Apple mobile processors are now made in America by TSMC | Hacker News Reader

Apple mobile processors are now made in America by TSMC

timculpan.substack.com

1,612 pointscolinprince824 comments

After all the wailing and rending of clothes, the industrial policy worked out great and we have top tier production here in the US, transferring knowledge from TSMC to a US workforce.

This is a significant win for the US, and just the beginning of the amazing industrial policy passed over the past few years.

US manufacturing is about to be reinvigorated, and we in the US are going to be building our own future both for chips and for energy security.

This is great news, and we should celebrate.

> US manufacturing is about to be reinvigorated

One thing that I think that frequently eludes people in this discussion is that US manufacturing has pretty much been constantly growing for the past century; just its share of GDP has fallen as other sectors have grown faster. And the share of the workforce has fallen even faster, as the actual manufacturing has moved towards higher value items and greater degrees of automation.

I think the actual outcome of this policy is mixed. I think it was a big case of corporate welfare that will result in somewhat increased chip production in the US. I think this is a win for national security. I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

> One thing that I think that frequently eludes people in this discussion is that US manufacturing has pretty much been constantly growing for the past century; just its share of GDP has fallen as other sectors have grown faster. And the share of the workforce has fallen even faster, as the actual manufacturing has moved towards higher value items and greater degrees of automation.

The last part is missing something important though. If we're measuring "output" in dollars and the US is doing the parts (like aircraft manufacturing) that globally don't have a lot of competitors, the high "output" is from high prices rather than high production, and then what we're doing is surviving in the markets where there isn't a lot of competition and getting killed in the markets where there is.

There are three problems with this. The first is that it implies the US isn't competitive in competitive markets, which is a sign that something is very messed up. The second is that the markets where other countries aren't competitive tend to get eroded over time. The US essentially had a lock on the auto market in the mid-20th century; not anymore. What happens when China starts making globally competitive aircraft?

And the third is that supply chains matter. If you give up on the low margin stuff instead of figuring out how to make it competitively domestically (e.g. via automation) then foreign competitors have a leg up when it comes to making the high margin stuff for which the commodities are inputs.

> I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

Relying on "free markets" actually requires free markets. If other countries are willing to subsidize their industries until they drive manufacturing out of the US, that's not a free market. It's the equivalent of a monopolist using dumping and tying to leverage their existing monopoly into new markets, but with a country acting as the monopolist and therefore being exempt from antitrust enforcement.

Doing the same thing in return is not likely to be an efficient strategy, but neither is the status quo. The main alternative would be to realize that the thing we've been calling "free trade" is not actually that and a country that subsidizes its industries until its US competitors exit the market has to be dealt with as an abusive monopolist, e.g. via tariffs and similar policy levers, since antitrust laws don't apply to foreign governments.

Do you have a source for this? In the past when I've been told this, the statistics referenced where based on whether a company was classified as being in in the manufacturing sector, not based on which jobs were classified as manufacturing. This included companies that were classified that way due to historical inertia, or based on their global industry but actually had little to no manufacturing in the US. Based on that I have a hard time knowing what to believe, and would love to be pointed to more accurate information.

> US manufacturing has pretty much been constantly growing

By dollar value, perhaps, but that mostly means the US makes lot of high-value microchips, a field that has made (well-documented) exponential progress over the past decades. It is still consistent with US manufacturing capabilities regressing in other key aspects, such as machine tools, injection molding, shipbuilding, consumer goods, and so on.

I think politicians hook onto manufacturing as an ideal of an industry which should pay well and has a low barrier to entry. If manufacturing grew 15% every year but employed 15% fewer people every year due to automation, it would be considered a catastrophe.

The reason most people want more manufacturing in the US is because they want manufacturing jobs. It is only within the last few years since the pandemic that we started to care about domestic manufacturing as a matter of national security

> I think this is a win for national security.

I think that's largely the point. Obviously, it's a balancing act, but when market forces create a situation that is incompatible with national security it really is the governments job to address the situation even if, as in the case with Chinese EVs, it means a bit of pain for consumers.

It doesn’t elude people, they just think that the 80% of manufacturing that isn’t high end shouldn’t have left and that it was criminal for it to have been shipped overseas at the expense of middle America.

Moreover, share of GDP has been falling because prices of manufacturing goods have been all falling vs inflation, pressured by foreign competition and by productivity improvements, while rest of the economy - services - face none of that: they usually (except corner cases like call centers) can't be outsourced abroad, and output there depends strongly on labor inputs (e.g. waiters) thus making productivity stagnant almost by definition.

So there isn't a problem about manufacturing, never has been. Problem is strictly about manufacturing employment, which is of course, inexorably falling and will continue doing so and every politician promising to reverse it is a blatant liar. It's falling much like farm employment has been falling 40-70 years ago, sure it's traumatic, ruins livelihoods of millions of families many of which will never recover, destroys not just their personal finances but their source of pride and sense of self in many ways. But just like it didn't result in decrease in food output back then (quite the opposite happened), it doesn't result in dearth of manufacturing products now.

Let me tell you a story:

Tektronix lifted the Portland Oregon region right up. Was called silicon forest.

At that time, Tek was funding startups its employees thought up after working for Tek, getting great education provided directly by the company as well as through college partnerships.

Tek also literally trained a workforce here by educating any of its employees and by doing programs with suppliers to do the same.

I am a product of that time.

A drive through this region in the 80's and early 90's was awesome! Shops of all kinds, Tek itself had COMTEK which could make damn near anything, and opportunities abounded!

Howard Vollem died and the MBA took over.

COMTEK was torn down, work was sent overseas, education stopped, startup funding stopped, and soon a drive through this region looked very different: hair nails and laundry.

While large scale manufacturing has grown, the rest has suffered huge!

Our military can't find the capacity it needs! And they, along with aerospace, are the best customers there are, with auto in some parts too.

The rest has been gutted.

That is what we need to fix. It matters.

If companies won't do what Tek did, and that is invest in the region and it's people, and they won't because getting max dollars at any cost matters more than sustainable business does, then we must have robust small to mid sized manufacturing.

Where else will our future skilled labor come from? And I left for higher end professional work and software. I can make anything I can draw, it was damn good at it too. Saw way too many places close and there's no way I can raise a family on that and I quit ... tons of us did.

The skills I have are rare and in high demand. Young people today can't get them like I did, and that adds right the hell up.

You think your arguments make sense. And you are not wrong. They do, but that is not the problem.

The problem is for your argument to make sense, a ton of people and manufacturing potential is lost and nobody seems to recognize the massive opportunity costs in all that.

And frankly if large companies aren't going to do it and get a return on that investment, then our government damn well should. We do really put our national security at risk doing otherwise.

> I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

Cuz letting it manage itself works so well

> I think it was a big case of corporate welfare that will result in somewhat increased chip production in the US.

I see two issues with this:

One, this is the same kind of subsidies and meddling with the markets that we accuse China of doing. If we are adopting state-led approach, it can be done in a serious manner without being hypocrites. But that would also require admission of some mistakes.

Two - why is it seen as okay to give tax breaks to an engineering company, why not give tax breaks to engineers themselves instead? Companies are imaginary, people are real, why not give incentives to individuals?

At least that money won’t be squirrelled away in tax heavens.

> One thing that I think that frequently eludes people in this discussion is that US manufacturing has pretty much been constantly growing for the past century; just its share of GDP has fallen as other sectors have grown faster.

> ...I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

I think the problem you're having is you're thinking of manufacturing in terms of dollars, like an aloof economist.

Yeah China has only ever really been a major player for about 3 decades. In a third of a century it has Actually SHRUNK American manufacturing to the point where there was genuine knowledge loss. It was cheaper to manufacture things in China so we used China, and now America doesn't even have the capacity to manufacture anything on the scale of China.

>just its share of GDP has fallen as other sectors have grown faster

I think this is inn-accurate. You're looking at a century of data but China only took 1 decade to overtake the US. We're now three decades in and the overall decline of American manufacturing is pretty evident.

> I think this is a win for national security. I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

National security and global freedom of navigation are essential preconditions for our current level of economic output and quality of life. In the long run, it’s not an either/or.

RE: National Security see his sobering thoughts on the supply chain: https://timculpan.substack.com/p/what-supply-chain-security-...

Just about all "made in USA" products today are "assembled in the USA from global materials/components." This usually means that the absolute bare minimum of work is done in the USA in order to claim US origin. For example, pillows that are sewn over seas from foreign fabric and the only part that is done here is inserting the filling and closing the final stitch.

The US was once a global leader in textiles and now virtually all of that industry is gone. The same goes for many many other industries.

This might be true, but the number of jobs has fallen. It kind of comes from our measurement of GDP. If you have a high-tech factory that employees 10 people to make a $10,000 unit price electronic, that can easily become more GDP per factory than one that sells nails and screws for a penny each with 100 employees.

As Henry Hazlitt said

The art of economics consists in looking not merely at the immediate but at the longer effects of any act or policy; it consists in tracing the consequences of that policy not merely for one group but for all groups

This might be true, but the rust belt is called that for a reason.

I also wonder what's the share of non-disposable products in US and other Western countries manufacturing.

"just its share of GDP has fallen"

The little "just" does a lot of work here. It means people in the manufacturing industry are being left behind.

If the rumors about Elon's robots are true, I see a lot of manufacturing coming back to the US.

However, I do not see this benefiting regular workers.

> "I don't think ... probably won't be a win"

Accidental double-negative, right?

that US manufacturing has pretty much been constantly growing for the past century;

Really? Every time I see this claim its based on some citing some statistical mismash that the person citing does not understand and cannot explain.

How much tonnage of merchant shipping does the USA build in 2020s versus the 1960s?

How many TVs and computer monitors does the United States make in 2020s versus 1990s?

How many tons of steel does the USA make in the 2020s versus the 1970s? Of tool steel?

How many nuclear reactors are produced in the USA in 2020s versus the 1970s?

How many railway rails graded for high-speed trains are produced in the USA in 2020s versus 1980s?

How many CNC mills are produced in the USA in the 2020s versus the 1980s?

How many artillery shells are produced in the USA in 2020s versus the 1980s?

How many jet engines are produced in the USA in the 2020s versus the 1980s?

How many car engines blocks are produced in the USA in the 2020s versus the 1980s?

How many computer hard drives are produced in the USA in the 2020s versus the 1990s?

How many motherboards are produced in the USA in the 2020s versus the 1990s?

(Also, adjust all the comparisons above for population growth, we should be comparing manufacturing production per capita)

If you think that these comparisons are misleading because there are 'quality changes' please tell me exactly how you quantify these changes in quality.

> US manufacturing has pretty much been constantly

Middle class manufacturing jobs have fallen off a cliff and completely destroyed huge swaths of our country. Take a tour around Bridgeport, CT sometime as a great example. The northeast is littered with towns like this. These executives and their buddies in Congress mortgaged our middle class for profits by sending all our industry to Asia.

> I don't think the government applying such levers to change how the market allocates capital probably won't be a win for economic output or quality of life.

We've got China cheating with their massively deflated currency, so how the market allocates capital is already screwed.

> transferring knowledge from TSMC to a US workforce.

No, no it's not.

When semiconductor manufacture moved to Asia, this was generally done under a "technology transfer agreement", which was an explicit agreement for US companies to transfer their (usually older) tech to an independent local company who would then be allowed to manufacture it and develop it. This is how TSMC started, by doing a deal with Philips to manufacture for them but also to trained on the tech and to be allowed to use it themselves.

This TSMC US fab (and Samsung's new fab) are not under such an agreement, it is directly run by TSMC with no explicit goal to transfer technology. I think it was a mistake for the US CHIPS act funding to go to such a venture without a clause for technology transfer back to a US company.

The workers can walk away with whatever is in their heads

When the island of Taiwan gets the Alderaan treatment, I'm not sure that will be an issue.

Why would TSMC ever agree to such a deal?

American companies already know how to manufacture older chips. It’s not like TSMC is light years ahead of Intel. They’re ahead, but not by so much that their older generation tech would be transformative.

Aren’t Philips and ASML both effectively under American control anyway? Is the TSMC part of the stack that special in terms of actual IP (versus more squishy organizational know-how)?

I actually think the extreme density and breadth of manufacturing in China is going to continue to outpace US manufacturing. They have multiple enormous manufacturing hubs connected by high speed rail lines over a wide geographic area. The US has no equivalent to the likes of Shenzhen and Guangzhou, where you can finish a PCB design in the morning and have the prototype in hand that evening. You can go to the Huaqiangbei Electronics Market and find exactly the right motor and controller for your specs, pick up specialized sensors, and build your next rev overnight.

I have lived in the Bay Area my entire adult life. We used to have Halted/HSC, we used to have Weird Stuff. We used to have Triangle Machinery Co in Santa Clara. Now everything is gone.

I think it’s great that we built a semiconductor manufacturing plant. That’s important for strategic manufacturing. But we’ve so thoroughly destroyed our manufacturing base, let the factories rot, and financialized property value that the “weird place with random electronics” can no longer even afford to do business. Starbucks makes more money, so in it goes.

US politicians love to shout about manufacturing. “Manufacturing jobs jobs economy growth.” But these people DO NOT understand how things get made. They have no serious industrial policy. They do not know the value of a high speed train connecting manufacturing centers. And even if they did, the entire apparatus of our government is set up to stop it.

Manufacturing workers need education. They need housing, transit, health care, maternity and sick leave. They need secure jobs and extra income that allows them time off to take classes to learn new skills.

I’m glad we passed the inflation reduction act, and the CHIPS act. We need that investment. But it’s going to take much more than that to “bring manufacturing back” and I’ve have seen time and time again that we do not have the vision or capability to move in the ways that would be required.

I hope manufacturing comes back. We desperately need it. But I’m quite frustrated that despite some marginal progress, the serious changes we need are not on the horizon nor seemingly beyond it.

> They have no serious industrial policy.

Ignoring how we define "serious", they do have an industrial policy. You just may not agree with the wisdom of the outcomes wrought by the regulatory regime. I don't know how you ever compete with the developing world that has a surplus of people and comparatively lax regulatory framework for everything from labor to the environment.

The policy is to move all the dirty work to someone else's back yard. It seems to work as long as the shipping lanes stay open and the other economies have something to gain (room to grow and raise standards of living).

Perfect comment, it's important to celebrate but more important to keep in mind it's a tiny piece of the public infrastructure and government inertia we need to do this correctly. We need exactly what you describe, and I want to bolster the mention of education. China produces 2x the stem phds we do every year. Sure, bigger population, but they also have a growing share of citations. (source for both is suleyman's book "the coming wave")

I came here to say something like this. I've worked in product development in the Bay Area for 30+ years and brought numerous products to market, mostly manufactured in China. There's nothing like the density and ability of manufacturing that's in China (and more broadly in other parts of Asia). In the US I've worked with great molders and toolers, PCB fab, machine shops, CMs, etc. but the ability to turn on a dime and get stuff done quickly in southern China is insane. In the Bay Area you see billboards for esoteric SaaS products and credit cards for startups, in parts of China, they are for molding machines and CNC tools. You drive by rows of roll up doors in the base of apartment blocks and each stall/shop is filled with bar stock, plastic pellets, CNC machines, injection molding machines, etc.

You'll also see people doing complex repairs of mobile phones sitting on a stool on the sidewalk. The level of skill and access to tools/spare parts that is endemic there is completely different than the US.

As they say, the perfect time to plant a tree was 20 years ago.

> and financialized property value

Do you think China has not?

I've long been of the belief that, much like uranium enrichment, supply chain integrity of semiconductors will become a national security issue. We've seen it already in reverse with export controls being placed on GPU cards, and of course there is a reason the NSA operates its own chip fab. The threat to western economies of (for lack of a better term) "poisoned" chips making their way into phones, laptops, industrial SCADA equipment, etc. is real if nascent.

On-shoring top-tier manufacturers like is absolutely a win not just for the high-tech manufacturing sector but also for the US and allied countries.

>This is great news, and we should celebrate.

Couldn't agree more.

I wonder how Taiwan feels about this. From their perspective jobs are getting offshored from their country because of massive subsidies, and the strategic shield of having most critical semiconductors coming from them is getting getting thinner. At the same time they can't complain because only the US could defend them from China.

I've always thought there's some geopolitical chess at here. The US can't abide being completely dependent on the island of Taiwan. So if TSMC wasn't willing to do this, the US might fund an alternative. This could leave Taiwan no leverage at all.

Now, with some US based production, TSMC is still in charge, and more resilient to disruption. So it may still be a very strong move.

The strategic shield isn't getting that much thinner: this fab is a generation behind last year's iPhone Pros and MacBook Pros.

Isn't this better for Taiwan because it strengthens their ally, The USA?

If China would just wipe out Taiwan's ability produce chips, and disables part of the US information tech supply chain, then it would be bad for Taiwan right?

One can only hope that the US learned from its mistake, and doesn't allow chip manufacturing to go offshore to that degree again in future peaceful times.

It wasn't really a mistake. At the time Taiwan and South Korea were advancing into semiconductors, the US was more concerned with Japanese domination of the industry so having 2 small countries as alternatives to compete with Japan in some sectors of the industry was beneficial.

It's a lot harder to go to war when countries depend on each other economically.

> US learned from its mistake, and doesn't allow chip manufacturing to go offshore to that degree again

I don't think USA made any mistake. It was always heavily invested in South Korea and Taiwan. Neither of them would even exist today without USA's investment, interest, and stewardship.

Intel is the one that made the mistake.

Meanwhile we're offshoring all of our software engineering jobs at a breakneck pace with no regard for the consequences on our future.

The whole same story is going to play out again and in 20 years we'll be panicking because nobody in the US will know how to write software anymore.

Agree that TSMC is good news.

> US manufacturing is about to be reinvigorated

I’d suggest you post a small/medium quantity machining RFP on MFG.com with a medium to high complexity. I’ve been quite discouraged that US vs China price differences are 5-10x. (My part was a custom M0.8 screw in quantity ~500.)

It seems that without a vibrant base of small businesses, it will be very challenging to truly reinvigorate US manufacturing. And that would require reforming the finance sector/allocation of capital that currently is skewing really heavily towards “scale”.

So the free market was the enemy all along, what we needed was state planning?

The free market actually works great under the assumption of governments not acting irrationally.

However, since China has expressed interest in war with Taiwan (not a thing advised by the free market), someone needs to address that.

In terms of economics, this is a net loss, but then again, the effects of war in Taiwan would be worse.

The free market is an innocent scapegoat that never existed in any government. As long as the State makes the laws, it is a form of state planning. The only different between Soviet Russia and modern Western countries is how heavy the hand of the state tries to move the needle of the market.

But I agree on the sentiment: everybody seems to have decided the state should control the market even further than it did three decades ago. Free market was never given a chance.

The free market often needs nudges in the right direction. The free market is a rule of thumb and not an actual Scientific Law. When left unchecked it ultimately eats itself

This is a move with swinging geopolitical implications. But the value and urgency of the reinvigoration of the American manufacturing centre cannot be overstated.

China's gonna be a bit salty though.

What I really want them building in America though is low-end AMD chips for development boards.

4-Core/8-Thread CPU, 4-Core GPU, 16 GB ram, sane IO, and however many of Xilinx's FPGAs they can put on it without overdoing it.

People would be able to make some pretty decent things with that.

Notably, this was started in 2020 with a $12B investment - https://www.tsmc.com/static/abouttsmcaz/index.htm

Then in 2022, TSMC invested another $18B and received $6.6B from the CHIPS act.

My bet, is TSMC was given a “you build in the US or we wont give you defense contract work” in 2018-2020 timeframe lol

https://en.m.wikipedia.org/wiki/CHIPS_and_Science_Act#Histor...

> My bet, is TSMC was given a “you build in the US or we wont give you defense contract work” in 2018-2020 timeframe lol

My bet is that TSMC recognizes they are a crazy geopolitical pawn. And is frankly playing their part.

Once China develops chip production abilities similar (but not necessarily better) to TSMC, they’re free to destroy Taiwan. Then they’ll be the sole cutting-edge producer, meaning that everyone will continue to do business with them despite their behavior.

TSMC and the US recognize this. If TSMC bring their tech to America, they’ll at least be safe to continue manufacturing (“for the shareholders”). It also is self-serving because it changes the geopolitical game. It increases the risks to China of an invasion, and favors to the US increase the odds of US intervention (good for their patriotism).

Finally, it’s pretty well established that the US defense industry prefers local factories for security purposes. They’re obviously interested in preserving this ability domestically, and most companies recognize that and accommodate.

> My bet, is TSMC was given a “you build in the US or we wont give you defense contract work” in 2018-2020 timeframe lol

TSMC is in an extremely precarious geopolitical situation; China's hardball is a lot scarier than Trump's. Expanding their geographical redundancy through billions in handouts is pretty appealing to investors.

I think you are overly optimistic, this is an older process, which means that next iphones will need to source their CPUs from TSMC fabs in Asia, not from US. There also will be no knowledge transfer, that was not part of the deal. It's more of a national security political message than a real change. I guess it's better than nothing. I wouldn't call it a significant win, but it’s a step in the right direction.

And fabs are not enough:

https://www.bloomberg.com/opinion/articles/2024-09-03/us-nee...

You can't lead in the energy transition or produce chips without the supply chain and critical minerals:

https://www.wsj.com/world/asia/china-harnesses-a-technology-...

https://www.politico.com/news/2024/08/29/biden-minerals-pric...

https://www.politico.com/news/2024/09/13/steelmaker-biden-cl...

There's still a lot more to do to actually make it work before you can celebrate a win.

...transferring knowledge from TSMC to a US workforce. This is a significant win for the US...

I get what you're saying and I agree, but there's some heavy irony in saying that considering that's exactly how TSMC started out but from the opposite side; Transferring knowledge from "the west" (RCA from US and Philips) with ITRI it evolved into a project of Taiwanese state which culminated in TSMC.

> we in the US are going to be building our own future both for chips and for energy security.

> just the beginning of the amazing industrial policy

Isn't intel one of the biggest companies in the world, makes chips for everyone and everything and based in the US?

If the marker for industrial success is supplying Apple Inc, Intel did until 2022?

It also helps TSMC, i think. More "allegiance" with the US while China is scrambling to catch up

It's also terrible/impossible news from a USD perspective if the US produces things nationally in a significant way (importing less, selling less printed USD in exchange for goods).

Printing insane amounts of USD to allow for systemic government over-spending and huge untenable government debt doesn't go hand in hand with not importing most goods. You can't keep your currency strong if you can't force others to buy your currency.

If you have a lot of production in the US, this is going to cause hyperinflation to come sooner.

In the long term it won't matter, the end result is the same, but if production significantly moves back to the US it will be very scary from a currency perspective.

I don't see the argument here. Importing less leads to selling less USD (yes), somehow leading to devaluation of USD? Is the implication that the dollar is strong because the US government "forces others" (foreign manufacturers) to buy it? Isn't that the opposite of the first thought, which implied that "selling less printed USD" was the reason that domestic manufacturing would be inflationary? I don't understand the causality, and it doesn't match my mental model ("a country that can build things domestically at a competitive price point should be deflationary because now there is more supply of stuff and equal supply of money"), so I think this could do with some expanding.

> US manufacturing is about to be reinvigorated, and we in the US are going to be building our own future both for chips and for energy security.

Don't count on it. For every high-end chip you need hundreds of commodity parts to support them, and nobody is investing in US factories to make $0.001 capacitors or $0.10 connectors. You just can't compete with cheap Chinese labor, so the US supply lines will never be able to equal a city like Shenzhen.

Unless the US is willing to get rid of capitalism and switch to a plan economy, most of those expensive high-end chips will just be shipped to Asia for assembly. So much for building your own future.

> most of those expensive high-end chips will just be shipped to Asia for assembly

I get the first part of your comment, but why wouldn't all the missing components be imported for assembly in the US? SMT lines in particular don't need that much cheap labour to operate. Even Brits can assemble PCBs!

"You just can't compete with cheap Chinese labor"

You sure can, we have cheap Mexican labor... and we have a much healthier trade relationship with Mexico.

Or how about instead of 0.001 caps and 0.10 connectors we pay the little bit extra and not sacrifice living wages and environmental regulations?

You don't need a planned economy to legislate that all of you source material must be produced in accordance to local laws.

Those components are not used in military products. Specialized vendors manufacture passives for the military. AVX, CDE and Vishay are just the first three I recall.

Hold your horses — this will only produce a fraction of the chips, and probably at a much higher cost.

It's a step in the right direction for the policy goals, but they've really just entered the woods with this one.

But it also makes us not dependent on a place that China has their literal sights focused on.

>and we have top tier production here in the US

TSMC's process that they are bringing to the US is 2nd tier. The crown jewels are being kept at home.

If Intel can get their act together, then we will have top tier in the US.

That's not really true, afaict. This press release[1] states that Fab 21 (Arizona) phase 1 is 4nm, which is not the best but clearly is enough to manufacture the A16, phase 2 opening in 2026 will be 3nm, and phase 3 will be 2nm or better. I'm not a semiconductor engineer so maybe there's some process nuance that I don't know, but it certainly seems that this is at or near the top of TSMC's process list.

[1]: https://pr.tsmc.com/english/news/2977

I feel like you are running around high fiving everyone for a job well done and the first chips aren't even off the line yet.

This is a huge milestone, but it seems a little premature.

Wave the flag and have colourful fireworks with hand at the heart and tears in the eyes, this is a glorious moment the children will cheer its glory in glorious essays!

Maybe you should visit the rust belt/midwest before SV starts patting itself on the back for single-handedly re-invigorating the economy.

I get the spirit, but flyover country is not doing great. Unemployment is rising and there is a severe lack of decent paying jobs. Chips are great, but everything else is made in Asia. Increased automation is making a ton of jobs obsolete and there is no solution in sight yet. Chips ain't gonna do it.

Maybe you should look at actual stats for what's going on in the economy before being completely cynical.

The investment in factories is absolutely massive over the past few years. The Inflation Reduction Act is bringing massive amounts of manufacturing into the US, starting with the lowest value add of assembly, and after that additional suppliers lower down the chain will be built up too.

It is not SV reinvigorating the economy, it's not happening in SV, it's happening in small towns all over the country. It's happening due to the bills that Democrats passed over Republican opposition, but because of politics, it's not being trumpeted as a partisan win in the towns where factories are being built.

You are completely correct. But the chips have immense strategic value. Not being able to manufacture them would be catastrophic in the event that China cut us off.

> After all the wailing and rending of clothes, the industrial policy worked out great and we have top tier production here in the US, transferring knowledge from TSMC to a US workforce.

> This is a significant win for the US, and just the beginning of the amazing industrial policy passed over the past few years.

I'm not sure these conclusions are justified. It's the 'seen vs unseen'.

This is (only a few years later than the rest of the world's) state-of-the-art manufacturing, built only with the expertise of a Taiwanese company, that relies on the technology of a Dutch company, that in turn purchased (and has since monopolized) its IP from another US company, twenty years ago, and only then because a number of other companies (notably Canon and Nikon, both in Japan) were excluded from using it.

It is not something to be celebrated. What TSMC and ASML are doing is amazing, but we could be so much further ahead.

If you look at it from a geopolitical angle, it's much to be celebrated. It means US can rely on its democratic like minded friends to help protect the supply chain of cutting edge chips, against the now very visible alliance of dictatorships (Russia, China, North Korea, Iran).

And make no doubt about it, there is a democratic alliance vs dictatorships here. Russia is aggressively sourcing artillery shells from North Korea, ballistic missiles from Iran, and financing and weapons from China. China incidentally is the economic caretaker of Iran and North Korea.

US accuses China of giving ‘very substantial’ help to Russia’s war machine https://www.politico.eu/article/united-states-accuse-china-h...

China’s Double Threat to Europe: How Beijing’s Support for Moscow and Quest for EV Dominance Undermine European Security https://www.foreignaffairs.com/china/chinas-double-threat-eu...

This also lowers the Taiwan risk, which was increasingly high after the China chip ban.

Yep. We should be investing our tax dollars in our economy and our people.

I think it falls back on lots of elites feeling that only “one special group” can do a thing. The pendulum swung too far into the “globalist”agenda and now it’s swinging back. I’m sure it will overshoot and we’ll be back to globalism within a decade, but I for one welcome anchoring more things to the domestic economy; the US is a huge country with lots of resources and lots of untapped potential (despite not being China big) and we could use some of that trickle down economy that the globalists have been hoarding for a while now.

Not only that, the US now also has some backup when China will invade Taiwan and take over TSMC.

>This is a significant win for the US, and just the beginning of the amazing industrial policy passed over the past few years.

I don't consider it a win. I consider it a loss. This is a desperate move by the US. Intel making better chips then TSMC is a win. The government strong arming Taiwan with "protection" from China in order to gain this technology is a display of American incompetence.

But then again maybe is't not about fair play. If the US wins by unfair means, it's still a win? A pathetic win but a win none the less.

There's a common pattern here, it's easier for them to import fresh meat than fix the rotting carcass back home

Whether that's fair or not, who really cares, what can we do about it

Someone in the thread mentioned "China's gonna be a bit salty though."

I see what salty China looks like now.

> US manufacturing

America is a continent.

Why did it take the United States government having to invest in US labor?

Why wasn’t the “free market” Capitalism allocating resources to the United States if in fact, it is the best place for this to happen?

Or is this just garden variety realpolitik nationalism?

"US manufacturing is about to be reinvigorated"

Oh i'm going to be downvoted into oblivion for this one.... but I think this is a win we can give to Trump. It was a hard focus of his, I think he put the right people in place to do it, and I think time is going to prove he was right to do it.

I didn't vote for him in 2016, but I think it's important to acknowledge it.

This is because of the CHIPS act. Other than Trump's constant whinging about China, he didn't have anything to do with this.

You really think it's a success to force Apple to lose money to make US politicians look like they are "doing something" about a world economy that is increasingly leaving the US in the dust?

Meanwhile in China, 1000 engineers (to one in the US) are building all kinds of electronics and embedded systems on shoestring budgets that truly force them to learn engineering. China's industrial policy architects are likely laughing at this big folly on the part of the US.

The worst is the 100% tariff on EVs which keeps the US in an artificial economy of gigafactory, high-end nonsense when the rest of the world will be getting true economies of scale from EVs which are actually simple, reliable and low cost.

It's deeply embarrassing that the US must suffer poltical rule of its economy along with the double embarrassment of seeing other nations do it so much more effectively.

Industrial policy should be measured in terms of person months of career acceleration (experience) per dollar. The US focuses on helping prolong the dominance of internal combustion engines and taxing high profile, high-end companies that do not offer skills that transfer well into the rest of the economy. How many Apples are there? Will forcing manufacture in the US suddenly result in another company doing 2nm process and competing with Apple? It's absurd.

I know of a variety of small and medium sized US tech companies (aerospace and 3d printing / robotics) that were almost sunk by US "industrial policy" becasue they relied on a small number of China-manufactured inputs that suddenly became unavailable, forcing unplanned re-engineering and work the companies could not afford. Sadly, one went under. Meanwhile, the US firms that import finished goods are thriving selling Chinese manufactured gear -- Chinese companies didn't have to pay US tariffs on the same inputs. Utterly absurd.

Politicians should stay out of the economy and focus on moving us closer to nuclear war and promoting the religion of American Exceptionalism.

> Industrial policy should be measured in terms of person months of career acceleration (experience) per dollar

Sounds about right but how would you come close to measuring that?

>“It's deeply embarrassing that the US must suffer poltical rule of its economy along with the double embarrassment of seeing other nations do it so much more effectively.”

The Chinese Communist Party exerts far, far more control over all sectors of their economy than US politicians do over the US economy.

https://www.cbsnews.com/news/xi-jinping-china-capitalism-60-...

> This is great news, and we should celebrate.

On the other hand, this is a protectionist policy that has been straining US' relations with its allies. That development means the US empire is a little less mutually beneficial, and a little more beneficial to the core.

Its success requires these allies not to reciprocate, and this is a long-term prospect that only time will confirm.

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.

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.

AppleTV, HomePod, a new display. Could be anything. There could also be government or corporate contracts requiring the mass production of a slightly older chip for something.

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.)

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."

[1]: https://archive.ph/LqV4M

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.

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.

Sounds a lot like corporate slavery.

When the machines need to be running 24/7, why do they not hire qualified workforce that runs in 3 shifts?

Or, hopefully, the engineer who is paid to fix it is paid for their time.

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.

If you believe there has been a decline in American work ethics, this actually seems like a good thing. Optimistically they could reach a good middle ground here.

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.

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...

I've seen those same articles, but also ones saying that was largely a ploy. There were billions of dollars in subsidies that took some time to lock down, and the reported problems with American workers evaporated right after the money was committed to TSMC.

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?

To build a factory from scratch you must first invent the universe

AFAIK TSMC does not manufacture the machines they use to create processors so in any case they would be moving the equipment into a facility.

yeah I mean this isn't factorio...

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.

Arizona actually has a lot of water because of several successful and ambitious irrigation projects in the last century.

So much so, that it has become an agricultural region for growing notoriously water intensive crops like alfalfa and pistachios.

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.)

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.

SMIC: "You got anymore if those Secure Foundry grants lying around?"

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.

I think the vector people talk about most in this context is denial.

If an adversary wants to deny access to a fab on American soil they'll need to deny access to dependencies or attack the fab itself.

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.

Also, some security measures requested by Apple to manufacturers in other countries are probably illegal in the USA.

That's interesting...do you have any specifics?

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.

Not a specific context in the article. Just wanted to see what (if) aspects it might have as I've seen some other comments around that.

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?

Do the chips get shipped to China for assembly?

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 they meant for assembly of the iPhones.

No. That was some misinformation. The chips are being fully packaged in the U.S.

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.

[flagged]

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".

Assembly of the phone or device using the processor.

At what cost meme, but literally, what's the cost of a chip made in TSMC US vs TSMC TW.

It's ok their profit margin will go from 50% to 47%, they already have so much money that they don't even know what to do with it anymore

I'm mainly curious if TSMC estimates that US fabs would cost 50% more is confirmed or not.

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).

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.

Exactly: https://en.wikipedia.org/wiki/Island_chain_strategy

So the US doesn't care about Taiwan, it cares about China. Taiwan is just a tool.

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.

My assumption is if they can do that chip, they can do others and there is less need to defend Taiwan and the massive TSMC fab there.

Good news doesn't have to mean overwhelming good news. Directionally, it is clearly good, not bad, and not nothing.

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.

(Some) "in small, but significant, numbers".

About time. It's easier to secure supply chains domestically.

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.

This happened significantly faster than I anticipated.

The question is where are these chips packaged? Potentially the wafers are shipped to the east for packaging, assembly and test.

People will be happy as long as it's not Taiwan (e.g. Singapore or Malaysia).

Packaging is much simpler

TSMC started sampling N5 in 2019 and full production in 2020. This means the US finally has a 5-6 year old TSMC node in the US.

Hardly a big win.

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.

Should other countries put tariffs on devices made with chips that profited from the CHIPS for America Fund?

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.

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..

This happening at the same time as germanys intel project freezes makes the fall seem even harder

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.

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.

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...

Yup. Just pass another tiktok bill to force sell factory to US buyer.

Holy shit they have 1,900 employees for that

It doesn't need to be a viable replacement. Even if it only ever makes chips that are 1-2 years behind, it's still a huge strategic benefit for the country.

Does it have to be or does it just have to be enough to be a deterrent to China?

I wonder if the strategy behind the CHIPS act is to have enough “backup” capacity in the US that it isn’t completely vulnerable.

China’s interest in Taiwan is about controlling the sea routes, not about chips.

How so? They also are extremely incentivised to make this happen. A war on your front door is not good for business.

Tsmc is mostly governed by Taiwanese who would like to maintain Taiwanese sovereignty

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

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.

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.

^https://news.ycombinator.com/item?id=39273830

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.

this is wonderful news. at the same time i hope this doesn't weaken the security posture of Taiwan

The dependency on Taiwan isn't going to go away any time soon, nor is the Taiwan Relations Act (which replaced the Mutual Defense Treaty between the United States and the Republic of China in the late 70s).

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.

It's really not about the jobs, it's about national security. The US needs the ability to fabricate chips on its own soil where the threat of China invading Taiwan isnt a concern.

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.

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.

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.

How much of this fab's supply chain still comes from Taiwan and/or China? Most especially, where does the fab process equipment itself come from?

> the fab process equipment itself come from?

Isn't there only one for this kind of scale: ASML?

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.

Besides lithography there are also several US suppliers for semiconductor manufacturing, Applied Materials, Lam Research, KLA.

Perhaps a stupid question - what exactly is TSMC contribution to producing Apple designs on ASML equipment?

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!

It does seem weird that there is this separation though. I would have assumed that there is a lot of overlap between machine design and operation.

This is a really sharp summary. I hope it's correct, because it was fun to read.

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.

That ASML is not undercutting TSMC and running off on their own should be telling? There's more to a running a fab than lithography.

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...

I wonder if the US plant ASML equipment also have a destruct mechanism like the Taiwanese plants have.

I don't see this a positive thing as European. Why not produce the mobile processors in Europe?

Isn't EU basically abusing every tech company for money every month? Why would Apple be inclined in investing anything in EU.

Plus, if US has skill shortage, can't imagine how bad it would be in EU.

Wonder how apple was able to curb the (assumable) higher cogs of producing this domestically.

This news could be bad for Taiwan

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.

Taiwan is still in charge of the fab.

Has either company verified this?

Excellent to see more manufacturers come to the land of the free and home of the brave.

The obvious question is when are they going to build 3nm chips here?

This is title seems to be quite the overstatement of the facts.

And then sent for packaging to Taiwan and assembly in China?!

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.

This may be a very expensive proof of concept… but it's definitely a concrete step toward their goal

Apple isn't the only company in the world that needs CPUs

America isn't the only country with a roadmap to manufacture 3nm silicon by 2026. If history has anything to say in the matter, it's likely that Samsung will have the US fabs beat on yield and price for a long time.

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”.

There's not actually that much uncertainty about what a TSMC chip fab does.

Maybe TSMC is sneaking finished wafers into the Phoenix fab at night and taking away blank wafers :)

It takes blank wafers in and produces finished wafers just like all other fabs. I would expect test and packaging are performed elsewhere.

good news like this is so rare. this makes me happy

Why can't America build a TSMC from scratch?

They did! It's called "Intel".

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.

Its called Intel Foundry

Wrong question. They'd still be buying Zeiss

This was build by TSMC, that's why it was completed in a few years. They have the know-how.

Have you even met Americans? We're terrible at this kind of thing.

Do you know who invented and developed transistors, microchips and their manufacturing processes?

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?

"Traditionally reputed" is a vague allusion containing multiple logical fallacies; can you cite some actual sources? Because it's nonsense. Quality is a factor of cost (and cost reduction), not the country of origin.

SOME of them are. A tiny fraction. At 4-5nm. But Taiwan is already making 3nm chips for Apple. Still better than nothing, I guess.

Curious to see when the US will force the TSMC to sell because it presents a national security threat.

Taiwan and TSMC are already USGov assets... why would they need to sell...

The new Apple chips are second generation 3nm. This 5nm stuff is old tech. Why are people celebrating?

Because all of these commentators are ecstatic to use their Made In America M4 processor in 2026!

...wait, M4 will be 3 generations old by then? W-well, a little reliance on Taiwan never hurt anyone...

very happy to see this!

USA

Are they hiring?

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.

Returning manufacturing to the US is a policy of both major parties right now.

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!

Why not? Are economists infallible? Even if there’s many of them they may have been taught the same material and risk groupthink.

What position is the US in if all their goods are manufactured abroad? What if the dollar stops being respected?

Yeah, optimizing for economic output isn't the only factor to consider. Having some degree of geopolitical independence and leverage matters when things go off of the happy path for whatever reason.

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.

There is a difference between having manufacturing capabilities and trade tariffs. You can in fact build your own chips AND trade with other countries for the same items at the same time.

Right, tariffs are another word for taxes that penalize importers.

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.

The era of people earning more from desks and mental gymnastics is over. The era of weaponization of supply and resources has begun.

All manufacturing will need to become local in some quantity for any country serious about its security.

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 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.

> 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!

> Both Trump and Biden handled the situation adroitly.

Trump started trade wars by raising tariffs. Biden passed the CHIPS act, the infrastructure bill, and the build back better act.

I wonder how bad the yields are and how long it'll take to get them up to being commercially viable.

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...

Once the US doesn't depend on Taiwan anymore, will they give up protecting and China will conquer it?

That's horrible news, as I was considering grabbing an iPhone... I guess only refurbished, but still, you can trust an American factory even less.

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.

Fortunately iPhones are not essential items you need to buy, so there's nothing forcing you to drop a hypothetical $2500 for one.

> 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.

Reply on news.ycombinator.com