Global Chip Shortage Threatens Production of Laptops, Smartphones and More


FBnil

geekologie is no more :(
Joined
Dec 14, 2012
Messages
4,861
Location
Yurp
This is quite a bad situation: You have hiccups in your chip supply, you build competing chip manufacturing factories all over the world. In 5 years, you have an oversaturated chip market. The factories are still new and need to recover the investment, and they would go broke if not, again, helped out by local goverment with a financial injection. The saturated market will make chip prices go down, but quality as well... so we are looking again to a situation similar to the one in the 80's. Where low quality chips died before the computer was thrown away (chips got extra hot, so you could find them, desolder and replace them. But now with microelectronics, good luck re-soldering the +40 microconnections...

(not totally the same of course, in the 80ties, there was a super demand of chips and low production. Now we will have a super production and medium demand)

it feels like this joke:
I had to buy a new car. Why, did you crash it? No, the ashtray was full.

So maybe by that point in time, new markets are found (that I.O.T - Internet of Things people have been talking about) or the OPPC Foundation (One Pyra Per Child Foundation).
 

TeDaDeS

Hardcore Member
Joined
Jan 15, 2004
Messages
1,558
Location
The Netherlands
Website
Visit site
This is quite a bad situation: You have hiccups in your chip supply, you build competing chip manufacturing factories all over the world. In 5 years, you have an oversaturated chip market.
When the new fabs are there you can scale up but also down, while currently we are already at maximum capacity.
While there might be a period of over-production I think it will settle. Every country should at least be able to produce what they need themselves.
Having these kind of situations might also spark some creative solutions.

What could happen is that due to the high rate of innovation products become out-dated really quickly. Intel is aiming to have their 20A nodes running in ~2025, which is really ambitious, but it could also mean an influx of cheap high-end chips.
While I'm against creating waste I also see that it would be good to replace older electronics with more modern, faster, and energy saving ones. And doing this cheap will make it available to a lot of people, so that's good I would think.

I also wonder what would be good about IOT. Probably because there are lot of unnecessary products that have an internet connection slammed onto it. Like a wifi-microwave, recently there was such a product that got bricked due to an automatic firmware update.
And why does it need firmware updates -> because it's connected to the internet. I can imagine it being handy to do some stuff via your phone, the way it's being done now doesn't always make me happy (feels like producing waste).
 

FBnil

geekologie is no more :(
Joined
Dec 14, 2012
Messages
4,861
Location
Yurp
see that it would be good to replace older electronics with more modern, faster, and energy saving ones.
Well that's the thing... if you go to a smaller nanometer technology, you need to update your factory as well. If the factory has not earned back the initial investment AND gathered enough funds to do such an upgrade, well, the state will have to pay again. Or you are stuck with the older tech microchips (which might not be that big of a problem, we also need lowtech microchips)
 

TeDaDeS

Hardcore Member
Joined
Jan 15, 2004
Messages
1,558
Location
The Netherlands
Website
Visit site
Well that's the thing... if you go to a smaller nanometer technology, you need to update your factory as well. If the factory has not earned back the initial investment AND gathered enough funds to do such an upgrade, well, the state will have to pay again. Or you are stuck with the older tech microchips (which might not be that big of a problem, we also need lowtech microchips)
While I agree, and makes sense, that fabs have to make enough money to keep production running there are some nuances.
If a fab doesn't make enough money they depend on investors or subsidies.
In this case countries / continents determined they should have this chip production capability, so for sure there will be a lot of money invested from governments there. And those will not directly return value by generating money, only by providing this capability.

High-end fabs usually don't upgrade, they build a new fab instead.
The production process determines the equipment, the equipment determines the fab. Each node has it's own production process.
You do have process improvements (sometimes called half-nodes) like 7nm->6nm, or for Intel 10nm->10nm++. But that is usually done with the same equipment.
Equipment can be upgraded/tweaked also, but it usually doesn't change the production process itself.

There are only a few companies running high-end nodes: TSMC, Intel and Samsung. They also still make enough money to invest into newer nodes.
And if they run out of money to invest, they indeed lose the race (look at GlobalFoundries for example).
And as mentioned: not everyone needs the latest nodes. Having a different production technology can also be done with existing nodes.
For example: if you make memory modules you have different demands. They use 3D memory, and stacking of dies for example.
While something similar is used for logic chips (CPU/GPU) the main driver there is to have more, faster, and energy saving logic on a chip.

So making money with a fab depends on the product they are making. The characteristics of this product are determined by the production process and used technology.
For example: AMD used multiple dies for 1 CPU, while Intel uses a monolithic die. AMD could leverage multiple nodes, the I/O-die would be made on a cheaper node. And having multiple dies means they high higher yields.
The node itself determines the costs and limitations of this product. It's up to the person selling the product to make it worthwhile. Marketing can make a difference there for example (see Intel again, they sell for a higher price).
Any country should be able to have a good mix of the right products produced on the right node, and scale-up/-down if needed. If not they should become creative, what they usually do is just tax imports of competing products to make their production more viable.
 
  • Like
Reactions: rSl

levi

Still fresh, damnit!
Joined
Oct 6, 2008
Messages
15,685
Location
Somewhere off the coast of the EU
This is quite a bad situation: You have hiccups in your chip supply, you build competing chip manufacturing factories all over the world. In 5 years, you have an oversaturated chip market. The factories are still new and need to recover the investment, and they would go broke if not, again, helped out by local goverment with a financial injection. The saturated market will make chip prices go down, but quality as well... so we are looking again to a situation similar to the one in the 80's.
You say oversaturated, I see competion where you can get a similar part made in many different factories around the world.
 

FBnil

geekologie is no more :(
Joined
Dec 14, 2012
Messages
4,861
Location
Yurp
You say oversaturated, I see competition where you can get a similar part made in many different factories around the world.
Yes, and good that the prime matters (silica) are also everywhere. Imagine the child labour having to work harder to mine a rare metal because the world now needs these to produce chips in different places of the world.
 

TeDaDeS

Hardcore Member
Joined
Jan 15, 2004
Messages
1,558
Location
The Netherlands
Website
Visit site
Yes, and good that the prime matters (silica) are also everywhere. Imagine the child labour having to work harder to mine a rare metal because the world now needs these to produce chips in different places of the world.
Are you referring to poly-silicon for solar panels, I couldn't find any source that wafers are produced with silicon from child labor, if you have a link I would appreciate that. It would certainly object to those practices.

I did encounter child labor in India, in a factory producing towels. I raised the topic with the child's mother, and she claimed she rather has her daughter working with her in the workplace then the child being at home alone.
If the child would be at home she would stay at home also, meaning she couldn't work. And having the additional income from her work and her child really helped them survive.
Since then I try to see if any claims fall into this category, but working in a mine I find implausible to be positive for the child.
 

FBnil

geekologie is no more :(
Joined
Dec 14, 2012
Messages
4,861
Location
Yurp
but working in a mine I find implausible to be positive for the child.
Unsupervised it goes into delinquency? At home bad guys come in and the child is raped? Giving it a work ethos, even if it's a shitty job? Their dad will be able to buy more booze, so happy dad is less getting hit? Or, my favourite, if you don't go down the mine they have to spend bullets and the other children will have to make up for the target the dead child is not making? (Congo - alledgedly in the Diamond mines, where they have guns just in case you take a rock with you...)
Axel Rose wrote a song about it: "Ohh, sweet child of Mine". (Ohh is H2O).
 

TeDaDeS

Hardcore Member
Joined
Jan 15, 2004
Messages
1,558
Location
The Netherlands
Website
Visit site
  • Like
Reactions: rSl

FBnil

geekologie is no more :(
Joined
Dec 14, 2012
Messages
4,861
Location
Yurp
New Intel fab in Magdeburg (Germany):
Magdeburg: https://en.wikipedia.org/wiki/Magdeburg
Source: https://www.intel.com/content/www/u...room/news/eu-news-2022-release.html#gs.tszosh

"Announced in March 2022, the 17 billion euro project will deliver computer chips using Intel's most advanced transistor technologies.", which probably is 20A (2nm) or smaller. Those probably use the newer High-NA EUV equipment.

That's a ~5 hour drive from ED's shop to Intel's fab.
Too bad they'd produce x64 instead of ARM
 

TeDaDeS

Hardcore Member
Joined
Jan 15, 2004
Messages
1,558
Location
The Netherlands
Website
Visit site
"Intel also announced partnerships with several companies aligned with this fund and focused on key strategic industry inflections: enabling modular products with an open chiplet platform and supporting design approaches that leverage multiple instruction set architectures (ISAs), spanning x86, Arm and RISC-V."
Link: https://www.intc.com/news-events/pr...el-launches-1-billion-fund-to-build-a-foundry

It was once a rumor Intel would buy SiFive: https://www.bloomberg.com/news/arti...ifive-is-said-to-draw-intel-takeover-interest
"Intel offered to acquire SiFive for more than $2 billion". This hasn't happened.
 
Top