Although I do think there will be an oversupply "crisis", I think the article is too optimistic about the timing. From what I have read the new foundries will start production in 2024-2025, and I don't see demand going down anytime soon, and that's not to mention some of the backlog that has accumulated. Although I am curious as to how amd and nvidia have been handling this. Do they have stockpiles of chips themselves?
There’s alot of stuff stuck in the shipping channel or backlogged.
Not the same commodity, but as an illustration, I just got a bunch of 4k computer displays direct from an OEM. Normally we get stuff like this 30-60 days after manufacturing. These devices were tagged as last July and received a week ago.
I’ve been told that lots of stuff gets stuck waiting for final package assembly. (Ie the monitor stand, cable or even manual) Although chips aren’t end user products that have to have accessories, many require packaging or other subcomponents that may be lost in the mail.
Additionally, contracts prioritize certain customers, and many manufacturers don’t have good processes to deal with diversions. If you order 20,000 widgets, they may not stop shipping until to hit some high water mark. So when the US government say “emergency, ship me your widgets now”, your order gets “stolen”.
COVID response activity is winding down, so I’d venture to guess you’re going to see a lot of cancelled orders and chaos. Imo, you’ll see prices of consumer facing IT gear crash in the June-August timeframe as schools and students are flooded with gear, only to surge again as component makers retool.
Also, like gas prices (quick to rise, slow to sink), I think you’ll see manufacturers keep prices high in the many markets that are controlled by little cartels. Why sell Ford some chip for $1 when they are paying $12 today?
> Why sell Ford some chip for $1 when they are paying $12 today?
That's a very good point. When you're competing to get your part selected for a new design, you want to sell it cheap, but once they've done all the work of incorporating your part and testing it, you want to make it expensive. It will be interesting to see how this plays out.
Yes they do. First, they normally buy complete electronics (functional devices with or without software), not just chips. Second, they do have strong contracts, spreading for 3, 4, 5+ years, specifying the volumes for each contracted year and also the piece price (among other details, of course).
To add to the logistical supply chain issues, the fleet of Anotov aircraft (including the AN-125 and AN-225) play a crucial role in the fulfilling a range global logistical needs, but the factory in Ukraine is reportedly damaged, leading to expected impacts to their serviceability and spare parts manufacture.
This will almost certainly have flow-on impacts to the broader shipping and commercial airline market, and further screw up delivery of your monitor’s user manuals.
Oversupply is just as much a crisis as undersupply. The lead time on ramping capacity up and down is long, which leads to a cyclic market which overshoots in one direction or the other. It's hard to predict the cycles.
However, an "oversupply" crisis means that either:
- prices falling, meaning costs aren't covered;
- equipment isn't running 24/7 as intended, leading to higher costs
This means businesses go under. There's a firesale which is nice, but the structural damage is far more harmful.
Personally, I'd like to see this industry a little bit more socialized. I'd like my chips produced in the USA, even if they cost a little bit more, so we're self-sufficient. I'd also like the supply chains to be in the US. I think the same goes for the EU, China, and hopefully soon, India and Africa. Having five independent supply chains costs five times as much, and I understand the efficiency argument for having just TSMC (and maybe Samsung or Intel), but I think both COVID and Ukraine highlight the risks of systemic failure.
We're just a little too over-dependent on each other.
I also wouldn't mind if my taxes paid for some excess, unused capacity, as they do with food production. That's also less efficient, but gives resilience.
We don't need to be self-sufficient everywhere, mind you. Most goods aren't essential. I do think each region should be self-sufficient for food, medicine, and now, ICs.
I have a tiny little side business making an electronic gadget. I've had to get good at substitution. Granted this is easier for me, I don't have to go through an engineering change order process, and can usually trust my gut on what's going to work. Still, it's a PITA.
Yeah it's a pain for those in the business. Thankfully it's just a hobby for me, but I've talked to professionals who were on their fourth board revision due to substitutions... can't be fun.
I feel like there’s a great opportunity here for electronic engineers with a good eye for supply chain dynamics to consult into all manner of industrial and consumer markets to assist with strategic re-engineering of their products.
