TSMC ups Arizona investment from $12B to $40B with second semi fab | Hacker News Reader

TSMC ups Arizona investment from $12B to $40B with second semi fab

cnbc.com

451 pointstotalZero282 comments

Well that seems significant.

Article says the chips in the 2nd fabs will be better than the first. And will meet all domestic US demand. Not sure what all that includes.

I wonder how domestic the full supply chain is. If the US is relying on China for rare earth metals or something.

The US has rare earth mines and even more deposits. They just got priced out through decades of lower prices for those minerals. (Which just shows that resource allocation even for scarce resources, follows financial incentives similar to other financial considerations. Stuff being left in the ground actually makes them available for another day when they’re higher valued.)

Indeed, but rare earth mining also has negative environmental impact even above and beyond normal mining operations. The US is partly paying China to outsource the environmental devastation.

No, China is absorbing the economic impact of destroying their own environment.

There is no blame to be placed else where.

No one is tricking China into destroying their environment.

I don't think previous poster said they were.

China demanded it even, by basically cornering the rare earth market via dumping

Mitigating that environmental impact is what makes rare earth mining so costly. China is quickly learning that environmental remediation [1] is much more expensive and their appetite for destroying their environment for profit is quickly disappearing, especially as the middle class grows and moves up Maslow’s hierarchy.

From the article:

> “The rare earth industry is a pillar industry here in Ganzhou and we must keep it running, but there’s still more to do to figure out a truly environmentally friendly method to pursue sustainable growth,” Zhang Guanjun, deputy director of the public relations department of the Ganzhou Party Committee, said in an interview. “Ironically, because the prices of rare earths have been so low for a long period of time, the profits from selling these resources are nothing compared to the amount needed to repair the damage.”

[1] https://e360.yale.edu/features/china-wrestles-with-the-toxic...

Is the environmental impact actually significant enough to warrant consideration in this circumstance? I see environmental impact talked about a lot when it comes to mining operations but I don't think anyone outside activists cares if the area around a mining operation is impacted as long as a town isn't poisoned or something else with human impact and usually the "environmental impact" is just the local ecology and the emissions involved. People expect the area around a mine to not be very nice.

> Is the environmental impact actually significant enough to warrant consideration in this circumstance?

Abso-fucking-lutely! *Smack* The answer to that question is always a resounding yes when we're talking about massive industrial processes. The qualification "as long as a town isn't poisoned" is fantasy at these scales. Most mining facilities require tons of water so they're built near waterways and those are almost always populated downstream.

The tailing ponds contain very toxic waste that, if breached, will basically salt the earth around them and anything downstream. The tailings from a much less toxic processes like iron mining, for example, caused a massive ecological disaster in Brazil when the Mariana dam holding back millions of tons of mining waste broke [1]. It followed the Rio Doce to the Atlantic polluting over 600 kilometers of waterways along the way, affecting over a million people. Five years later people are still getting sick, fish stocks are way down, and farming along the river is crippled [2].

A few years later another dam in Brazil broke, once again contaminating hundreds of kilometers of waterways [3]. The Ok Tedi disaster in Papau New Guniea [4]. The Marcopper disaster in the Phillipines that still hasn't really been cleaned up decades later [5]. Aberfan in the UK [6]. History is full of these kinds of tailing disasters effecting millions of people, often the poorest with the least recourse.

Edit, forgot to add final point: Those Ganzhou tailing ponds are much more toxic than those above from lesser metals and coal. They're dangerous to approach without full PPE and much of the surrounding area has been polluted to the point where nothing will grow without remediation. Even assuming there's no accident, the environmental damage is already significant enough to cause political pressure in an authoritarian regime.

[1] https://en.wikipedia.org/wiki/Mariana_dam_disaster

[2] https://news.mongabay.com/2020/11/deadly-anniversary-rio-doc...

[3] https://en.wikipedia.org/wiki/Brumadinho_dam_disaster

[4] https://en.wikipedia.org/wiki/Ok_Tedi_environmental_disaster

[5] https://en.wikipedia.org/wiki/Marcopper_mining_disaster

[6] https://en.wikipedia.org/wiki/Aberfan_disaster

> I wonder how domestic the full supply chain is.

The supply chain necessarily includes some Japanese and Dutch companies for their specialized tooling that cannot be sourced from American companies. Many of the spare parts and raw materials can be sourced domestically.

However, it is ultimately unavoidable that many of the constituents within these spare parts, repair kits and raw materials would be manufactured/synthesized in China.

(I worked on the supply chain systems and processes at Samsung Austin Semiconductor for almost a decade.)

For at least part of this, ASML has had a large facility in the Phoenix metro (Tempe) for quite a few years. Assuming that's one of the Dutch companies you're referring to.

TSMC is going to benefit from the microchip industry that's already in Phoenix thanks to Intel's enormous presence here.

What are some of the companies in the microchip industry in/around Phoenix?

There are a lot: https://www.chipsetc.com/semiconductor-companies-in-arizona....

The article is wrote for tech illiterate audience.

They will use EUV for the first factory (2024). And High-NA EUV (next gen from ASML) for the second factory in 2026.

4nm,3nm are just marketing terms, and means nothing, since they stopped naming nodes according to their actual size since 16nm.

I don't think a tech illiterate audience has a clue with 4nm or 3nm mean

i dunno, they can google it pretty easily

But words are just, like, socially-defined metaphysical constructs that only occasionally and coincidentally overlap with physical reality, man.

"Physical reality" is also a construct, philosophically saying there's only information.

How much rare earths do you imagine a semi fab uses?

Chips primarily require silicon, which is not a rare-earth metal (it is rather abundant), but it takes a lot of expensive equipment and energy to purify it and slice it into wafers.

TSMC currently sources their wafers from Taiwan, Japan, Singapore, Germany and others:

https://investor.tsmc.com/static/annualReports/2005/pic/E-3-...

(edit: that PDF is out of date, but the wafer supplier list largely isn't)

No...chips are made on Silicon. What is required to make them into semiconductors requires a lot of special metals, gases and other chemicals. They source those things from around the globe...including Ukraine for things like noble gases.

>> https://www.thermofisher.com/blog/mining/whats-so-rare-about....

Rare earth metals aren't actually rare. They are called that because they were once thought rare.

https://www.thermofisher.com/blog/mining/whats-so-rare-about....

I've read Spruce Pine, North Carolina has the world's purest source of natural quartz best suited for making silicon wafers.

https://en.wikipedia.org/wiki/Spruce_Pine_Mining_District

> Not sure what all that includes.

Possibly a lot of currently obsolete military equipment that can no longer be produced due to lack of domestic parts suppliers.

I recall reading that modern chips use argon-neon-flourine lasers, (roughly 75% of the world's neon goes into the semiconductor industry) and ~50% of the world's neon comes from Ukraine.

This is because the Haber process, necessary for fertilizer, involves industrial quantities of forced air, and the USSR put adapters on their Haber process plants for gathering inert gases for titanium welding back in the 1970's, and those are now used to gather materials for semiconductors around the world. I don't know enough whether US or other Haber plants could also be adapted to gather up these noble gases, but presumably it would be more expensive than using what the Ukraine provided, or else it would have been done before.

That is one issue among many, obviously, but it suggests the scale of the problem for making all domestic supply chains.

If 50% of the world's neon comes from Ukraine, then 50% comes from outside Ukraine.

According to [1] 50.1% comes from the top 3 countries. And the worldwide market size is about $220 MM/year.

Let's say that these TSMC plants will consume fully 5% of one year's worth of worldwide production (this is an ultraconservative estimate, as the whole computer industry's share of the neon gas market is 50%). That would be about $10 MM. Even if the neon gas price spikes up by a factor of 10 or 20, it wouldn't move the needle when it comes to a $40 BN investment.

[1] https://www.futuremarketinsights.com/reports/neon-gas-market

I've had to look this up because everyone also brings up this point without any proper details.

The exact numbers are not known. It's supposedly 45%-55% (at the beginning people were claiming 70%). It's limited to 2 factories. One in Odessa and one in Mariupol.

Both of them shut down somewhere at the start of this whole endeavour. The Mariupol one is already in the hands of Russians. It wasn't that close to Azovstal, so it's due to reopen. The one in Odessa will not reopen until the Russians either take it or Ukraine agrees to the deal judging by all the missile strikes. But for the short term future its out of commission.

The fundamental part here is this:

If this is really such a big deal as everyone claims it is, then people will inevitably have to find a compromise with Russia.

But I will repeat again one thing that people don't seem to grasp. Arounds 30% of the worlds commodities are exported by Russia, another ~30% are exported by the US. The difference between these two is that Russia exports its own commodities and the US exports not only its own, but also its "little colonies" like Trinidad and Tobago commodities. You cannot just turn off the tap on 30% of the worlds commodities.

Alternatively Russia could fuck off and leave.

Oh, but you can. You see, there's this mechanism called price. If there's not enough of something, prices go up, and sources that were uneconomical become all of a sudden profitable.

> If this is really such a big deal as everyone claims it is, then people will inevitably have to find a compromise with Russia.

Why? Better to defeat them militarily and solve the problem once and for all - even if the world has to get by with slightly less advanced chips for a while.

Solving problems once and for all is very hard. Our track record with that is less than stellar. I agree with you, Russia needs to be defeated. But that is for moral reasons, not because it is overall the best economic solution for the world. Rather in spite of the economic consequences.

The biggest semiconductor equipment manufacturers in the US are Applied Materials, LAM Research and KLA. AMAT is a pretty big company with a wide portfolio of products but it doesn't do lithography which the Dutch company ASML has the only EUV product.

I've been trying to build a habit lately of finding the primary source of news like this, especially since news agencies seem to be very hit or miss with whether they link to their primary source.

So for whoever is interested, this is the original announcement by TSMC themselves: https://pr.tsmc.com/english/news/2977

Thanks, that's a really cool habit! Has this lead to any changes re how you perceive news? How far does news tend to deviate from the sources you find?

The most common pattern I’ve noticed is that news publications will typically include small snippets from the primary source, and then fill the bulk of their article with what amounts to opinion commentary.

It ends up being something that’s technically factual (at best), but spends most of its time trying to convince you of how to feel about the info, rather than just relaying information to you.

This has also inspired me to build a news aggregator site that consists entirely of primary sources like government agencies, company/organization announcements, and statements directly from the person in question on their verified social media.

Onshoring is a logical result of improved automation.

When labor costs lessen, transportation and logistics become the primary concern. Wouldn’t be surprised if 100 years from now most final good assembly is done locally.

Of course sometimes raw materials and inputs are more expensive to transport than the finished good, so there will always be cases where long distance transportation of finished products is still preferred

Onshoring in this case is a logical result of improved tax incentives. Almost every cost is higher in the US – not just labor.

Improved automation does reduce costs, but those cost reductions are even better in places where manufacturing is already cheap.

Semiconductors have a very high value to weight ratio, so shipping them globally makes more sense than it does for other items.

The cost difference between US and China for a fully automated factory is negligible. Even considering real estate costs and taxes.

No tax incentive needed to onshore in the long run

While semiconductor fabs are highly automated, the machines are complex enough and do extreme enough things that they break a lot, requiring human intervention. MTBF for many of the machines used can be measured in hours, it's not like you can set up and go home. TSMC has around 65k employees overall - a cursory search did not yield their makeup by job breakdown, although from 1997 through 2003 around 50% of their employees were 'factory' workers via [0].

I'd expect high-margin items like semiconductors to be more amenable to onshoring anywhere - certainly Intel has almost all of its fabs in highly developed countries, but I think this is hardly a cost efficiency move - rather it's a hedge against the failure of globalization and the risk of the US losing access to advanced semiconductor technology in case of strife in Taiwan. Russia has already lost this access. TSMC took in about $56B in '21, of which about $28B was 'cost of revenue', the amount they needed to spend on supporting this revenue (marketing, legal, production, e.g., not building factories or R&D). There's certainly room for increased production costs while maintaining a profit.

[0] - https://www.researchgate.net/figure/TSMC-personnel-structure...

Which is driven by concerns over the safety of Taiwan's fabs.

The highest costs in Semiconductor fabrication are the specialized machines needed. Most of them are at best dual sourced. Many are single sourced. The fact a technician in the US can make $100k when their counterpart in China would make $50k doesn't matter when the EUV Lithography machine is hundreds of millions of dollars.

Who makes those expensive machines? Are they made in-house by companies like TSMC? Or are they purchased from others? And whats the geopolitics of that part of the supply chain (made in taiwan? made in china? made elsewhere?)

In many ways, ASML:

https://www.asml.com/en/products/euv-lithography-systems

ASML is the very high end, but there are many other manufacturers as well. Canon and Nikon both make lithography systems used in many wafer fabs.

I really doubt anyone else than ASML can do 4nm (which this factory is planned to target).

Interesting 18min CNBC tour of ASML and their unique technology: https://youtu.be/iSVHp6CAyQ8

The machines are made by other companies. Lithography is the most expensive now and its made by ASML in Netherlands is the only solution for EUV lithography. But there are many steps in manufacturing and a machine for each. The major equipment manufactures are spread across US, Europe and Japan. In the US the three big equipment manufacturers are Applied Materials and Lam Research and KLA.

The costs are probably driven by extreme engineering requirements (temperature, pressure, corrosive chemicals, low contamination) and R&D costs to meet these requirements. The production volumes are not high either. Its not like they are building millions of these machines.

Asianometry on YouTube is a good source for this kind of topic.

https://www.youtube.com/watch?v=5Ge2RcvDlgw

Capex amortizable across the life of the fab and usually treated beneficially for tax purposes (though these fabs are so tax-special that may be meaningless). Opex is with you forever and affects the economics of the business much more profoundly.

The semiconductor industry is somewhat special in that it requires nearly constant investment in not just R&D but also in upgrading existing fabs to keep up with the cutting edge. And these are well into the billions of dollars.

This is not a problem for players who have dropped out of the race like GloFlo. But its definitely the case for TSMC, Intel and Samsung. Your last three serious players who aren't working for the PRC.

the cost reductions are necessarily worse in the places where costs are lower. I think you didnt do that math right. You might be right that the logistics for shipping silicon are still so good that this does not reach the amdahl limit, but it is not a consequence of your argument.

"Tax incentives" like writing off Capex have been there for the taking. So yes, the improved tax incentives with bipartisan support translates to stability for investors.

The benefits of JIT manufacturing outweigh everything else though. The waves of oversupply and undersupply in the semiconductor world are something that only JIT can fix, and transportation costs are only low if you go against JIT and try to bulk ship items.

It would make more sense to put the mega-fabs where electricity is cheapest in the Americas, and that certainly isn't Arizona.

Isn’t solar the cheapest energy per KwH? A sunny place like Arizona with tons of empty space should be able to cheaply deploy lots of energy.

The cheapest energy is from a hydroelectric plant built generations ago.

Fun fact: That's why Boeing first built their planes outside Seattle - the electricity needed to refine the aluminum was cheap.

These days, you can see that effect in the large datacenters that have sprung up near the Columbia River in Oregon.

I'd guess TSMC's siting decision had a lot to do with tax breaks, labor pool, and cheap land.

Probably also hedging for geopolitical risk? Not sure how only having facilities in a country at least somewhat likely to be invaded in the next few years affects the stock price but it probably doesn't hurt to put one in the middle of a country very unlikely to be invaded and very likely to want to purchase chips built domestically.

Fun fact most experts consider the US uninvadable: https://en.wikipedia.org/wiki/Invasion_of_the_United_States#...

Another fun fact. There were incursions on both US coasts during WWII. Neither were significant:

Long Island, Florida, and Maine invasion by nazi spies swimming to shore from German submarines: https://www.fbi.gov/history/famous-cases/nazi-saboteurs-and-....

Japanese in Attu and Kiska islands in Alaska: https://www.history.com/.amp/topics/world-war-ii/battle-of-t...

Cheap, seismically stable, land.

I think that depends how much build-out you have to do in order to install the kind of capacity required for a $40B semiconductor fab, which would hypothetically need some kind of major energy storage infrastructure or backup source in the event that solar is the main source of electricity.

> Large semiconductor fabs use as much as 100 megawatt-hours of power each hour, which is more than many automotive plants or oil refineries do. [0]

If a fab is running at that consumption rate for twenty hours every day for one year, it will consume approximately all of Arizona's 735,000 MWh of annual utility-scale solar, wind, and geothermal net electricity generation. [1]

Contrast that with the annual hydroelectric generation of Itaipu Dam in South America, at 79,440,000 MWh in 2019. [2]

[0] https://www.mckinsey.com/~/media/mckinsey/dotcom/client_serv... [pdf]

[1] https://www.eia.gov/state/print.php?sid=AZ

[2] https://www.power-technology.com/projects/itaipu-hydroelectr...

> Large semiconductor fabs use as much as 100 megawatt-hours of power each hour

Ooof, this kind of writing where the author doesn't understand units just gets my goat. Just say it uses 100 MW! Geez.

I prefer it as written, because the additional information of timeframe for measurement indicates that the author is not talking about momentary/peak power.

Which indicates the author knows nothing about how semiconductor fabs use power. Power usage is stable with a few percent...and 100MW is really on the low side for this fab. If all 6 phases are completed as they are expected to be...it's more like 250MW.

I think we can forgive the author for the low estimate considering that this project didn't exist publicly at the time of writing.

How does power usage remain stable in a fab? Please say more about how that works.

Much of the power is used for keeping air clean with temperature and humidity controlled in a very tight range. More energy is used for water cooling, waste treatment and other facilities. The rest of the power is used for production tools...which when a fab is running are generally consistent users. The power usage will scale with wafer loading, but again advanced fabs like these will run at near 100% capacity for years at a time.

I'd still much rather see "100 MW continuous" over "100 megawatt-hours of power each hour". And the former actually tells you a lot more than the latter, as with the latter it could be bursty (e.g. much more than 100 MW momentary draw), which would be a lot harder to supply. Fortunately for everyone the power draw is actually steady.

Here is a solar plant in California, which broke ground over a decade ago, that produces 550 MW of power, for $2.5B. Build even a small version of one of these and its 100 MW power needs are met.

https://en.m.wikipedia.org/wiki/Topaz_Solar_Farm

The 550 MW "nameplate capacity" is peak. The capacity factor of 26.6% means that it actually only produces about 146 MW on average. Nuclear power plants generally have a capacity factor above 90%.

146 MW on average is still more than this facility needs.

Not during the night and not when it is cloudy. Averages are vey bad numbers for something which is on 24/7.

Do the plants typically run at night? Also, doesn't the capacity factor account for these variations already?

No, you would need a large storage facility as well.

A nameplate capacity of 550 MW with a capacity factor of only 26% almost certainly means there are a lot of times when it isn't generating any power at all, or at least none to speak of.

Right, but if those align with the times where the plant is not used at all (night), does it matter? That being said, I have no idea what the normal operational schedules are for these types of facilities.

I'd wager a large sum that if you're investing $40 billion in a plant, you're going want it running 24/7/365.

And definitely not if the clouds suddenly blow in when the process is at a critical stage.

Electricity usage is much, much lower at night in Arizona because everyone turns AC off due to temperature drop. Plenty of baseload generators that can be better utilized at night with increased manufacturing.

Kinda. Depends on the season. In the summer, its still 80 degrees at night in Phoenix.

Nope...once all phases are built they will be using ~250MW of power.

Solar farms are one of the fastest electricity generators to build. AZ will have no problem stepping up production in the face of demand.

Your EIA source is for generation in the month of august only. Total solar generation in Arizona is much higher than that. Very roughy 1,500x nameplate capacity = total annual generation in Arizona. 5,600 mw capacity in 2021, 8,400,000 mwh annual generation.

https://en.wikipedia.org/wiki/Solar_power_in_Arizona

All fabs have battery backup regardless of power source, although with solar you would need a much larger one indeed. Either way, the scale of solar power generation easily supports these new fabs.

https://www.eia.gov/electricity/generatorcosts/ Cost is ~1,65B per GW of solar. 30% subsidy from government = ~1.1B per GW to generate ~ 1,500,000 mwh. They’re dirt cheap. Unfortunately battery storage, while it exists at this level, is more in the range of 800M to be safe (1600 mwh of storage). https://www.tesla.com/megapack/design

So all power infrastructure would be less than 5% of the cost of the plant, and that’s overbuilding solar needs by 4x (meaning even on the crappiest day power wise you’ll have more than enough) and having 16 hours of backup power which is also overkill.

There's also a nuclear plant nearby in AZ:

https://en.wikipedia.org/wiki/Palo_Verde_Nuclear_Generating_...

Nuclear + Solar seems like it would be a pretty ideal mix in a hot climate, especially given that peak demand for air conditioning coincides with solar maximum output, and Nuclear could handle some of the base/overnight load.

It is but you also need controllable components with the ability to quickly change output performance when building solar as your peak source. Usually this means gas turbines, in some cases this could also mean battery.

Do these plants not require tons of water?

from the wikipedia link - they use sewage/greywater for cooling:

> The power plant evaporates the water from the treated sewage from several nearby cities and towns to provide the cooling of the steam that it produces.

This treated sewage is the water that Vegas now provides for general use, and what AZ will need just for its people in about 5 years. Wasting it on chip production is going to go over slightly better than the Saudis wasting groundwater on alfalfa.

They could literally truck Evian water in if they wanted to. Water is extremely cheap compared to just about every substance on earth, and it’s hard to imagine that water costs can be a problem for this type of operation.

You're assuming that the severe undervaluing of natural resources in the US will simply continue. Try taking a field trip to a place (including the US), where people can't drink the water that comes out of their tap, and you'll get a better idea of where this leads. In no universe can TSMC use bottled water, much less Evian, and be competitive.

This may not be true today. But when I first arrived in US in AZ, I was told Phoenix boasts highest numbers of golf club concentration in entire US, despite high water requirements of a golf club and being in middle of desert.

I suspect the ongoing Western US drought would have worsened the water situation.

Arizona has the reputation of being a desert, but in reality it is a state that is half-desert and half-mountain, and has a tons of water. Enough to waste on ubiquitous golf courses and alfalfa farms. There is a canal system taking the water from the mountains to the drier south where people live.

But the population pressures are realigning who pays what for the water now, so I expect the alfalfa farms are going to go, and chip fabs will take their place.

The thing you don't mention is that Arizona is wasting other people's water -- nearly 40% is sourced from rivers originating in other mountainous states

https://www.arizonawaterfacts.com/water-your-facts

I don't mention it because it's not true. The idea that if a river originates in one state, then that state owns all the water from the river is not how ownership of river water works. And in America, rivers often form the border of states, so there are always competing claims about which state gets which percentage of the river. These are generally settled through treaties and agreements among states, and those agreements were made also for the Colorado river. Arizona is only taking its share of the water as per the treaty, not someone else's share.

And it's share of water was already insufficient a decade ago, which is why the water table has run dry in many rural areas, forcing people to truck in bottled water or pack up and leave.

Do you mean the Colorado River? The one that flows through the Grand Canyon? The canyon in Arizona?

Yes, the Colorado River whose headwaters comes from the state of Colorado.

That's kind of a ridiculous way to claim "ownership" of water. Especially water that naturally flows into the state.

"Has tons of water"? LOL It is 5 years from rationing. Vegas solved their problem 20 years ago, things in AZ have gotten 10x worse.

> Arizona has the reputation of being a desert, but in reality it is a state that is half-desert and half-mountain, and has a tons of water.

The mountainous areas are 'high deserts'. Most of Arizona's water is underground. It's being pumped out faster than it's replaced. Old wells go dry all the time, because newer wells are deeper.

"The Sonoran Desert's bi-seasonal rainfall pattern results in more plant species than any other desert in the world. The Sonoran Desert includes plant genera and species from the agave family, palm family, cactus family, legume family, and numerous others." - https://en.wikipedia.org/wiki/Sonoran_Desert#Flora

> There is a canal system taking the water from the mountains to the drier south where people live.

The Central Arizona Project pumps water straight east from Lake Havasu (738 ft elevation) to Phoenix (1086 ft elevation), then south to Tucson (2388 ft elevation).

https://en.wikipedia.org/wiki/Central_Arizona_Project / https://www.cap-az.com/water/cap-system/water-operations/sys...

The Salt River Project [SRP] buffers water from the rivers that run through the Phoenix area, and diverts the rivers into canals for delivery to the Phoenix area. Before the dams, the Salt and Verde rivers ran year-round. SRP modernized/expanded the Hohokam's ancient canals.

https://en.wikipedia.org/wiki/Salt_River_Project#History

> Do these plants not require tons of water?

If you’re referring to the Palo Verde Nuclear Generating Station, it’s downstream of water treatment plants and is effectively evaporating the water from the effluent stream [1], in other words it’s cooled by 100% recycled water. Probably a significant amount of the evaporated water returns to the local area.

[1] https://en.wikipedia.org/wiki/Palo_Verde_Nuclear_Generating_...

Seismic activity

Also, I would think that lots of water is needed, and Arizona is pretty dry.

> would make more sense to put the mega-fabs where electricity is cheapest in the Americas, and that certainly isn't Arizona

Geologically stable. And at least for business, politically stable. Same reason Walt Disney is frozen in Phoenix [1].

[1] https://www.alcor.org/

Ah, the lesser-known Pascal's Wager business plan.

The Disney thing is an urban legend, right?

> Disney thing is an urban legend, right?

I’ve heard the claim repeated by an Alcor employee. But they may have been repeating a myth. (Alcor definitely exists, and claims it chose Phoenix for the aforementioned reasons.)

Walt Disney was cremated. https://www.snopes.com/fact-check/suspended-animation/

It's right there on the death certificate: https://findadeath.com/wp-content/uploads/directory/w/Walt_D...

The legend I heard was that only his head was frozen. I believe Alcor does provide head freezing services and at cheaper cost than whole-body.

ASU has a good optical engineering program too, electricity is no the only variable.

This has got nothing to do with improved automation.

This isn't onshoring - onshoring would be moving to Taiwan - this is the Taiwanese Semiconductor Manufacturing Corporation.

This is onshoring - to the shores of a country whose national sporting league called The World Series.

There is a distinctly USA centric view of the world :/)

MLB isn't a national thing. There are two countries (US and Canada) who have teams.

Honestly I am not sure if that is sarcasm or not :-)

It is not sarcasm. How could something strictly be a US thing when both the US and Canada partake?

It’s onshoring from the perspective of an American company that contracts out to TSMC fabs, such as Apple, AMD, nVidia, Marvell, Qualcomm, or other major TSMC clients. Sony is also a TSMC client so if TSMC built fabs in Japan, Sony would be onshoring.

    Hsinchu, Taiwan, R.O.C., Nov. 9, 2021 - TSMC (TWSE: 2330, NYSE: TSM) and Sony Semiconductor Solutions Corporation (“SSS”) today jointly announced that TSMC will establish a subsidiary, Japan Advanced Semiconductor Manufacturing, Inc. (“JASM”), in Kumamoto, Japan to provide foundry service with initial technology of 22/28-nanometer processes to address strong global market demand for specialty technologies, with SSS participating as a minority shareholder.

Ref: https://pr.tsmc.com/english/news/2880

People are cheap in semico in compare to building/equipment

Semiconductors are probably one of the lightest products relative to their costs.

So as you alluded to, the long distance logistics of delivering the finished products, have a negligent impact on it's price.

Huh? This is re-shoring by political influence (probably pressuring Taiwan/TSMC) and financial incentives. Both TSMC and Samsung[1] are buildings Fabs in the US. Both Taiwan and South Korea are dependent on US military for their own existence. This is anti-competitiveness at its best.

[1]: https://www.anandtech.com/show/16483/samsung-in-the-usa-a-17...

This + the recent news that Apple is planning to diversify production out of China (1) is hard not to read into regarding the US' view of where US/China relations are heading . . . any geopolitical armchair experts have a view?

(1) https://www.wsj.com/articles/apple-china-factory-protests-fo...

Policies like this actually move pretty slowly. This is still a reaction to Ukraine in february, not planning for anything in particular, just the possibility of a fucked status quo.

Or maybe even a reaction to the pandemic. Which means a couple years ago at least.

I think not being able to procure n95 masks because mask production was all off-shored, particularly to China, was a real wake up call that supply chain security is integral to national security.

Ukraine was the moment that showed that: "No two countries that both have a McDonald's have ever fought a war against each other." is a false idea.

Hong Kong made it clear that the idea of increasing prosperity in China will not lead to it's liberalization, nor can China be relied upon to be bound by documents that are "historical documents that no longer have any practical significance."

> Hong Kong made it clear that the idea of increasing prosperity in China will not lead to it's liberalization, nor can China be relied upon to be bound by documents that are "historical documents that no longer have any practical significance."

This is a great point, it is underestimated I think in how what happened in Hong Kong impacted perceptions of China. Was the first clearcut example that once China was in they would not respect prior agreements (like around having 2 systems in Hong Kong) which led to a wake up call for what letting China have a big political role in your society would mean.

I read it as an interest exchange between US and China. US has the upper hand here so managed to grab some good stuffs fron TW while China probably got the permission to go in the future.

You don't need to be an expert to see the writing on the wall. The 2nd in power will always want to be first. Kingdoms are built and die on power transitions and China has appeared desperate to become the leading super power for as long as I can remember.

edit: Can ignore this comment, I see an earlier thread that discusses water with lots of replies. Sorry about that.

The geopolitical context for this is wild; and in some ways is Taiwan selling its crown jewels and sacrifying its strategic relevance for unclear returns.

https://asiatimes.com/2022/12/us-mulls-scorched-earth-strate...

US mulls scorched earth strategy for Taiwan

US strategy to blow up Taiwan’s semiconductor fabs to deter China might do more harm than good

The US is mulling disabling or destroying Taiwan’s semiconductor factories in the event of a Chinese invasion. This stark change raises questions about its capabilities and commitment to defend the island.

The "mulls scorched earth strategy" part is completely solid game theory.

Always make it clear for your opponent that your worst-case scenario will result in severe blowback if at all possible. Then you can a) try to ensure that scenario never happens and b) mitigate the outcome or response if it does.

But it's only solid game theory if one side doesn't have escalatory dominance.

For example, during the cold war, the US always had far, far fewer troops in Europe than the Soviet Union, and didn't have anywhere near the sea-lift capacity to transport the quantities of men and material necessary across the Atlantic to defend Europe. But we had nukes, and so we had plans to nuke Germany in case Warsaw pact forces invaded and made sure the Soviets understood those plans. Nuclear weapons nullified the escalatory dominance of the Soviet Union in a land war against NATO.

Now, the question is, who has escalatory dominance in Taiwan? It's clearly China. All the US can do is blow up the chip fabs, but China wants Taiwan for reasons that have nothing to do with the chip fabs, and would still want Taiwan even if it had no chip fabs. Even if it had no industry or population whatsoever. Whereas the US only wants Taiwan for the fabs, which China can destroy whenever it wants.

Unless we actually put nukes in Taiwan or commit to defending Taiwan with nukes, China will have escalatory dominance and not the U.S.

So that leads us to the second question -- what happens when you threaten to escalate but the other side has escalatory dominance? Then you end up making the situation worse, because you are in effect provoking the other side to begin a chain of events in which you are guaranteed to lose. That is not solid game theory.

I'll confess that I don't understand Taiwan's importance to China. They can put a certain amount of their resources into taking Taiwan -- which will absolutely not gain them access to TSMC's production capacity under any conditions, due to a web of Western dependencies that runs deeper than any rabbit hole -- or they can put the same resources into improving their own economic strength. Why wouldn't they choose the latter?

The reason I say I don't understand any of this is that the same reasoning should have applied to Russia. If they had spent a fraction of the effort and energy they've historically devoted to harming other countries on improving their own country instead, they would be way ahead of the game... yet they didn't, and don't show any indication of changing course.

Is the same irrational thinking present among China's leadership? If so -- and their pathological obsession with Taiwan certainly suggests that it is -- then things are really going to suck for everyone involved.

Really you need to understand your opponent's utility function before immediately leaping to making summary judgments that China is not acting in its own interest.

For example, knowing that mainland China and Taiwan are the remnants of a civil war, in which mainland China has wanted to absorb Taiwan long before it had any chip fabs, as China views itself as the nation of the Chinese people as a whole, with Taiwan a historical part of China. This is the Chinese view. You may not agree with this view, but that doesn't give you the right invent some fake view and ascribe it to the Chinese, one that is easier for you to understand. 600,000 Americans died in our civil war and great slaughter and economic destruction was committed in order to prevent a portion of the US from seceding. Was that irrational? Maybe to someone who can only understand the North Invading the South because it wants access to Southern Cotton, but that wasn't the payoff function.

This idea that Taiwan = place where chips are made is the equivalent of viewing the Confederacy as where cotton was grown. It is not the Chinese view. Chips wont be made in Taiwan in the future, and they weren't made there in the past. Taiwan is much more than this, it is a population of Chinese people living in a region that was viewed as a historical part of China, and thus China views Taiwan as a rogue province first, and a chip maker second. Just because you view Taiwan as a chip producer does not mean that China does, or that China even cares about the chip production nearly as much as the "rogue province" part.

Being able to step outside of your own values and understand someone else's values is table stakes for doing this game theory exercise.

Moreover, the assumption that the military development of China which is ostensibly done to become strong enough to conquer Taiwan (but which would probably happen anyway) somehow comes at the expense of Chinese "development" -- is just not how China (or anyone else) views development. Most people view development as both military and economic. China's rise necessarily includes military power, and it's not at all clear that their military investment is so high as to cause their overall rise to be slowed. Certainly the US is more than happy to spend a trillion on defense each year and invade some country ever few years, yet we attribute much of our development to the growth of our military-industrial complex.

> If they had spent a fraction of the effort and energy they've historically devoted to harming other countries on improving their own country instead, they would be way ahead of the game... yet they didn't, and don't show any indication of changing course.

Either indulge in propaganda that vilifies the behavior of your enemies and attributes to them irrationality, or understand your opponent well enough to be able to predict their behavior but not both.

So this explains the "not being able to understand the behavior of X" -- you have to tear away the delusions first and view the world through a clear eye, and then you will be able to understand.

For example, Russia isn't the one who invaded or attacked 125 countries in the last 30 years. Is America acting against its own interests by doing that? No, they may be acting against your interests, but they are not acting against their own interests. Each of those 125 attacks against foreign nations had some reason, and were consistent with the view of people who view themselves as basically decent and seeking to protect their own legitimate interests. The US did not do that to "hurt other countries". The moment you stop believing that, you will find yourself unable to understand the policy of the United States. America will become a riddle. But worse, when you engage with America, you will find that they behave unpredictably, and so your geopolitical strategy will fail. The exact same is true for Russia, or China, or Iran, or Israel. They are all basically decent people in their own moral framework trying to look out for their own political interests in a rational way. No better or worse than the United States.

To be tall enough for the ride that is geopolitics, you must assume good faith, a desire to meet legitimate interests, and basic decency in their own moral framework on the part of your political opponents. The moment you slip into caricatures and cartoonish villainy, then you become blind to others and lose understanding. You seal yourself off and are not able to navigate the treacherous waters of the world, which requires you to both know yourself and your opponent.

For example, Russia isn't the one who invaded or attacked 125 countries in the last 30 years.

Ah, whataboutism, then. My mistake, I'll ask somebody else.

Ban me if you want but this is the dumbest comment reply I have ever seen. Chatbots regularly display more intelligence than this. I feel bad for anybody who has to interact with you in real life.

Please don't respond to a bad comment by breaking the site guidelines yourself. That only makes everything worse.

https://news.ycombinator.com/newsguidelines.html

Please keep tedious tropes and canned arguments like "whataboutism" [1] out of HN comments. They're reflexive reactions [2] that aren't compatible with the curious conversation we want here.

[1] https://hn.algolia.com/?query=whataboutism%20by:dang&dateRan...

[2] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...

Honestly, what other response is appropriate for the post I replied to? (A rhetorical question to which I already know the answer: "No response at all.")

Is that person engaging in good faith, in your opinion? A giant wall of text with a Kolmogorov complexity of maybe six or seven words, explaining how whatever happens is all everyone's fault but Russia's and/or China's, and that none of us have the moral authority necessary to question the outcome?

Then maybe you should read about the history of China (and Taiwan). And how it happened that Taiwan seperated from China.

>> The US is mulling disabling or destroying Taiwan’s semiconductor factories in the event of a Chinese invasion

I think this completely bogus. Firstly, the geopolitical tension in the region isn't about to go off anytime soon -- Taiwan just elected pro-China KMT party last week.

Second, Taiwan is one of the largest investors in China. In fact, most of CCP's chip initiatives in China are spearheaded by Taiwan-expats (ie, ex-TSMC/UMC engineers) dreaming of striking rich in China. Take for instance Mong Sang Liang, a Berkeley PhD and former head of TSMC R&D, now co-CEO of SMIC in China.

Third, those fabs are both capex and opex intensive. These fabs are not self-healing perpetual machines -- those expensive equipments need constant babysitting by various foreign vendors just to ensure they are running 24x7. I doubt that Taiwanese would agree to such an extreme plan.

Don't these fabs need like an insane amount of water? Why would they build it of all places in arizona which as far as I know barely meets drinking water requirements (at least at some cities) and is basically mostly a desert?

I came here for this comment. Thank you for bringing this up.

Although Phoenix is one of America's most progressive cities with regard to minimizing per-capita water usage, the whole prospect of the city of Phoenix was ill-conceived from the beginning.

Of course that doesn't matter now -- the city is there. But I hope that TSMC knows what they're getting themselves in to.

An article from August 2021: Water shortages loom over future semiconductor fabs in Arizona https://www.theverge.com/22628925/water-semiconductor-shorta...

> Chipmakers are setting up shop in Arizona as drought worsens

> Major semiconductor manufacturers looking to expand in Arizona will likely be spared from water cuts induced by an unprecedented water shortage in the Southwest, at least for now. As part of the scramble to end a shortage of another kind — the global dearth in semiconductor chips — both Intel and TSMC plan to open new facilities in Arizona. But they’re setting up shop just as one of the worst droughts in decades grows worse across the Western US.

> ...

> While Intel recycles much of its water, more fabs will mean it will need to send even more water through its systems. The company says that Arizona has been “vital” to Intel’s operations for more than four decades. The state is already home to its first “mega-factory network” and its newest semiconductor fab. Intel used more than 5.2 billion gallons of water in Arizona in 2020 — roughly 20 percent of which was reclaimed water, according to its most recent corporate responsibility report.

> ...

> TSMC said in an email to The Verge that for now it doesn’t expect the water shortage to have “any impact” on its plan to build a new fab in Arizona, although it says it will “continue to monitor the water supply situation closely.”

I was also wondering about the added energy needed to keep the fab cool enough to work in during the punishing summer months in phoenix. Seems like an odd choice.

If only there was a way to turn energetic photons into electricity?

Thanks for sharing that. At the very least, Arizona has "water budget" in the form of severely curtailing water-intensive agriculture.

I am hopeful that the politicians there are smart enough to realize that semiconductor fabs are an industry whose water needs are worth prioritizing more than farming alfalfa in the desert.

The problem is that these agricultural water rights are old, and the folks who hold them are often disenfranchised. Using a ton of water in the desert to farm is their whole livlihood; we don't just pull people's generational careers out from under them any more, like we used to. Not without a chance at an alternative.

Farmers are selling their farmland to developers who then turn it into housing. Everyone wins: Residential water use is significantly lower than agricultural, the farmers get rich, and more housing is being built.

Should people live in an area with no water?

It feels like we have so many problems caused by people just not living in the locations in the USA where there is lots of rainfall and water. I have tons of water, my state just uses lakes filled by rainwater. We make electricity with it too.

https://commons.m.wikimedia.org/wiki/File:Average_precipitat...

I always find it interesting that people keep cramming into the western and specifically southwestern part of the USA. Like the original settlers just kept going that way due to the human natural urge to keep going further and hope things would be better over the horizon. I often wonder what America would be like if the western US was settled first. I imagine people would have noticed how awful and non-conducive to human life it is and kept on trucking to the promised land in the East.

I hear you, but I feel like this argument is fraught. We can't (shouldn't) really be telling people what they can/can't do.

We can restructure incentives, though. (e.g.: taxes). Perhaps this is just kicking the ethical quandary down the road.

A more appropriate term would be disincentive, in your example about taxes. You are more or less telling them what to do by hitting them in a wallet. Sin taxes and enriching government to change behavior the government wants is practically synonymous and not appropriate in 99% of cases.

The problem is less people living in areas with no water and more people growing water-intensive food in areas with no water. 80% of water usage in California, which has 40 million residents, is agricultural.

Should people live in an area that needs to burn fossil fuels for heating? Energy efficient heat pumps, which AC is, only work on a relatively narrow band of temperature differentials. You can't really use them in freezing conditions.

There's lots of hand wringing about living in hot, dry climates. It's just odd to me because where people live, in general, are not particularly amenable to the number of humans living there.

Heat pumps have come a long way. https://youtu.be/MFEHFsO-XSI gets into the energy efficiency of heat pumps and their temperature tolerances.

Cold can be mitigated by enhanced R-value insulation in a single application. Water use is continuous.

https://www.energystar.gov/campaign/seal_insulate/identify_p...

I don't really see a hot/cold stratification in this chart- https://www.statista.com/chart/12098/the-us-states-with-the-...

And even then, the difference in costs seems quite small. Alaska is $332 and Georgia is $310.

If you live on a 20' by 50' plot that gets six inches of rain per year, that's 500 cubic feet, or enough water for ten people to drink all year. And that's low rainfall even by Arizona standards (Phoenix gets 7 inches while Tucson and the plateaus are wetter), a small plot of land (1/43 acre), and ignoring any reuse or importation. The Southwest has plenty of water for people to live their lives, as long as it's utilized efficiently. It's agriculture that isn't suitable.

It's also worth recognizing that the western United States was settled first, as people came down the Pacific Coast after crossing the Beringia land bridge during the previous ice age. California in particular has always been quite densely populated, and the only (IIRC) indigenous irrigation systems were built by the Hohokam near the Colorado Plateau.

EDIT: https://repository.si.edu/bitstream/handle/10088/23978/SMC_8...

>The population [of California] cannot be tabulated by tribes, but there can be no question that it was several times larger than any other area north of Mexico and that the destruction has been correspondingly greater.

And food still shows up in the grocery stores. Yay!

These plants will be 100% recycling on water and Phoenix gives a lot of water to Tucson to store that they could pull back as needed. Central Arizona Project is a good read. reply

There have been people living where Phoenix sits today for at least 1,500 years. The canals in Phoenix were originally dug by the Hohokam 1,400+ years ago. Dutch settlers used what remained to design the canal system that's used today. Here's a cool article that describes some of what they built http://www.azheritagewaters.nau.edu/loc_hohokam.html

Some people said because Arizona is in relatively stable seismic region in the US. But I don't buy it. TSMC fab in Taiwan is also located in seismic prone region. Compared to the cost of the fab itself, seismic isolation system is relatively affordable.

They need seismic isolation regardless of where they build. Cutting edge fabs require such precision that a poorly shielded USB port can cause enough noise to create problems, let alone the vibrations from distant earthquakes. (Source: previous work with electron microscopes)

Fabs are expensive because every part of building it from the electrical wiring to the seismic isolation to the HVAC system needs to be perfectly tuned to remove all sources of noise. Most of these costs are superlinear if not exponential with the magnitude of noise.

I heard a story about an Intel fab in Arizona that would always produce bad silicon at a certain time a day. After some investigations it was determined that a train passed by at that time every day causing enough seismic activity to disrupt the manufacturing process.

Logistics risks around the plant are not nearly as easily done. There is an inherent benefit to not being exposed to seismic and weather issues. AZ isn’t a bad choice.

I was curious about this a while back. My cursory research showed that these facilities are actually phenomenal at recycling water to the point that they're not actually guzzling water at the rate you might expect.

...deputy spokesperson Nina Kao said via email that “approximately 65% of the water used in the Arizona fab will come from TSMC’s in-house water reclamation system...

https://www.azcentral.com/story/opinion/op-ed/joannaallhands...

Water can be piped around cheaply. Worst case they pump it in from nearby states

Arizona actually already does this, in profound quantities. https://en.wikipedia.org/wiki/Central_Arizona_Project

The problem isn't the pipes, obviously.

It's the fact that the 1922 agreement that divided up the water in the Colorado River (upon which Arizona depends) incorrectly calculated the amount of water available in the river every year.

The Colorado River Compact divides up 20 million acre-feet of water. Modern analyses show that the Colorado river's average flow is about 15.5 million acre feet, and has been less than 13 million acre-feet during the last decade or so. These reductions curtail who gets what water.

https://www.5280.com/how-the-100-year-old-colorado-river-com...

And the Colorado River is the only water source in North America right?

People have a fetishization with hyping up disappearing water sources as leading to mass forced desertions of population centers, which is not rooted in any kind of reality.

We can already pipe oil from texas to Canada with ease, we can do the same from Canada to Arizona with water. Or wherever else. The economics and viability of it is obvious and clear.

Residents of water scarce areas would pay marginally higher taxes to support this

> People have a fetishization with hyping up disappearing water sources as leading to mass forced desertions of population centers, which is not rooted in any kind of reality.

I have never heard this. Where are you reading that?

Disappearing water sources are a reality. Time and time again, long pipelines like what you suggest just aren't viable.

I'd suggest you enjoy some fact-based reporting on the topic: https://grist.org/agriculture/drought-water-pipeline-cost-we...

The near- and medium-term solutions we need are solely in use reduction. Period.

Oil pipelines have already proven the economic viability.

The choice of abandoning trillions of real estate vs a 10s of billions infrastructure project is obvious.

Article cites the cost as $14B which is effectively nothing vs the alternative. Even $100B is nothing. These two TSMC fabs are $40B alone

So much fearmongering from economically illiterate people. Use your brain

I hate to break it to you, but your "economically illiterate people" might actually understand some fundamental social and political realities that you've not yet had the chance to reckon with. Not to mention the physical realities of water resources.

Your comparison to oil pipelines is nonsense. Where to start: privatization, a global commodity market, relative scarcity, massive extractive capex.... Sure, we built oil pipelines easily in the past. How difficult has it become? Surely the capital exists. Why don't we build more?

Nobody is talking about abandoning real estate.

The tragedy of people like you is that you provide a half-compelling distraction. People (cities/farmers/industry) need to use less water.

The longer it takes to get people to realize this, the harder it is to fix.

Why would we practice conservation when we can spend $30-50B in capex amortized over 50 years via marginally higher taxes on the populace with minimal opex to solve the problem permanently?

Thats why they are economically illiterate. The obvious and more cost effective choice will win in the end, and it doesn’t involve austerity and suffering and repentance for high water use

Because it wouldn't solve the problem permanently.

The folks who wrote the Colorado River Compact thought they were solving the problem of dividing the river's water permanently, too. They made it worse.

I don't disagree that economics are a critical part of this conversation. But I don't understand how you're making the argument that enabling more water use, not less, is cost effective.

Even amortized across 50 years, $30B is more than the $0B required if we just use less water.

None of this even comes anywhere near discussing the potential deleterious ecological (and economic!) impacts of draining water-rich ecosystems by piping their water elsewhere. Look to the phase-out of leaded gasoline as a great example of how externalities matter.

I don't think you appreciate just how much water the northern part of the continent has.

Actually, I do. I've lived there.

Generally speaking, substantial alterations to hydrological regimes can cause deleterious cascading ecological effects. Just look toward the literature on the hydrological (and consequential ecological) effects of dams (necessary to divert water) for examples.

One of particular interest is dams' tendency to reduce the frequency and severity of flooding events, which are characteristically necessary for much of the functioning of floodplain ecosystems such as those you're describing.

Such reductions can wreak havoc on biodiversity in these regions.

I only meant that the region which has an abundance of water which largely goes to waste can sell their excess via pipelines to the lower 48 wherever there is a dire need.

water problems are actually energy problems. With cheap energy, you can clean all water from residential areas and you can pipe water in from where it is plentiful.

Water is not usually created/destroyed it is moved around and contaminated.

Over the years, I've learned that "fundamental social and political realities" are often salesmanship driven solely by agendas. There is the reality and there is fantasy. The reality is we walk over more water (1000x) than there is in all of our rivers and lakes.

Humans need water. We don't need to use less of it. We need more of it and there is plenty of it. For example, farms need water to grow food. Humans need food (and water) or we die. And water isn't going anywhere. It doesn't get destroyed. As you probably know, there is a water cycle. It falls out of the sky, bubbles up from the ground, and can be found by drilling for it, just like our most critical energy sources.

>>Oil pipelines have already proven the economic viability.

For oil. Water has far lower value density.

For water. Its proven we can build many thousand miles of pipeline for 10s of billions.

Which is vastly cheap enough and sufficient to solve this problem. Much cheaper than alternatives

I love how arm chair engineers have solved the water crises in the south by suggesting people "just build pipelines!".

Just the lack of understanding of like 15 different aspects of the problem with that idea and the confidence to defend regardless is kind of amazing.

You're right, piping water from point A to B is perhaps one of humanity's most challenging problems.

What's really sad is that so many people are brainless enough to believe that this is not possible to solve, and easily so, once incentives drive it forward.

Your comment does remind me of many of the 0.1-0.5x engineers I've worked with though. We need hundreds of people to maintain a mobile client, it's not possible to do it with a handful!

I'm sure those "0.1x" engineers found you a treat to work with. One of the more impressive combinations of abject ignorance and misplaced confidence I've seen in a long time. Believe it or not, but there are indeed people who've designed large volume water piping systems on Hackernews..

Where is the water coming from? It’s not coming from the Great Lakes or the Mississippi River. Where are these enormous volumes of untapped water that we can pump over the Rocky Mountains?

How about the Columbia? Average discharge is 265,000 cfs, which has been pretty consistent over the past 100+ years.[1] That's approximately 200 million acre feet just flowing into the Pacific every year.

For comparison, the Phoenix metro area uses 2.3 million acre feet of water per year (40% of which is used in agriculture).[2] So a bit over 1% of the Columbia's annual flow.

Another comparison: the Colorado "in its natural state" averaged just 22,500 cfs, less than 10% of the Columbia.[3]

I'm not saying that piping water from Oregon to the southwest is a good idea, but it's not as if we don't have water in the west. It's just a lot further north than most of the population centers.

[1] https://en.wikipedia.org/wiki/Columbia_River#Discharge [2] https://askabiologist.asu.edu/questions/how-much-water-are-w... [3] https://en.wikipedia.org/wiki/Colorado_River#Discharge

> We can already pipe oil from texas to Canada with ease, we can do the same from Canada to Arizona with water. Or wherever else. The economics and viability of it is obvious and clear.

Oil and gas are far more value dense than water. Pipeline transport adds somewhere around $5-20 per barrel of oil, or $0.03/L. Some quick googling suggests Phoenix residents pay around $0.00022/L. That's a factor of 142x.

If they pay $0.00022/L then its really not that scarce after all huh?

True scarcity would drive prices up and increase cost viability of transport, obviously. The reason its not done right now is because its not an imminent problem

it’s an essential for life and the consumer price reflects heavy subsidies

They do need a lot of water, initially, but it was pointed out that it's similar to filling a swimming pool [1]. Yes you need a lot of water at first, however it's not as though all that water is just being dumped outside to evaporate. It's cleaned and recycled, and recaptured from the air within the plant.

While Arizona doesn't have an abundance of excess water, they do have pretty strict usage guidelines and rules in place from the Department of Water Resources [2]. Assuming we can trust the government's call on this (are they being objective and fair, or tilting the reports to bring in Billions of dollars of corporate revenue and jobs) then these fabs should have no problem being supplied with water or impacting others in the area.

[1] https://www.cnbc.com/2021/06/04/why-intel-tsmc-are-building-...

[2] https://www.arizonawaterfacts.com/water-your-facts

This seems wrong. Didn't TSMC have to throttle down production recently because Taiwan had a drought? Why would they have had to do that if the plants don't need new water after the "initial fill"?

Apparently they've figured it out: TSMC tackles Taiwan drought with plant to reuse water for chips https://archive.ph/XIff9

Yes, exactly. In places like Taiwan TSMC just didn't need to account for recycling/treating the water - I'm sure it's cheaper not to do so. In Arizona it's a requirement so it's built into the cost of investment.

It depends on how many typhoons fill the reservoirs though. This year we only had a few.

Access to plentiful, cheap water was a major reason why Micron chose Upstate NY for their new fabs.

Yes, but the cost of ultrapure water is essentially represented by the cost of electricity used for purification.

Arizona paid them more in subsidies than the other states would

They need a lot of water initially but recycle it as they operate.

Intel and ASML already have facilities in Arizona. I'd assume TSMC chose partly based on that allowing them to hire out of the local talent pool, along with incentives from the state.

While water does seem like a serious issue, it also feels very Monday-morning quarterback: of course the largest semiconductor manufacturer in the world knows that Arizona has low rainfall. What I'm curious about is what makes Arizona preferable over, say, Idaho or Tennessee. Proximity to California? No freeze-thaw cycle? Solar energy?

Energy might be a big one. Solar is plentiful here in a state where cloudy days are rare, and there is also the Palo Verde nuclear power plant

These plants will be 100% recycling on water and Phoenix gives a lot of water to Tucson to store that they could pull back as needed. Central Arizona Project is a good read.

>Why would they build it of all places in arizona

for the Cynical person: Ensuring a future crisis is a sure fire way to get a nice Big Government Bailout in a few years to pay off the rest of the Capital costs that were not already covered by CHIPS....

I remember reading that making chips requires a lot of water, something that Arizona is running out of, will this be a problem?

If I was conspiratorial, I’d say this serves as a excuse for building more pipelines for getting more water from other states.

What is the longest and (large enough) volume water piping system in the world? From my recollection its not long enough to significantly make inroads to another state.

edit: I think I might be wrong, looks like China has done significant waterworks across their country.

Why isn't it the same as say an oil pipeline? Those can get really long.

Oil is significantly more valuable than water - the added cost per gallon of water would be astronomical.

1 oil barrel is 42 gallons and costs around $75 today. I pay about $6 per CCF of water which is 748 gallons. It might be worth doing this with oil, but even if you scaled it up, it would be hard not to have the transport cost more than the actual water.

It would be cheaper to just desalinate it.

Of course Arizona doesn't really have that option. But California does, and Arizona already sources water from the Colorado River, which California does too.

Anyways. I think moving California off the Colorado River would do a lot to ease the water problems in the southwest.

California has had trouble kick starting desalination. It is an energy intensive process, and the California electricity grid (especially in water poor SoCal) is still heavily dependent on fossil fuels (even some coal plants in other states feed into the LA metro). Furthermore, desalination intakes can negatively impact marine life, unless subsurface wells are used. Those are more expensive and companies pursuing desalination are not eager to foot the bill for them. Perhaps most importantly, desalination (even without subsurface wells) is more expensive than current sources of water in Southern California. I've not even mentioned the challenges in achieving safe exhaust of brine.

No, Arizona is not running out of water. https://www.arizonawaterfacts.com/water-your-facts

Is it plentiful enough that extracting highly-pure water is an economically-solved problem in Arizona?

Arizona is not running out of water. There's a lot of fearmongering surrounding water levels at Lake Mead or with the Colorado river, but Arizona doesn't get the majority of its water from those sources. It also uses 10x less water with a population of 7m as it did in 1950 with a population of 0.7m.

Arizona has a state-of-the-art water portfolio that uses lots of reclaimed water and is not reliant on micro or even macro-climate trends. A megadrought has impacted the area for the last few decades and only now are Arizona cities talking about the needs for conservation and possibly cuts in the next decade or so, but if climate models hold up the drought will be over in that time frame anyway.

Astounding amount of semi investments. $500B in this list alone https://semiengineering.com/where-all-the-semiconductor-inve...

I'm interested in the financial aspect of this. Assuming US Fabs are more expensive than in Taiwan that then must mean lower margins or higher prices.

I wonder if Apple for example has said that they are prepared to pay a premium to ensure diversity of supply?

The iPhone SoC was made by Samsung in Austin for many years, so this has already happened. At the time everyone assumed that Apple wanted to diversify away from Samsung for other reasons which were not necessarily financial.

I bet the Samsung SoC's were a lot cheaper though as less sophisticated. Samsung -> TSMC got them away from a competitor, possibly cheaper (US -> Taiwan) and onto better roadmap.

This is quite different and possibly a further sign of Apple spending to diversify supply chains.

It was probably a bit cheaper, but Apple was Samsung's anchor customer since mid 2000's. I don't think Apple used Samsung's 10nm or 7nm -- ie, there was no "dual sourcing" or "diversifying" supply chains to speak of. I think Apple's move to TSMC was largley driven by Apple's China/Taiwan first outsourcing practices.

> At the time everyone assumed that Apple wanted to diversify away from Samsung for other reasons which were not necessarily financial.

When Apple was dual sourcing chips from both Samsung and TSMC, the Samsung version of the A9 had a ~10% reduction in battery life compared to the TSMC version, although the performance was nearly identical otherwise.

https://www.tomshardware.com/news/iphone-6s-a9-samsung-vs-ts...

The article you linked actually reports the Samsung version lasted longer than the TSMC chip (the opposite of what you've stated).

Samsung was on a 14nm fab process.

TSMC was on a 16nm fab process.

My bad. I should just say that there were the same silicon lottery issues we see when we buy a chip from a single vendor, eg Intel today, compounded by the fact that the process nodes were different too.

The popular narrative at the time was that TSMC chips had the better battery life, but it depended on those silicon lottery results.

>What we know is that there isn't enough information currently out there to accurately determine whether the TSMC or Samsung A9 SoC has better power consumption, and more importantly just how large any difference might be. 1-on-1 comparisons under controlled conditions can provide us with some insight in to how the TSMC and Samsung A9s compare, but due to the natural variation in chip quality, it's possible to end up testing two atypical phones and never know it.

https://www.anandtech.com/show/9708/analyzing-apple-statemen...

However, when the chips fell as they may, Apple single sourced the A10 at TSMC, and they would have had the largest number of performance data points.

dual sourcing? I thought TSMC's 16nm was just transitional to wean off Samsung's US operation completely in favor of TSMC in Taiwan. Did Apple "dual-source" at Samsung/TSMC 10nm? or 7nm?

Apple doesn't care that much about cost of chips. Their margins are wide enough that even a substantial increase in chip pricing won't significantly impact their profits.

I don't really think this is true. Apple's margins are so wide because they care about the cost of all their components.

this is profoundly wrong; Apple and other hardware brands have not only practiced material source cost discipline to the extreme, but their internal culture promoting and strengthening that resulted in a nerdy work-a-holic with no life to become CEO of one of the largest companies in the world, specifically because he would, and did, put whole companies out of business with suffocating negotiating tactics over portions of a penny. This is exactly in the pattern of Dell Inc, where the saying about commercial supply partners was "those who do business with Dell, go out of business"

source- electronics recycling business in Silicon Valley for a few years

Apple is known for squeezing their suppliers to last penny.

My understanding is the that the two biggest financial advantages for building domestic fans are the CHIPS act subsidies and lucrative defense department contracts that are required to be sourced domestically.

Not unlikely that there’s a significant political aspect involved here too though.

Does anyone else wonder how the water issues that we are seeing become more pervasive in the south west will impact long term investment? I have family that lives in PHX so I'm familiar with the area and much of the development. Though I cannot help but conclude that at some point water shortages become a massive problem in that area. My naive understanding is fab plants recycle (re-use?) much of their water but ignoring that, fab plants need people and people need water.

My understanding is also that fabs can mostly re-use water. I don't know as much about AZ but in NV and CA most of the water is "wasted" on low-value agriculture and doesn't go to people.

I guess this is TSMCs getaway if things will get sour with China.

Remember when Warren Buffet invested into TSMC and someone said he must know something we don't?

TSMC has had a license to print money for a while now given how huge ASML’s backlog is.

Is ASML even going to be able to meet the demand created by all these fabs opening up across the US (and rest of the world)?

The stories I've heard about what it takes to build a single machine would suggest to me that scaling up would be a huge feat and I could only imagine take years to accomplish.

Then again, perhaps TSMC and their peers had predicted this kind of growth and already placed their orders long ago?

Maybe they will just move a chunk of existing fab?

ASML has now $32+B worth of backlog and they are still able to make only ~40+ EUV per year.

I keep wondering if this is different than when Foxconn announced a $10b investment in Wisconsin but carried through less than 10% of that?

Judging by the pictures, TSMC is building a lot of expensive-looking stuff out in Arizona. I know they're both Taiwanese companies but this project seems a bit different.

https://asia.nikkei.com/Business/Tech/Semiconductors/TSMC-to...

Buffet probably knows little or nothing about the fab industry and probably didn't make this investment decision.

TSMC is upping their investment in the US, but I agree with Morris Chang that the US operation won't be as lucrative -- ie, thye won't be able to maintain 50+% profit margin. Further, TSMC really didn't have any competition in the last 2-3 nodes and Samsung was largely absent in 7/5/4nm, but Samsung is going mass production with 3nm GAA pretty soon and TSMC is going to have to put up a good fight to maintain their dominance.

I think it makes it more likely that China is going to swallow up Taiwan without even firing a shot. Wargame simulations are giving a dim prospect for defending the island. US has removed fighter jets from Japan. And public opinion in Taiwan for resisting China has slightly shifted toward accomadation. I hope I am wrong.

>"I think it makes it more likely that China is going to swallow up Taiwan without even firing a shot."

I am not sure about "without even firing a shot" part but golden parachute type people and important staff will most likely be relocated to the US right before that and the rest of Taiwanese will have to suck it up. Yes very sad story and I hope it does not happen.

I think about half Taiwan exports go to the mainland. That is a signficant pull towards people favoring unification, if China really wants to play hardball. (And they will.) It all depends on the Taiwanese will to fight.

If China really thinks their completely untested military can easily pull off the largest amphibious assault in history, they deserve the results. It would be an absolute bloodbath. Wargame simulations are ways for the pentagon to get more money from congress… the reality is that invasions aren’t easy to hide and never go quite as planned.

If you disagree just go read up on the wargames and predictions for a Russian invasion of Ukrainian. Most people thought they’d reach Poland in days, if not hours. And they had the second most capable military in the world, on paper.

It’s also important to remember that a war game that your military wins isn’t a very useful war game. The whole goal is to find blind spots in military strategy and planning. “Nah we’re good everything is great” is how you end up getting worked by Ukraine.

> If China really thinks their completely untested military can easily pull off the largest amphibious assault in history, they deserve the results. It would be an absolute bloodbath.

For which side? No one knows.

There would be so many factors at play, both military and economic, that it is almost impossible to predict how things would play out. All anyone knows is that China's capabilities have increased many times over in the last 20 or so years, and that they are increasing every year. However, no country on Earth has fought a war on this scale for decades (with the possible exception of Russia and Ukraine right now), so no one knows what the outcome would be.

That's why I said "without firing a shot". Yes, a full-on invasion would most certainly be complicated. But the war excercises China is conducting are done for the purposes of intimidation, getting concessions before it even comes to that.

> US has removed fighter jets from Japan.

They’re phasing out older variants of the F-15 C. They will be replaced by F-15 EXs when they are ready, and until then the old F-15s will be replaced with F-22s

Well, that's good to know. I hope you're right. The story I read neglected to mention that information.

> And public opinion in Taiwan for resisting China has slightly shifted toward accomadation.

Huh? No it hasn’t. Taiwan doesn’t want to end up like Hong kong. They know the 1 country 2 systems doesn’t exist. Taiwan is already an independent country. They just want to keep status quo so China will leave them alone.

There is still widespread anti-unification support, but, as with everything, it's a bit more complex. Recent elections took a bit of the wind out of its sails. Chinese military excercises have taken somewhat of a toll. I know they don't want to end up like Hong Kong, but these public shifts matter if the people don't think they are being supported internationally.

https://www.bloomberg.com/news/articles/2022-11-27/taiwan-el...

Local elections don’t have any influence on national elections. Green always loses more to blue. The issue at national level isn’t anything to do with anti unification or anything like that. Alot of people here feel like real issues are not being dealt with. Like the cost of living continues to go up and isn’t addressed.

Well, I hope you're right. I'd hate to see Taiwan fall. It will be up to the Taiwanese on how hard they want to fight. Unfortunately, if they sense it's a losing proposition, many may end up throwing in the towel. That's what I fear.

In a sense.. Currently western liberal democracies have a monopoly on the most advanced microprocessor manufacturing tech.. And they plan to keep it that way indefinitely.

This "sours the milk" so to speak and it makes it easier to take TSMC off the board.

For one, if there was ever any real consideration by China of forceful takeover it would be made clear TSMC would not be a spoil of war. This makes a scorched earth threat even more credible.

For two, it disincentizes playing games with Taiwan as a pawn based on the wests reliance on chips flowing out. Knowing the chips are flowing in large quantities from other countries removes the lynchpin status.

Third, it stabilizes supply in the chance that something does happen. Maybe a natural disaster. Maybe a blockade.

It's interesting to see semi fab factories becoming a hot thing here in the states with an entire campus being rolled out in Ohio along with expansions and further growth in the southwest. Can only mean good things for workers as more options become available.

Are the 2 plants side-by-side? Or some distance apart?

Edit: I found another article that says they are at the same site.

Why Arizona why not NY state

Thats why Buffet bought that stock a week ago..

big investment eh? guess tsmc knows china is going to invade soon.

Good.

It’s good news geopolitically, and the construction will produce a lot of jobs. But people should temper long-term job expectations.

This huge of an investment in a tech where TSMC will face competition from fabs in lower-cost countries almost certainly means the company believes it can use extensive automation to avoid having labor costs sink their ability to compete on price.

(Yes, there may be a window where they don’t have to compete on price, but you build fabs for 5-10 years of production)

> It’s good news geopolitically

I actually think the opposite. It will be good for global supply chains but not geopolitics, because it reduces the reliance on Taiwan thus opening it up for assimilation.

That fab will not be even 1% of US semiconductor consumption.

And it will come to a halt within weeks if consumables will stop coming from Taiwan.

Nope. Chips clearly have a national security component. Ensuring local supply reinforces national security. Which reinforces geopolitical commitments through making those commitments more credible.

Edit-removed a comment

The geo in geopolitics means the whole planet, not just the USA. Destabilizing Taiwan could have major effects

> Nice try though.

I think you shouldn't use that type of wording on hacker news. It's clearly against the rules.

I think it actually increases assimilation with the US instead. One of the things that improved US relations with Japan when there was fear that they would take over the US was when they started building a bunch of US-based manufacturing.

That's describing something different. In the 80s the US was worried that Japan would dominate the US economically. The dynamic here is that Taiwan is worried that China will dominate Taiwan militarily. The assimilation being referenced above is not between the US and Taiwan, but between the Taiwan and China.

The US has no such fear about Taiwan. Taiwan has a fear that they will be invaded by China if they do not have industry that is too important to China to be interrupted by war.

Are you saying China is more likely to invade Taiwan if the destruction of TSMC’s Taiwanese fabs could be mitigated by buying the same products from TSMC’s US-based fabs?

I believe what they are saying that the US is less likely to risk war with China to protect Taiwan if they can get chips sources domestically.

I phrased that wrong. I meant that they want to be important to the world in general (and the US), not specifically China.

Even if it were bad for Taiwan, geopolitics is about more than Taiwan.

But I’m not sure this is bad for Taiwan: today, China can wipe out TSMC overnight. With these fabs there will be Taiwanese revenue streams coming from mainland US, which streams can support Taiwan government and military.

Walk me through why it’s bad for Taiwan to be more economically diversified?

Well for one the US won't be so inclined to defend TW so much. Sure, there will be posturing, maybe even a proxy war like in Ukraine, but generally speaking the US will not be sending troops anymore, now that the golden goose has nested on its shores.

I'm guessing Taiwan received quite a carrot in order to get their approval for the TSMC expansion.

No carrot I'm afraid :D I think this decision was 100% made by TSMC and the US. Both get what they want: - continuity of operations and safety for TSMC - locally sourced chips for the US and most importantly - much diminished importance for Taiwan, meaning they do not have to defend it with troops (which is the most important aspect of war - see Afghanistan, Vietnam, and so on).

Or conversely the US will commit even more resources to the defense of Taiwan now that their own domestic supply of chips is safe.

The US had no strategic interest in Ukraine yet we’ve all seen the reaction. The US will gladly defend Taiwan and create the hill for the CCP to die upon.

Now please explain why would the US send troops to defend TW if there's nothing actually valuable there anymore? Ukraine was hugely important for it's position next to Russia, and so is TW. And still, no US troops on the ground in Ukraine, and there won't be any in TW. This is a smart way of giving up on TW without actually admitting it, and there's nothing to be ashamed of, really.

They might do it because of the value in denying China access to Taiwan. Even assuming that Taiwan blows their semiconductor fabs to prevent China from getting them, there's still all the know-how and all the other stuff that the US wouldn't want China potentially getting its hands on, like American weapons. There would also be the consideration of the risk to Japan (and other friendly nations in the region) created by allowing China to just take Taiwan without at least Ukraine levels of backlash.

You seem to be asserting a few contradictory things:

1. Taiwan can only be defended with US troops on the ground

2. Ukraine has no US troops on the ground and is being successfully defended

3. The US would only put troops on the ground to protect valuable things like TSMC fabs

4. (by implication) if China attacks Taiwan and destroys TSMC fabs tomorrow, the US would have no incentive to defend Taiwan because the fabs are gone

I don't see it. Ukraine looks likely to survive as a country without US troops on the ground. Ukraine, as you note, has no fabs or similarly valuable assets. Why can't those exact same conditions play out in Taiwan, which is also much much much harder to invade with ground troops than Ukraine was?

1. The topic was US defense of TW. How exactly would it be defended by the US without troops? 2. Ukraine is defended by the Ukr army - with copious amounts of ammo and weaponry yes from NATO but no troops. At the same time the US never said they would send the army into conflict to defend it; when it comes to TW, despite the strategic ambiguity or whatever it's called, the general belief is there would be US boots on the ground. 3. The US would send its soldiers if there was something valuable obviously - be it fabs or oil or whatever, but not just to fight China for a random island in the Pacific. Otherwise they would have done it already on the countless atolls that are being fortified. 4. Yes, pretty much that. Just that it would not be obvious, but rather a long war of words would precede it... Again, nothing wrong with it. As long as no lives are lost I guess it is for the better.

The US isn’t giving up Taiwan- it’s the perfect opportunity to humiliate the CCP. China is in way over their heads if they actually believe they can achieve a naval landing on a fortified island without air superiority. It would be an absolute massacre.

Additionally, the US could simply embargo Chinese supply lines and dare them to engage.

I volunteer you to go fight in Taiwan to "humiliate the CCP".

Far beyond the value of TSMC, Taiwan is the lynchpin of the first island chain [1]. Taiwan remaining friendly to US interests is crucial to the integrity of the first island chain and thus the US island chain strategy[2]. In the event of conflict with China, control of the first island chain would allow the US to effectively interdict all maritime trade to China by denying access through the chokepoints created by the first island chain. This is frankly much more valuable to the US from a strategic perspective than TSMC (though TSMC is obviously very important).

You can also note the importance that Chinese planners place on this as well by looking at the Belt and Road Initiative, in particular the land-based projects that aim to connect China to European markets via rail [3][4].

1. https://en.wikipedia.org/wiki/First_island_chain

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

3. https://en.wikipedia.org/wiki/New_Eurasian_Land_Bridge

4. https://en.wikipedia.org/wiki/China%E2%80%93Central_Asia%E2%...

> geopolitics is about more than Taiwan

That’s obvious and wasn’t stated otherwise.

Absent any form of fab redundancy in the global economy, the fabs make Taiwan fragile in the eyes of the Western powers. Taiwan is safer when it is not a pressure point for the entire Western economy because it's easier to defend territory that can withstand a few hits.

It may be a useful talking point to get the public to do the right thing on Taiwan, but frankly there's no meaningful economic reliance on TSMC/Taiwan and policy makers know this.

> This huge of an investment in a tech where TSMC will face competition from fabs in lower-cost countries almost certainly means the company believes it can use extensive automation to avoid having labor costs sink their ability to compete on price.

What countries are you imagining that will suddenly become competitive in semiconductor manufacturing? As far as I know there are no up and coming companies of note. (Companies in China are not of note as they have been permanently prevented from further advancement.)

Yup, this isn't making the cheapest t-shirts possible. The cost of labor is nothing compared to the revenues. For a fab you want a stable government and environment, and good tax/trade situation, and reliable, quality labor.

"Permanently prevented" -- semi manufacturing is an engineering problem, not magic. Engineering problems can be solved, unless you think Chinese people are inherently inferior engineers.

It‘s not that easy to duplicate the institutional knowledge of companies like Zeiss, for example, otherwise they would have a lot of competition already.

Reply on news.ycombinator.com