view it is inappropriate for the government to prohibit or tax peaceful, voluntary activity such as trade.
The US grew economically due to the interstate commerce clause prohibiting states from imposing tariffs on each other, and now we are supposed to believe that Trump and Perot are economic geniuses because they want to subsidize coal extraction and tax EVs so that Americans have to pay double?
Most of the big wars started because countries got protectionist and isolated and had no economic reason not to fight each other.
Yes it is heartless, because it's talking about complete dissolution of communities and swaths of the country due to corporate decisions that are all made under certain regulatory regimes and strategic policies of foreign nations (in many cases enemies). Look at H1B, I've never seen a good justification for it based on working in many large corporations, there's absolutely no reason we can't be finding actual citizens that could fill these jobs and build their careers.
From what I've seen most economists have extremely short-sighted thinking. Their theories are almost comically naive.
Yes, economically in the short term (< 50 yrs) putting your eggs in your few specialized industries will give you big economic growth. But, in the long term this is economically extremely risky - because you're relying on remaining competitive in those few, high price, more advanced industries. If that happens to change, you're screwed.
And it CAN change due to geopolitical factors (something economists don't understand). A dictatorship of the future can 100% make more efficient supply lines than you. Even somewhere in-between and you can be screwed - just look at the Chinese automobile industry.
For decades, the automobile industry has been the darling child of the US. This has, and will continue, to no longer be the case. The reality is China subsidizing their industry and providing top-down support means they can make better cars cheaper. The only reason this hasn't completely fucked that portion of our economy is because we don't let them in.
We can't keep outsourcing all our manufacturing while we sit on our asses and rely on our darling child industries to grow.
Take a look at what happened during the global communist revolutions. Those communist countries were scary to us because they have the potential to make more shit and make it cheaper. They can out manufacture us.
Luckily we were not completely braindead (and the tech did not exist) to outsource our manufacturing to them. But if we did, it could have been catastrophic for our economy in the long-term.
A few points to consider:
- Electric vehicles are inherently much cheaper and have way fewer moving parts. Just because an entry level internal combustion vehicle costs $25K doesn't mean an EV has to. But with 100% tariffs it can!
- Every day that American workers spend building heavily government subsidized internal combustion powered vehicles is a day we fall farther and farther behind. All those low-end "hoverboards" that everyone bought a few years ago, all the electric scooters. The engineers who design those in China are the ones designing low-cost EVs that utterly out-compete what the US can do. US policy to subsidize mediocrity (Tesla, over-priced, over-hyped, impossible to maintain) HARMS the entire US economy. How many people need to pay an extra $500 to $1000 a month in payments that are effectively a subsidy of outdated tech? Most people with a car payment are doing just that.
Meanwhile we keep getting into wars over petrol which is why we don't keep track of how much we spend on the military because nobody cares, of course it's worth it to keep the oil flowing!
Economics is about information. Price is a function of supply and demand. As much as governments may wish that internal combustion tech was competitive with low-cost EV tech, it's not. As much as everyone wishes healthcare was free, it's not. We have to choose our subsidies wisely. US industrial policy is a disaster and it is fraught with so many misconceptions.
If it's really a national security issue, where is the US stockpile of raw steel, copper, lithium, 555 timers, etc.? Politicians would rather rant and impose tariffs and get photo-ops near coal factories than actually do something simple and strategic that would take away the possibility that a conflict would disrupt crucial supply chain.
Economic specialization is a good thing. Economies are not so simple as importer and exporter. Most companies are both importers and exporters. China's government knows this and adopts sensible policies like subsidizing oceanic shipment of goods so that shipping costs of the $25 electronic device aren't $100. This lets an engineer build and sell something and learn and grow.
China has an economically-aware industrial policy, the US has a backward-looking, short-term, electorally driven one.
A strategic triumph for both the current and previous administrations. Both Trump and Biden handled the situation adroitly. These may not be the absolute bleeding edge tech but it’s a proof of concept that we can wean ourselves from Chinese tech if it becomes necessary.
It comes at the cost of many, many Chinese jobs in the midst of a devastating economic downturn there. They are victims of repeated local and geopolitical malpractice by the current emperor.
EDIT: User lotsofpulp pointed out that we don’t make any strategic chips in China. That is of course true. I meant that the game of economic chess played by the current and previous administrations has been highly effective in reducing China’s options.
>It comes at the cost of many, many Chinese jobs in the midst of a devastating economic downturn there
As far as I know, TSMC does not make chips in China.
TSMC Fab 10 is in Shanghai. Probably not cutting-edge stuff coming out of there though.
> Both Trump and Biden handled the situation adroitly.
I don't remember any news like this during the Trump administration. I do remember the Foxconn plant that didn't open though!
Biden continued Trump tariffs and raised many of them, appropriately.
https://taxfoundation.org/research/all/federal/trump-tariffs...
The yields are pretty good - TSMC’s Arizona Trials Put Plant Productivity on Par with Taiwan https://finance.yahoo.com/news/tsmc-arizona-trials-put-plant...
Apple seems to have pretty tight quality control. Thought it is true, they are likely starting with lower scale production of an older chip to work the bugs out of the system.
That the term "electoral college" does not appear once in this entire thread is telling.
Now it does as a result of this comment. Is that also telling?
I can't believe iPhone chips, almost the supreme luxury good, are considered worthy of Federal subsidies.
Surely a better path would have been to slap imported silicon with tarriffs at least equal to their gov't subsidies?
(Unpopular opinion: The people that spent the last 30 years giving away US & EU manufacturing to the Far East - no doubt with plenty of "10% for the big guy" type deals behind the scenes - should all be shot.)
This is troubling news, as we could soon be paying $2,500 for an iPhone within the next three years. The original reason for outsourcing was to keep costs down, and now, with this trade war, it's clear consumers will bear the burden.
While some may see the return of manufacturing to the U.S. as a win for national pride, the reality is more complex. The high cost of U.S. labor, combined with excessive bureaucracy, leads to higher production costs, which ultimately get passed on to consumers. There's nothing inherently beneficial about manufacturing in the U.S. other than symbolic gestures tied to identity politics.
Most consumers want affordable, high-quality products, not overpriced goods that may be touted as "Made in America" but offer no real value beyond that label. Instead of focusing on where products are made, the priority should be on ensuring that they are durable and not part of a system of planned obsolescence. We want iPhones that last longer, not cost more, yet U.S. manufacturing may drive up prices without offering real improvements in quality or longevity.
Unfortunately, the consumer is losing in this scenario—stuck paying for rising costs while receiving little in return. We need to reassess the real benefits of domestic manufacturing and whether they justify the inevitable price hikes. It’s clear that without a shift in strategy, we're moving toward a future where innovation is stifled by political posturing and unnecessary cost inflation.
Chip manufacturing is critical for national security, which is to say world security, if you like the Pax Americana (and you should). This is not some trade war thing. My only actual concern about this is that it may make the US less willing to intervene if the CCP invades Taiwan, and we absolutely should intervene if that happens.
I get the national security part, but not sure about world security part.
Why should I, as an example, who is neither a US citizen nor Taiwanese nor Chinese, should trust a chip being manufactured in the US vs. somewhere else?
I'd say it is neutral in regards to world security, not better.
I suppose it depends on where you are from and your politics, but I think many people outside the US would feel safer with chip production in the US than under Chinese control. I don't think most would really jump at the chance to buy the same chip from a US manufacturing plant vs. a Taiwanese one, but if China were to make a move on Taiwan, I'm not sure the world's computing resources would be particularly safe. (Not to mention, I wouldn't be surprised if the secret back-room plan was to raze Taiwan's chip manufacturing capability to the ground if it looked like China was going to win a takeover of the island.)
Even ignoring the specific players, having critical advanced technology manufactured in more than one place increases world security. What if, say, a catastrophic earthquake were to significantly damage Taiwan's chip manufacturing? Having expertise and working, active manufacturing elsewhere is a good thing.
I mean that the world is best off if the US continues to maintain the global maritime order, and this means there being no credible way of cutting off the US military from being able to mass produce weapons.
Sure, this costs US taxpayers a lot, but whatever, it's worth it.
> and now, with this trade war, it's clear consumers will bear the burden.
"Now"? The trade war has been on since what, 2017?
> There's nothing inherently beneficial about manufacturing in the U.S. other than symbolic gestures tied to identity politics.
While I think that argument can be made in general, if you consider certain sectors and certain products, the calculus changes. Onshoring chip production is a matter of national security. Not necessarily in the "big bad China will take over Taiwan and put backdoors in our chips" sense (though that's certainly a concern), but in the sense of not being dependent upon an adversarial state for fundamental advanced technology.
> Most consumers want affordable, high-quality products
Sure, but that's not sustainable. You end up playing "chase the country with the worst worker protections". This isn't the case of chips (yet?), but there are quite a few things where China used to be the go-to for manufacturing, but production has moved elsewhere because costs went up, and it's cheaper to stop doing it in China. The long-term end result of all this is that everywhere has labor costs that have gone up enough that offshoring doesn't really buy you all that much.
Of course you can say, "okay, maybe that's true, but at least I can get my cheap iPhone now, and moving production to the US hurts that now, rather than decades from now". And I'm somewhat sympathetic to that. But ultimately Apple may just have to change how it prices things if it costs more to make iPhones. They already make solid profit on each unit, and perhaps they'll just have to make do with less of a markup.
> Instead of focusing on where products are made, the priority should be on ensuring that they are durable and not part of a system of planned obsolescence
I feel like Apple is a pretty bad example for you to use here. I had to replace my perfectly-functional, four-year-old Pixel 4 last year because it stopped getting software updates after three and a half years. Meanwhile my wife has a six-year-old iPhone that will update to the latest major version of iOS tonight, and it will likely keep getting updates for a couple more years. My new Pixel 8 will supposedly get major OS updates for seven years. If I break the screen on my phone or the battery gets bad, I can get them replaced fairly affordably. These are improvements!
Apple's repair situation is worse, but that's a choice Apple has made. If they wanted to focus on repairability, next year's iPhone would be the most repairable phone on the market. But they don't want to do that. Moving manufacturing around is orthogonal to all that.
I feel like I've read a few articles on Bloomberg and/or NYT (drawing a blank on the exact source) that a very large portion of the workforce was taken directly from Taiwan and the American workers were having a hard time adopting to the Taiwanese way of doing things (long hours, on call all the time, constantly stepping outside your predefined roll, etc.). Is this currently now, or will it in the future, affect the overall success of the factory? (It also might simply be untrue for all I know.)