I'm in charge of sourcing components for multiple companies right now. I believe they could double (100%) prices and companies would pay it. There's chips that I'm sourcing from obscure places in China just to get the part and paying whatever price they ask.
Designers are getting overwhelmed because they're having to redesign boards for components that are available, only to realize by the time they finish the changes, the part they redesigned for is also out of stock.
There's been some talk of a Chip-induced cold war, and I would say that's not crazy given the market conditions. I also love the guy who posts here all the time saying he works for the company that makes the machines that makes the chips and they can't make more machines because they don't have the chips that go in them. XD
Concur. We are currently paying a chip broker $6.30 for an Infineon chip we typically buy by the 100,000s for $0.47. Cuz without them, our customers cannot sell their $10,000,000 product.
24 volt digital output drivers.
We're also getting crushed on Xilinx FPGAs, certain Cypress CPUs, AD MUXes, Relay switches... It's something new every week.
Venor Vinge's The Children of the Sky is hard sci-fi touches on this issue of how to bootstrap a civilization to the point where the technology necessary to restore a highly advanced spaceship becomes available. Humanity will need this skill if we are ever able to venture out of our solar system.
I've played enough survival games to know that all you need to do is fashion a pickaxe out of stone and wood, use it to mine some metal, build a forge, and everything just snowballs from there
For instance, knowing that Corn doesn't supply absorbable niacin without nixtamalizaion, and diets dependent on corn tend to have large bouts of pellagra, is something that took hundreds of years to work out.
Walking into a corn fed post-apocalyptic future knowing that soaking your corn in ash water and hulling it prevents pellagra and improves the viability of your group is some handy knowledge.
This is also why rebuilding a web page after deleting it is so much faster the second time around. You've solved problems on the first go round that weren't apparent when you started.
> This is also why rebuilding a web page after deleting it is so much faster the second time around.
Counter-intuitively though, while a lot of people think rebuilding software from scratch will be a lot faster, in practice it's not because people tend to have to re-invent and re-discover a lot of things, and bolt a lot of other stuff on top as well. Plus over-engineering, plus analysis paralysis, etc.
So making a carbon copy of e.g. a website: sure, but for a lot of software it doesn't work like that.
That said, I do believe rebuilding software will work out better in the long run because learning from the mistakes of the past. It'll be interesting to see what e.g. Google's Fuchsia will turn out to be, the first big / sponsored OS made from scratch in a while.
And Firefox / Rust is another interesting one; it was a massive investment and risk, but I think it's paying off.
I have the same experience in general programming. You start doing something but then you run into a dead end. Then you look for other routes to your destination and find them. If you do it the second time you can avoid those dead-ends, which you didn't even know about when you started.
I think this means we should keep log-books like ship-captains do so we can avoid treacherous waters on subsequent trips.
It's a nightmare. A broker in HK recently quoted me $71/ea for an industrial temp rated 10/100 Ethernet PHY (KSZ8081MNXIA) ... That is a roughly 50x markup. We're making a design change instead. I am now buying qty 500-1000 of any new part that is getting designed in, before the design is even started.
Our Analog Devices sales rep told me to expect another 12-18 months of everything being completely fucked. That was based on a conversation he recently had with TSMC. It's going to be super fun.
Oh yeah, I've told my manager it's only a matter of time until we get burned. So far we have had decent luck with Heisener and Win-Source but prices keep climbing.
Well.. we were just looking at Heisener for a Cypress chip, but can't decide whether to pull the trigger. Have you placed an order and received the product from them?
Yes, we successfully purchased a small quantity of Kinetis microcontrollers which were supposedly tested prior to shipping, those are going into assemblies right now so we'll see if they actually work.
Heisener is just a middleman so every transaction is different unfortunately.
> by the time they finish the changes, the part they redesigned for is also out of stock
One of the things my last employer did to get around availability problems was to buy a vendor's entire supply of a chip we could use and then redesigning the product to use it. IIRC, the supply was projected to be enough to last a year. Hopefully, a year from now it will have proven to be a good decision :-)
Similarly, my company just redesigned a product after running out of an IC that's been on order for many months, still without an estimate on delivery by the distributor.
We redesigned it to use another chip we had in-house, intended for another product, but that product sells for much less and is in less demand, so...
We initially thought we could move pins around in firmware, and the parts have the same footprint, so it seemed possible, but they were different enough that we had to rev the PCB. Now we're waiting for PCBs and crossing our fingers that they will test out.
Um, I just did the same thing the GP did (buy up all the stock) for a product in development but one of our older products is getting hit by the same issue described by the parent!
Bosch recently opened their own chip fab in Dresden, primary focus for now is on automotive chips and those required for their in-house tools and appliances:
This makes a lot of sense as Bosch make so many automotive parts. In Europe, it doesn't matter whether you buy a BMW, a Peugeot or a Volvo, under the hood they are all using Bosch electronics to control the engine and many other parts of the car.
"service economy" was always a meme being sold to the gullible masses, manufacturing capacity is the only thing that matters for an economy. Manufacturing has always printed the money
It amazes me how the US allowed itself to be nearly deindustrialized after manufacturing effectively won them WWII and global dominance. Now the US can't even make ventilators or face masks
US Manufacturing is still going. In fact, we manufacture more than ever. Folks conflate manufacturing with manufacturing _jobs_. Our manufacturing is more efficient then ever which is why you constantly hear this false refrain that US manufacturing is down or gone.
It seems to me that if you can make a product of a given quality very cheaply, that probably indicates that you're very good at making it.
The fact that I can't get competitive production quotes in the USA for most products, like electric motors, injection molded plastic parts, printed circuit boards, stamped metal parts, castings, etc, suggests that the US is also not particularly efficient at making those parts. There was a time when the cheapest place to go for such things in production volumes was actually the US.
The price quotes you are getting also include the opportunity costs of saying no to other people wanting to pay for that manufacturing capacity.
The people I know that are running manufacturing aren’t exactly hurting for business. Now is just not the time to lock in high quantities of low-cost stuff.
This article is about TSMC raising prices for what little capacity they have unallocated. We can all agree that they are probably the best in the world at what they do, yet there is no way you will get “competitive production quotes” from them for commodity chips. Jim Bob’s Plastic and Stuff out in Fort Wayne is no different. Now, Jim Bob is probably going to cut you a deal way sooner than TSMC is, but I know of a certain mega-big car company that shut down an engine plant for a week because Jim Bob said last years prices aren’t this years prices - and you probably aren’t as big as mega-big car company.
people always point to this dumb metric, "we're producing more than ever" compared to ourselves, while ignoring that the number in comparison to the rest of the world has dropped massively. Our share of global manufacturing is tiny compared to what it used to be
I've worked for manufacturing companies my entire 30+ year career. It's not a dumb metric. It's usually pulled out in response to the silly complaint that "the US doesn't manufacture anything anymore."
Sure, our share of global manufacturing is smaller, but that's hardly surprising when many more countries up their manufacturing game. A more interesting metric would be US percentage of total worldwide manufacturing as a function of sophistication: I'd rather be building complex medical instruments than toasters.
Most toasters suck. I have to buy a new one every other year or so, for some stupid electrical or mechanical reason. Right now, for instance, the latch on our current toaster -- the one that latches on to the handle you press down to lower the bread and close the circuit to start the heating elements -- isn't latching any more. Fucking low-grade crap! I mean, I'm not asking for a self-lowering magically-correct-degree-of-toasting toaster like in the old TC video -- I just want a stupid mechanical latch in a product I bought a couple of years ago to keep stupidly mechanically latching for six or ten more years. That is not too much to ask.
So, stick to your complex medical instruments. Or, if you switch to toasters, please don't build the same shitty crap as everyone else.
>I'd rather be building complex medical instruments than toasters.
Why? In what way is it better to have a handful of elite workers creating a handful of units sold to a handful of consumers instead of having many ordinary workers producing many units for many consumers?
What happens? Well I suppose the economy collapses and the proletariat rise up and usurp power, or something fantastical.
What’ll more likely happen is we’ll further automate repetitive low skill tasks, and / or start making more durable goods, or otherwise ingenuity our way forward.
Post-industrialized countries make it impossible to for those countries arriving late to the party to advance any further, by way of global environmental agreements and glass bead "carbon credit" payments, thereby ensuring a permanent pool of slave labor.
Eventually it becomes more profitable to automate. See: Sony's automated PS4/5 factory. Formerly these would've been assembled in some Foxconn plant by tens of thousands of low wage workers.
I have a sneaking suspicion that's fundamentally impossible, as some of the prerequisites for developed economies (in the modern, not-industrialization sense) seem to require predatory exploitation of developing economies. And the global inequality of capital ensures there's a power disparity sufficient to enforce it for the former, to the detriment of the latter.
But I guess it's really a question about post-scarcity economics, and whether they're possible?
Because there won't be "many ordinary workers" producing those toasters. Low margin products like that will have their manufacture automated to the n'th degree and they'll be built by "a handful of elite workers" supervising the automation.
Or, you know, they'll send it to the lowest-cost labor supplier they can find to build by hand and that won't be in the US!
That's because the rest of the world is no longer a bombed-out postwar wasteland or its near-enslaved colonies as it was in 01945. "Producing more than ever" is what you needed last year when you couldn't get N95 masks or ventilators, not "producing more than the rest of the world". But something went wrong and the masks and ventilators were not forthcoming, something that looks an awful lot like an absolute decline in manufacturing abilities.
Other people said this above, but I will add my own anecdote I heard from a factory owner, that he had the ability to retool immediately back then to make medical masks, but between the uncertainty around whether people would keep buying them and at what cost, he was unsure it was worth the risk of switching away from a profitable business to something that would take less ability to make and less certain profits. I don’t know what he ended up choosing to do. But sounds like ‘not forthcoming’ is also possibly an indicator that skills, and the demand for them, is high—and not in decline.
> Our share of global manufacturing is tiny compared to what it used to be
You mean compared to a time when most developed countries' production capacity was significantly limited because they had been bombed out in WW2 and developing countries were a long way away from being competitive?
tl;dr if you focus on Intel's monopoly money-printer years, multiply it by a hedonic adjustment due to Moore's law, and mix it in with the rest of manufacturing it's enough to hide de-industrialization in your statistics.
> But it turns out that Trump’s story of US manufacturing decline was much closer to being right than the story of technological progress being spun in Washington, New York, and Cambridge.
Claiming Trump was really right, is usually sign of a writer trying to get people to read a bunch of opinions that don't match reality presented as "bold independent" stance
> Worse than the Great Depression: America’s manufacturing jobs implosion
So they're straight up going to max bullshit mode?
And they say if we ignore one of the most productive manufacturing sectors (computing) the data looks worse, Is that at all surprising?
US is smart enough to retain the cream of the manufacturers. I think some smaller western countries are really in the danger of a complete de-industrialization.
It's not for lack of trying. It's hard to compete on the global stage and it takes more than wishful thinking to maintain a competitive edge in a world full of people trying to top you.
Not really, it's usually the least profitable part of the supply chain.
That American businesses would gladly hand over their profits to poorer countries is irrational at face value.
TSMC is more than a manufacturer, they are a provider of very sophisticated in-demand tech for which the manufacturing process itself - is high tech, high stakes, high barriers to entry.
They are a little bit like Boeing in that the 'making' part is really hard. Less so like Flextronics.
Depending on how countries react to this, we could very well be in the opposite situation in 3 years, where we have a vast glut of chips and parts of every kind.
TSMC is not going to charge their customers 'prevailing market value' i.e. they could spot price at 5x their current prices, because it'd be gauging and the supply change would react long term to that 'bad behaviour' by TSMC. Big supply chain relationships are longer term so those things matter.
Because of the huge players involved with enough cash do influence (i.e. Auto, Apple, Samgsung) maybe we're in for a strategic alignment beyond just a fab shop in Texas.
One big caveat to all of this is that though manufacturing itself is low profit, a lot of things around that can become profitable. Making moulds, adjusting designs, managing supply chains, optimizing designs etc. - all of this becomes 'know how' that moves over to the manufacturing side and if they are smart about it, they'll leverage it.
Also, manufacturers can possibly develop some market power which they can try to flex as well.
Hopefully we see a shift to more high quality intelligent manufacturing in the Western world by leveraging tech, robots etc. and convincing consumers to pay higher prices for some things.
Service economy doesn't imply deindustrialization. The provision of a service can be highly capital-intensive.
>>Now the US can't even make ventilators or face masks
The US led the world in the speed of its vaccination rollout. As for manufacturing, its world-leading biopharmaceutical infrastructure allowed it to produce the most shots of the most effective covid19 vaccines, of any country in the world.
"Manufacturing economy" was a meme sold to gullible nostalgic masses
Work is work, money is money. There's nothing fundamentally more important about manufacturing then other types of jobs. The US still has tons of manufacturing but unsurprisingly its on the backs of much fewer jobs, many of which require more education and specialization then manufacturing economy of old used to supply
> There's nothing fundamentally more important about manufacturing then other types of jobs.
the idea that manufacturing is the only work that matters is especially bizarre on a social media website for tech workers.
microsoft or oracle or any number of other software products absolutely deliver value to customers just as much as a machine stamping out parts, those companies couldn't work without that software just like they can't work without an assembly line or a delivery truck or whatever.
Luckily you can buy computers from other countries
Same with food.
It's not anything special.
A way to earn a living/have an economy as a country is as good as any other.
Moldex masks are made in the US. Their unique construction also protects them somewhat from counterfeits since the 3M masks are the low hanging fruit everyone is copying.
Capitalism contains the seeds of its own destruction. The relentless drive to shave fractions of a cent off costs in the short term creates the conditions you describe in the long term. In other words, it's like being penny wise and pound foolish on a global scale.
Fake News... that is not capitalism, that is a by product of decades and decades of tex based incentives that push people in to market index funds, mutual funds and other institutional investment where it is all managed funds
These managed funds are driven by quarterly earnings people expect to see...
This intern drives public companies to focus on quarterly earnings
Privately held companies almost never look at quarterly earnings, they look at a 5 year picture... It is mainly public companies majority owned institutional investors (401k's, pensions, etc) that focus on quarterly earnings
So once again what people blame capitalism for, is in reality government putting their hands on the scales and tipping the market in ways that have unintended consequences
> "service economy" was always a meme being sold to the gullible masses, manufacturing capacity is the only thing that matters for an economy.
Service economies work great in places where food and housing are free, so money is only needed to buy luxury goods. They only work poorly in the US because capitalists have leverage over service workers.
Service economy is kinda fun when you really think about it. Rather few sectors really make sense. Food, you can't make living just by selling food to each other, the ingredients and hardware has to come from somewhere. Hospitality quite similar. Banking? Banking and finance is just extracting money from others, and actual goods produced is quite minimal... Same can be said from quite many other field like law. Some of the work does increase productivity, but you have to have some base line of it to increase...
Because all of their competitors are also raising their prices. Why would companies reduce prices and profits when the demand exists even with higher costs?
If you look at prices for the stuff a lot of wafer capacity is going towards, particularly Flash and RAM, you'll see that the prices have gone up 20%+ over the last year.
I think people like to pick out TSMC because of the politics surrounding Taiwan and China, and maybe because a lot of companies use TSMC for their IC production and like to complain. People don't notice as much when other ICs like flash, ram, micros, etc are slowly increased in price.
People like to pick out TSMC because they're the winner of the foundry wars; GlobalFoundries bowed out of further process node development, I think two years ago, and Intel is trying and failing to equal TSMC's performance with their in-house fabs. GF and Samsung and other players are still around, but only TSMC offers the highest-resolution parts, so they're your only option if your part needs current fabrication technology, and they may be your cheapest option even if it doesn't.
There are three companies (TSMC, Samsung and Intel) that compete at the top end of chip fabrication in the whole world. One of them (Intel) didn't make chips for anyone but themselves until about 4 months ago. Preparing a semiconductor for a given companies fab process takes time. You cant just call them up and ask for however many units of your design. Its a months to year long process of your engineers working with their engineers to optimize your circuit design to work best with their methods.
Also there is only one supplier (ASML) in the entire world who makes a critical machine (extreme ultraviolet laser photolithography machine) needed for these cutting edge processes. It is unable to match the increased demand for their tools. And even if they were these super advanced might as well be magic machines have a lead time in years themselves as do the factories they go in to.
If you want to start marking modern semiconductors or increase your existing capacity to do so it isn't as simple as throwing a couple million at it and having it running in a few months. Its billions of dollars in a massive sterile factory using machines made by a handful of suppliers taking up acres of land and consuming impressive amounts of electricity and water to do so. Semiconductor fabs are massive sprawling industrial campuses https://images.anandtech.com/doci/13749/intel_kiryat_678_678...
Many missing chips for automotive or devices are not gpu or top of the line level but simple microcontrollers. Those can be done with 130nm to 40nm process for the most modern. 130 nm has been used for pIII or amd athlon64 and dates from 2001. Those need near UV, not extreme UV.
> One of them (Intel) didn't make chips for anyone but themselves until about 4 months ago.
actually iirc they've tried to get into the custom foundry business a couple times now (see: "Intel Custom Foundry" that got discontinued a couple years ago) but they never really attracted any major customers.
Part of this was due to the fact that Intel's design rules (the "rules" that govern chip design: how small wires and transistor features can be and so on) are completely different from every other foundry, so while porting from TSMC to Samsung or something like that certainly isn't easy, it's at least straightforward, whereas Intel's stack is basically totally different from everyone else's. One of the advantages Intel has is that the chip design and the process are tightly intertwined, if there is a problem with the chip design the designers can go down the hall and talk to the process people and see if they can come up with a solution at the process level. This obviously also became their downfall when 10nm slipped, and likely also it severely reduced their ability to sell their foundry services to other customers.
According to that article they were apparently also rather secretive about their process (seeing it as part of their "special sauce" I guess) and I've heard many customers doubted Intel's commitment to the program as well as their neutrality. And I think the final piece of the puzzle was probably price - I'm guessing Intel was asking a royal sum compared to GF or Samsung or TSMC. After all... they're intel, and their process is the best!
> One of the funniest stories I heard was about the first copy of the Intel 14nm design rules Atera got from Intel. They were heavily redacted, which is something I had never seen in the foundry business. After many delays Intel put their own implementation team on the first 14nm Altera tapeout and the result was a very competitive FPGA chip. If not for the continued delays, Xilinx would have been in serious trouble as the Intel 14nm FPGA, based on my experience with customers, beats the Xilinx 16nm in both density and performance.
They are now trying to get back into the custom foundry business of course. Different management, and a very different business situation for Intel - if they can get other customers to help finance their node development that's a big win for them at this point.
ASML is trying to increase the world's capacity by ramping up their production of EUV machines to 50 per year. [1] The ASML CEO stated it's probably not possible for the world to produce more than that. These machines are ~$150 million ea and building one is insanely complicated. [2]
Their max is about double that (other kinds of systems, and service.) That’s about a p/e of 60 which is too high but what this market has done to many stocks. There is some expectation of growth, too, of course. ASML will keep pushing their capabilities and if trends continue, they should eventually be selling individual machines worth an order of magnitude more.
What he's making is basically caveman technology compared to actual cutting-edge nodes.
No offense to his accomplishments, but the modern fabs like TSMC are so advanced you cannot just make the things they make without billions of capital investment. It just is not possible.
Unfortunately, many of the chips in short supply are on pretty small nodes, so you need extremely specialized equipment limited to a few of the big players to make them.
Certainly there's some stuff that can be made in larger nodes, but if you want something in 5nm you basically have one game in town.
Building a new factory is a 10+ year effort for a competitor. (They have to do the R&D to get on-par with performance, and physically build the factory).
The shortage won't last that long.
But in the meantime, anyone who owns a suitable factory can pretty much set the prices as high as they like. I'm really surprised the price hike is only 20%, rather than 10% per month for the next 3 years.
Which government? Taiwan, where it matters? This is a major part of their economy; they all win big. I bet they'd sooner surrender to China.
Some other government? This wouldn't have much impact on chips, except to make it a lot less likely anyone makes chips in that country ever again — which isn't to say no nation would do it; economic history is full of this sort of thing, particularly as nations start going the way of Venezuela.
That just disrupts the profit incentive of competitors.
The entity that manufactures the product is compensated for it, meaning the money isn't being siphoned off by a random third party. Price controls grant a wind fall to consumers who are in no position to build another semiconductor plant. It's basically a wealth transfer away from the semiconductor industry into the consuming industries. If you want semiconductor prices to drop then either grant a subsidy or let the government build a plant that produces way more than the market really needs.
A single shortage plus price increase might not be able to pay for a whole plant, but any plans to expand capacity will be shifted forward rather than be delayed. If the profit motive exists the construction of the new plant will happen one year sooner and it will be bigger in anticipation of future shortages.
Probably the limiting factor is one needs to make sure the customers of chips stay viable. I mean, one hopes the shortage will eventually be over and foundries will want to make sure they didn't gut their own market.
From my experience no one cares if they are as long as they more or less work.
Source: I'm a software engineer who has had to fix so many issues due to unannounced component changes from suppliers.
"Oh you just replaced the specific USB chip with something different that has a whole lot of errata (Bugs in chip logic that you have to work around in software) and didn't tell us? GREAT! THANKS!"
Oh man. I just did a board bring-up recently and didn't notice that the flash chip I was having trouble with was a previous version than the one we specified. I only just barely managed to get those boards working, and best as I can tell, it was the conductivity of the particular flux I was using to reflow the part that was nudging an NC connection to actually use it's internal pull-up. When the factory couldn't follow my steps, they finally realized that if they were going to substitute that chip, they'd need to add a 10k pullup resistor on a pin that had a use on that revision but not on the one we wanted. What's worse, is even if I'd noticed that the chip was older, looking at the datasheet for it that I had access to, nowhere did it mention a need for a pullup on that pin. Only the factory datasheet mentioned issues with the internal pullup and suggested using an external one.
Designers are getting overwhelmed because they're having to redesign boards for components that are available, only to realize by the time they finish the changes, the part they redesigned for is also out of stock.
There's been some talk of a Chip-induced cold war, and I would say that's not crazy given the market conditions. I also love the guy who posts here all the time saying he works for the company that makes the machines that makes the chips and they can't make more machines because they don't have the chips that go in them. XD