These types of articles/pronouncement always tend to make a connection that is nearly never warranted. That is, the article quotes this tweet that says "The modern economy rests on a single road in Spruce Pine, North Carolina. The road runs to the two mines that is the sole supplier of the quartz required to make the crucibles needed to refine silicon wafers." But the first sentence doesn't really follow from the second.
That is, even if all the silicon wafers are made from this quartz from NC today, it's not really because that is the only possible option, it's because that quartz is the highest quality that can be procured at a relatively low price. But if that factory and quarry went away tomorrow, I have no doubt that the worldwide semiconductor industry would quickly adjust to a different quartz supplier.
I think it's still a very cool factoid, but the subtext that "the whole world economy 'hinges' on a teeny NC town" is false.
"...Mollick makes it clear in his social media thread that, yes, fully synthetic techniques are available to create similarly pure quartz. However, any sudden closure or interruption of the mining at Spruce Pine would likely cause "pretty catastrophic" disruption (and extra expense) for a few years as the industry scales up manufacturing."
I understand that 25% of all neon production was coming out of Mariupol, Ukraine (and another 25% out of Odessa) prior to the war. It would be interesting to see how the industry coped with a 50% loss of supply.
Neon supply disruption started earlier with the Donbas war, so companies were already adapting to the neon squeeze. Approaches used include developing neon production capacity outside of Ukraine, optimizing process control to reduce neon consumption, and recycling neon gas that would formerly have been vented to the atmosphere:
"Gigaphoton Announces Neon Gas Rescue Program" (2015)
Even then, I disagree with the characterization in that last sentence.
When billions (trillions?) of dollars are on the line, people and organizations tend to move with a degree of urgency not seen in the normal course of business.
Even with billions on the line, semiconductor shortages can take years to resolve.
Consider the 2020–2023 chip shortage, which hit automakers especially hard. When COVID19 struck in 2020, car sales plunged by 30%–90%, inducing many automakers to cancel future chip orders. Then when demand rebounded unexpectedly quickly, automakers found that there were not enough chips to buy, nor enough available manufacturing capacity to make them.
This shortage of chips was severe. Every major automaker (including Toyota,Volkswagen, Daimler, Ford, Honda, and GM) were forced to idle manufacturing plants. In 2021 alone, the chip shortage reduced auto production by ~11 million vehicles, costing an estimated $210B in auto revenue(!). In response, the White House convened 3 summits and signed the CHIPS and Science Act of 2022.
Building extra semiconductor manufacturing tools takes years. Even in late 2022, lead times of various tools were as high as 24-30 months, and lead times of some chips were still over 100 weeks.
Literally hundreds of billions of dollars of automotive revenue (tens of billions of profit) were sitting as a prize to anyone who could make some extra chips for automakers. And this is a relatively “easy” task - automotive is only a small ~10% slice of the semiconductor market, and doesn’t even need the leading edge technology nodes, where production is even more bottlenecked. Yet even with this massive prize of billions of dollars, the industry couldn’t match this demand in under 3 years. It’s not for lack of skill or effort — these companies have hundreds of thousands of employees, a large fraction with advanced degrees — but simply because the modern semiconductor supply chain is fantastically complex.
When designing a supply chain, cost trades off against resilience. If you minimize cost (e.g., single sourcing crucibles from the lowest cost provider), you lose resilience (e.g., it may take you months to recover from that single source going offline unexpectedly).
One easy way for the automakers to solve their problems would be to exploit Moore's law, and move all of their designs to a 28nm planar node of transistor production (which are somewhat more durable than the smaller FINFET nodes).
Because many of them were still at 180nm (or higher), they wasted the silicon material with a much larger die size.
Spend the money to stay current. Nobody is going to mass-produce vacuum tubes for you overnight.
I think the automotive chip shortage actually shows the opposite of what you think it does. The industry largely shut down, a global pandemic causing massive shipping issues for a huge range of components hitting a complex industry focused on just in time supply chains and complex logistics combined with poor forecasting of demand resulted in a 12% reduction in output over a year. Which then quickly caught back up.
That’s so wildly far from catastrophe it’s laughable, economic swings frequently have larger impacts on car sales. The pandemic hurt the automobile industry less than several others, companies didn’t fail, and things quickly caught back up.
So yes disruptions mean a temporary reduction in output and possibly a significant period catching up, but suggesting it would take an industry years to deal with X issue just doesn’t line up with how companies actually behave.
As I understand it, the "legacy" production capacity that the automobile manufacturers canceled/forfeited was (re)sold to other customers. When car sales picked up, the order books were full and there were were no immediately-available fabs.
The foundries did not appear to have as substantial of a downturn.
I agree that it wouldn't be catastrophic, but we just saw a semiconductor shortage during COVID. For a lot of us end-users, that mainly meant that we couldn't get a Raspberry Pi, but it did affect some industries where a lot of Micro-controllers (including simple 8-bit AVRs) and even stuff like ADC/DACs weren't available at all or only for way jacked up prices for a while. I think the automotive industry also suffered from that despite all their money because they couldn't just swing a magic wand to make a chip foundry appear in their backyard.
But in the end, everything was fine because as you say, eventually everyone is going to move.
We would probably never hear about it if for a completely made up example an extra 10k people died because some medical device was being made without some microprocessor that made it more effective or reliable.
It’s possible for things with large scale effects to have large impacts without it becoming well known by people not working in the affected industry or directly impacted.
As another example, I (and many other hardware engineers I know) spent 2021 doing essentially remediation on current designs to keep things in production (by designing in new parts). If there hadn't been shortages, we could have spent that time designing new products that will now either never be released or come to market a year later than they otherwise would have (preventing their use for that year).
It's a good point - I was more talking about society as a whole, since we didn't collapse. And we wouldn't collapse if some Quartz mind in North Carolina closes either.
But yes, on an individual level, it can be severe. Like the employees of car manufacturers who lost their jobs because they couldn't make cars for a while because they didn't have the chips needed.
Silicon wafers aren't made from that quartz; the crucibles in which silicon is melted is. The wafers themselves are made from metallurgical silicon that is subsequently converted to trichlorosilane, then repeatedly distilled to remove impurities, and then converted back to elemental silicon.
If we didn't have sufficiently pure quartz for the crucibles we could oxidize some of that purified silicon back into silicon dioxide and use that instead.
"A critical component in the design was a diode that would allow electrons to flow in one direction only, requiring the use of pure silicon. Mark's research at DuPont pinpointed a way to purify the element, and he himself grew the first crystal of hyper-pure silicon. This forged the way for DuPont to embark on a silicon production program that over the next 50 years led to the development of radar, silicon transistors, semiconductors and the Silicon Valley computer phenomenon."
> I think it's still a very cool factoid, but [...] is false.
Technically, that falsehood is what makes it a factoid, since it resembles-but-isn't a fact, just like how a spheroid is not a true sphere, a meteoroid/asteroid isn't a true meteor/star, an android/gynoid isn't a true man/woman, and almost every other -oid word out there.
Now, there are some philistines out there who claim that popular misuse means the "correct" definition of factoid has changed recently... But either way, I still find it amusing that everyone agrees that "A factoid is a small fact" is itself a factoid.
We don't have the vocabulary to separate two ways people use "critical": "required for something's functioning", and "not substitutable".
A tire is critical to a car's functioning, but there's plenty of brands to choose from if one tire company goes bust. Ofc, substitutability is a matter of degree - cost/time.
Any language can explain that difference easily. Even if all the article said was that this factory was "critical" to the semiconductor industry, I wouldn't have commented. Saying "the semiconductor industry hinges on a quartz factory..." or "The modern economy rests on a single road in Spruce Pine, North Carolina" is a very different type of deliberate hyperbole.
Those are true statements in the exact same way, though! "Hinges" and "rests" are used in the same way as "critical". Ofc, this requires some interpretive thought - and that may be a lot to ask of a general reader.
There is a similar fallacy that I've seen many engineers & teams fall into. The occupy "critical" roles which without, the whole chain would break and everything would fall apart
However often times the project only needs that one critical role and can replace it fairly easily. (the on-ramp off-ramp costs are lower than the cost of redundancy or other political costs)
Ironically, "non-critical" roles can be more important from a business narrative or core competency pov.
E.g. "the AI expert/team" leaving breaks the narrative of being an AI company even if all they do is non-critical path exploratory research. The inability to get new hires breaks the narrative of a growth-company etc...
There is more leverage in being able to direct narrative energy than in being nominally critical
> E.g. "the AI expert/team" leaving breaks the narrative of being an AI company even if all they do is non-critical path exploratory research.
I find myself in this position (but, not AI). That sole researcher doesn't just have "narrative" power. I make prototypes which are fleshed out into products by more capable engineers. Without me, the team is a bunch of engineers with great intuition on engineering and some familiarity with my niche -- but not the depth of knowledge required for keen intuition. The result would be either stagnation, or very inefficient research as an engineer attempts to step into the research role, hoping to accidentally rediscover decades of specialized knowledge.
An AI company without an AI just won't be able to execute at the pace of an AI company with that expert. And likewise, us niche experts don't go far without a solid engineering team to back us up.
There's a whole school of "Industry X is totally dependent on far flung facility Y" journalism. ("your cellphone depends on coltan from African wars" from a few years back is the worst offender - not that people shouldn't be concerned with such wars but concern should be based on a reasonable lens).
That said, the world has experienced quite a few supply chain disruptions in recent years and a big factor is large-scale, highly capitalized enterprises wind-up preferring situations of not producing at all to situations of increasing or continuing production at a lower profit rate for an emergency situation. (I remember covid vaccine production was limited by drug companies unwillingness to invest in new gene printing machines, US hospitals didn't buy US made masks because neither the mask manufacturer would make/accept short-term contracts at temporarily higher prices, etc. We all whine about AMD's unwillingness to compete directly with Nvidia with a Cuda-clone system).
Which is say, yes markets generally take risks into account but some amount of fragility seems to be a risk markets are willing to pay.
This comment reads like someone that has already forgotten about the supply chain nightmare of the past couple of years. Yeah, And, So to the idea that other suppliers can, could, might be available.
How long does it take to suddenly have those suppliers ramp up their production to the same levels? How long does it take for a new shipper to be found from that new location? How long does it take to make any changes to the processing because it's the same but different as far as what is coming in?
Switching to a new vendor when you're comparing shopping at Walmart, Target, Amazon is not any where close to the same thing as receiving raw material from a new vendor for an industrial process.
These comments about the Covid supply chain issues only further prove my point:
1. Most importantly, the pandemic was global. Pretty much every country was affected, so it's not like if you wanted to switch supply chains from one place to another that the other place didn't also have significant pandemic issues.
2. I never said their wouldn't be an adjustment, I'm just arguing that you always see this type of hyperbole like "The world economy hinges on one tiny town in North Carolina!!", as if this is the sole source of vibranium like in the Black Panther movie. Even at the height of the pandemic, with major supply disruptions, I'd hardly call a 12% decline in vehicle output or a Christmas without a PlayStation some sort of "catastrophe". If this factory disappeared tomorrow, yes, a ton of people in the semiconductor industry would scramble, and you might see some prices rise short term, but the average person would probably not even notice.
you're like a fence post. it does NOT matter if there are other providers of the mineral. if all users of the mineral are getting it from one supplier, there will be disruptions when that supplier is no longer available. it cannot be described any more simply than that.
you're monday morning quarterback analysis is just off. the industry has decided that this supplier's material is the best for them for reasons. therefore, the world economy <of this material> does hinge on it whether you want to call it hyperbolic or not.
> you're like a fence post. it does NOT matter if there are other providers of the mineral.
You are arguing that your frame of reference is correct. This is not an argument you can win, because it requires all parties to agree to that foremost. Also, don't call people names. It's against the rules cough guidelines - https://news.ycombinator.com/newsguidelines.html. Name calling doesn't strengthen your position...which you don't need to do anyway.
Peak oil is defined in terms of production. When total global production decreases, the moment of maximum production will be peak oil. Unless of course it's later exceeded, which could easily happen.
Why this will happen, when this will happen, and what the economic consequences will be, we can only speculate about. But it's the amount, not the price.
You are correct, of course. However, most of those using the "peak oil" scare tactic try to pretend that once "peak oil" is reached, there will be a sudden and drastic drop in oil production.
I think that's what the person you're responding to was alluding to.
> I think it's still a very cool factoid, but the subtext that "the whole world economy 'hinges' on a teeny NC town" is false.
Let them have their mistaken day in the sun, I've no doubt that the market would adjust pretty quickly rather than, you know, stop making semiconductors forever if this place disappeared off the map.
That is, even if all the silicon wafers are made from this quartz from NC today, it's not really because that is the only possible option, it's because that quartz is the highest quality that can be procured at a relatively low price. But if that factory and quarry went away tomorrow, I have no doubt that the worldwide semiconductor industry would quickly adjust to a different quartz supplier.
I think it's still a very cool factoid, but the subtext that "the whole world economy 'hinges' on a teeny NC town" is false.