ASML CEO: Industrial conglomerate buying washing machines to rip out semiconductors A large industrial conglomerate is being forced to take increasingly desperate measures to satisfy their needs for chips, according to ASML, a manufacturer of chip-making equipment. Talking on an earnings call this week with financial analysts, Peter Wennick, CEO at ASML, said that "width of demand" for chips was " …
...that discretionary spend will go on travel instead of phones, computers and white goods.
People have been conditioned over the last 2 years that home = good, people and outside = bad. As such they will stay at home, work from home and play on their PCs and phones rather than travel.
> People have been conditioned over the last 2 years that home = good, people and outside = bad.
Bicycle sales shattered records during the pandemic. Outside was highly promoted as the place to go. Crowded, indoor public spaces are a very different matter. That means people are less inclined to take the bus/train/plane, and instead to drive, alone. Many people are doing such travelling, and it has been driving up fuel prices as a result. They were at record highs even before the Russian invasion of Ukraine.
This post has been deleted by its author
This post has been deleted by its author
This post has been deleted by its author
It's a good story, but why doesn't the conglomerate just offer a barrow load of cash to the washing machine vendor for their chip inventory. The washing machine vendor's pipeline is no worse affected than by having their retail stock hoovered up, and the conglomerate is not left with a pile of shiny crap metal.
Other options include:
* said conglomerate needed these chips ASAP and couldn't afford to wait for a supply agreement to be thrashed out.
* said conglomerate didn't want to get into a commercial agreement with the washing machine manufacturer due to being competitors or having some other business relationship issue.
* said conglomerate may very well be overall a large company, but the specific product for which these parts are needed is something they only build in small quantities, so it's somewhat less newsworthy than the anecdote suggests.
why doesn't the conglomerate just offer a barrow load of cash to the washing machine vendor for their chip inventory
In the medium and long term, the washing machine manufacturer wants to manufacturer washing machines. They can't do that without their chips, they'll either have to diversify into a new market (expensive, slow) or go out of business.
My local washing machine repair man used to collect PIC chips for me. The older machines had them socketed in as I guess they were learning to program them and needed to upgrade them during production as the worked out bugs and stuff. Now you can get one for love nor money!
If they can get the commercial chips to last until semiconductor shortage ends (maybe in a year or two), that may be considered a win, even if they have to replace all of those chips during a future maintainance or something.
Especially if these are some costly industrial gear, for which there should be fat margins.
So they're installing salvaged commercial grade components in industrial specification equipment? I predict quite a lot of failures.
Typically the only difference between industrial and commercial spec is the temperature rating. For many devices that is a complete non-issue. If so why would there be a issue?
This proves it then, washing machines are significantly more complicated than they need to be if the chips harvested can be used in EUV machines, but then I suppose the washing machine maker NEEDS to get its jollies off knowing how often you wash your underwear and selling that to the maker of those japanese vending machines.
Can you print out all the various predictions that company makes, preserving the paper as well as possible, so you can pull it out of archives in whatever time-frame they predicted, and check how far off/on the mark they were?
If you leave said stories as pure digital, perhaps a second copy into a specific directory called "Recorded Predictions" or some such.
I think there might be a decent story in how well their predictions fare. Perhaps broken down into "Nearest the Mark" and "Completely Fekkin' Wrong" sections to make it perfectly clear just how well/bad they do.
Predictions are like arseholes: everyone has one, everyone knows they're full of shite, but people keep expecting rainbow sparkly faerie dust for some damned reason...
The problem is that automakers build their cars needing separate chips to control every little function. There's no reason that a property designed car couldn't have a single system controlling the rear seat air conditioning, front seat air conditioning, heated seats, heated steering wheel, cabin lighting and so on instead of each needing their own chips.
Surely the engineering required to make that happen would be well worth it and give you a competitive advantage over other automakers who don't need so many chips per car, or are selling de-specced models to reduce the number of chips? I know that can't appear immediately but hopefully that engineering is well underway. But I suspect not, some beancounter ran the numbers and showed the payoff, while significant, would be a few years off, and was vetoed by some exec who figures he'll probably be working somewhere else by then so he cares only about his bonus this year and next.
The problem is that automakers build their cars needing separate chips to control every little function. There's no reason that a property designed car couldn't have a single system controlling the rear seat air conditioning, front seat air conditioning, heated seats, heated steering wheel, cabin lighting and so on instead of each needing their own chips.
There's no reason why they couldn't, but also no reason why they should. MCUs have long been just another electronic component from the designers perspective. If a chip is 50p it's easy to justify if it saves other cost and weight (e.g. that of wiring), allows flexibility for design variants (e.g. 3 door, 5 door, van versions) or keeps the component generic in nature so it can be used on different models.
That's before you consider the benefits of a distributed system. If one chip runs one electric window there might be a hundred lines of code in there and it's trivial to analyse. If one big chip it's doing all the windows, the heaters, the demisters, the central locking and so on it becomes a lot more complex. The interactions become a lot less predictable as a result. Would you be happy if your ABS stops working at a key moment because the passenger happened to be winding down their window at the same time?
Automotive safety critical systems are managed separately to in-car entertainment, air-con, etc. It is not the 50p chip it is the £50 ECU sourced from a 3rd party that the chip sits in.
My recent experience of one UK automotive is that they think the digital platform of a vehicle is an Electrical Engineering problem and the IT Profession has nothing to contribute so they are now trying to solve software management problems like Configuration/Version/Build/Release Management, etc. that IT solved decades ago. They are literally reinventing the wheel (pun intended). But frequently do not realise a wheel is even required.
Distributing the controllers around the car and allowing them to communicate on the CANbus was such a huge step forward in automotive technology that manufacturers will never go back.
This network-like approach hugely simplifies automotive writing over an old fashioned electrical loom with a dedicated write for each function. There are huge savings in cost (due to quantity of materials and ease of manufacture), and of weight. Reliability is hugely improved. An old fashioned wiring loom designed to support modern automotive technology would be a truly monstrous thing if it was even possible at all.
This post has been deleted by its author
I'm pretty sure the company in question is Krone. They make agricultural equipment, most notably forage harvesters.
Heard stories through their dealers for a couple of months now about them buying washing machines because they can't get the microcontrollers otherwise. They need to get their machines out the door ASAP as the European harvest season is starting in earnest, and £500 for a washing machine on a £500,000 harvester won't make much difference to their bottom line.
No-one uses 4-bit or mask ROMs nowadays.
At minimum, they'll have an 8-bit or 16-bit processor with on-chip flash memory for the code.
Some of them probably have 32-bit ARM cores. The ARM Cortex M series are cheap, and there are several reasonably cheap microcontrollers that use them.
At the high end... well, my tumble dryer has WiFi and an app to tell you when it's done. Lots of computing power. (That I don't use, because why would I want the security risks of connecting my tumble dryer to the Internet).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
