News TSMC Hikes Price of Chip Production: CPU & GPU Costs Set to Rise

The oligopoly working as intended: companies have practically no choice about where to go to get their chips fabbed, so the only 2-3 games in town can arbitrarily jack up prices for whatever capacity they can squeeze out of their fabs and only add the strict minimum manufacturing capacity required to keep up with booked orders.

The EU and other countries/unions really need to get cracking on national fabs to secure supply diversity and availability as it becomes increasingly difficult for independent fabs to keep up with tech and demand.
 
  • Like
Reactions: ubercake and Jim90
It's called supply and demand.

BTW gloom and doomers: increased profits is what attracts new players to enter the industry thus increasing competition.
 
It's called supply and demand.
For supply-vs-demand to work properly, you need to have a working competitive market.

In a healthy competitive environment, every supplier has an incentive to maintain spare manufacturing capacity to seize new market opportunities and steal customers from other manufacturers. If you refuse to supply something for a price the market will bear, one of your competitors will.

In an oligopoly situation, manufacturers have every incentive to increase capacity by the absolute minimum amount required since their customers have nowhere else to source their parts from and the prices go out of control.

The various component shortages likely wouldn't have been half as bad if there hadn't been so much semiconductor manufacturer consolidation over the last 20 years.
 
  • Like
Reactions: MMorris666
It's called supply and demand.

BTW gloom and doomers: increased profits is what attracts new players to enter the industry thus increasing competition.

Not if sales start dropping due to unacceptable product price increases - that is, unless our pay packets increase by at least the same percentage, and for the vast, vast unlucky majority of us, that never ever happens!
 
The various component shortages likely wouldn't have been half as bad if there hadn't been so much semiconductor manufacturer consolidation over the last 20 years.
We need a simple multi-national law to prevent mergers / acquisitions once the # of players in any industry hits this ___ floor.

IMO, if an industry has 9 or less major players, there should be a BAN on All Acquisitions and Mergers.

International and/or domestic.
 
  • Like
Reactions: ubercake
Everything in the US is going up--and I mean rapidly. Inflation is being felt in all market segments. Much of that is because of increased transportation costs that can be pegged to the Biden administration making the US dependent on foreign sources for oil, again. Prior to Biden's meddling we were 100% energy independent!
 
To be honest, does it really matter anymore? I've practically lost track of how many times I read about price increasing. At the end of the day, it just dampens demand. I can't even be bothered to look up for any new hardware at this point. Just going to sit back and use my existing devices and wait it out. They can increase all they want, but at some point, it may come back to bite them hard.
 
Well, Intel might just benefit from this
Intel can if they choose to. But if competitor is selling their CPU at XX price, why do you think Intel will want to undercut competition in a significant way, assuming Alder Lake will make them more competitive? In terms of using Intel fab, the same applies. Why would Intel want to undercut competition, i.e. TSMC, when they know that all their competitors are increasing prices?
 
For supply-vs-demand to work properly, you need to have a working competitive market.
Incorrect. You are conflating how you think capitalism should work "properly" with the economic law of supply and demand which works even in "improper" markets.

In a healthy competitive environment, every supplier has an incentive to maintain spare manufacturing capacity to seize new market opportunities and steal customers from other manufacturers. If you refuse to supply something for a price the market will bear, one of your competitors will.

In an oligopoly situation, manufacturers have every incentive to increase capacity by the absolute minimum amount required since their customers have nowhere else to source their parts from and the prices go out of control.
Excess capacity is driven by demand projections and the cost to carry said capacity far more than the number of competitors in an industry. NO ONE predicted a global pandemic and therefore demand projections were off regardless of how competitive the industry was.

The various component shortages likely wouldn't have been half as bad if there hadn't been so much semiconductor manufacturer consolidation over the last 20 years.
Just in time manufacturing is a far great cause of the shortage than consolidation. The issue is mostly upstream suppliers, not the chip makers. Further disproving your contention is the fact that the shortages are worse in older process nodes which have many more manufacturers than the leading edge manufacturers.
 
  • Like
Reactions: TJ Hooker
Not if sales start dropping due to unacceptable product price increases - that is, unless our pay packets increase by at least the same percentage, and for the vast, vast unlucky majority of us, that never ever happens!
A) Sales are not dropping due to unacceptable product price increases. TSMC will still sell every chip they can make at their new prices.

B) Good jobs are rare and valuable, so if you want one, you need to develop rare and valuable skills. People will gladly pay you more for such skills, but no one is going to pay you more simply because something you want became more expensive.
 
Excess capacity is driven by demand projections and the cost to carry said capacity far more than the number of competitors in an industry. NO ONE predicted a global pandemic and therefore demand projections were off regardless of how competitive the industry was.
There were on-going years-long supply shortages of many components even before COVID became a thing, covid merely pushed a supply chain that was already on the verge of collapse over the cliff.

When you have no meaningful competition in an industry, you have no meaningful pressure to build capacity ahead of demand since you aren't worried about your competitors stealing your clients and vice-versa when neither has the spare capacity to actually do so and the whole industry gets screwed over by the smallest supply chain disruption.

When you have real competition, the overall spare manufacturing capacity is much greater since every competitor needs enough spare capacity to accommodate customers switching between suppliers if they want to bid on new contracts - can't bid on contracts too large for your spare production capacity.
 
  • Like
Reactions: MMorris666
There were on-going years-long supply shortages of many components even before COVID became a thing, covid merely pushed a supply chain that was already on the verge of collapse over the cliff.

When you have no meaningful competition in an industry, you have no meaningful pressure to build capacity ahead of demand since you aren't worried about your competitors stealing your clients and vice-versa when neither has the spare capacity to actually do so and the whole industry gets screwed over by the smallest supply chain disruption.

When you have real competition, the overall spare manufacturing capacity is much greater since every competitor needs enough spare capacity to accommodate customers switching between suppliers if they want to bid on new contracts - can't bid on contracts too large for your spare production capacity.
No meaningful pressure - you know, other than profit - to build capacity. It's funny that you argue that they are both greedy about profits in their pricing and yet unaffected by profit in building capacity. You can't have it both ways.

What you are failing to grasp is the time and expense of building chip manufacturing capacity. The plants cost billions and take years to build. Build too much capacity and you go out of business. That is a greater threat to the company (and executives' jobs) than potentially losing future market share.

Further, the optimum point in manufacturing is being right at full capacity. Otherwise, you are losing money on wasted capacity. That has a real, you can look at the money spent on it, equipment cost. Willingly moving away from that is only done after a careful calculation that the potential future rewards outweigh the absolutely known future costs.
 
  • Like
Reactions: TJ Hooker
Further, the optimum point in manufacturing is being right at full capacity.
Which is exactly the problem: when there is no meaningful competition, there is no meaningful pressure to build more than the absolute minimum amount of capacity needed to meet baseline demand, so you are perpetually in a situation where the most minute supply chain disruption screws the entire industry.

If private/listed companies won't build sufficient fab capacity to ensure stable supply because they don't want to carry the hit to their profit margins, then governments should step in with their own state/union-owned and operated fabs and make the private/listed companies worry about losing profit margins that way.
 
Which is exactly the problem: when there is no meaningful competition, there is no meaningful pressure to build more than the absolute minimum amount of capacity needed to meet baseline demand, so you are perpetually in a situation where the most minute supply chain disruption screws the entire industry.
You should read my entire post as I already answered this above.

If private/listed companies won't build sufficient fab capacity to ensure stable supply because they don't want to carry the hit to their profit margins, then governments should step in with their own state/union-owned and operated fabs and make the private/listed companies worry about losing profit margins that way.
Europe is talking about doing this. BUT it won't materialize until long after the private companies' CURRENT capacity expansion plans that have been in the works for YEARS are completed. See my above comments about the time required to build a fab.

Also, it won't work out nearly as well as you or they hope, as is the case will every. single. government. intervention. Including those supposing to "increase competition" and make capitalism work "properly".
 
Europe is talking about doing this. BUT it won't materialize until long after the private companies' CURRENT capacity expansion plans that have been in the works for YEARS are completed. See my above comments about the time required to build a fab.
Yes, building fabs takes a while. The question is whether those fabs will bring total capacity at least to par with backlog that will have built up by then. If they don't, then it will still be the exact same situation all over again where companies that want to get stuff made have nowhere to go until someone else with spare capacity becomes available.

Meaningful competition is impossible without meaningful spare manufacturing capacity and supplier diversity.
 
Which is exactly the problem: when there is no meaningful competition, there is no meaningful pressure to build more than the absolute minimum amount of capacity needed to meet baseline demand, so you are perpetually in a situation where the most minute supply chain disruption screws the entire industry.
Huh?!
If there is fierce competition why would you build more then the minimum facilities and be in danger of having them sit idle since having more competition does the opposite of guarantee you more business.
The only reason you would do that is if those fabs would be able to do something that no other fab could, which would be a monopoly again.
 
  • Like
Reactions: TJ Hooker
Huh?!
If there is fierce competition why would you build more then the minimum facilities and be in danger of having them sit idle since having more competition does the opposite of guarantee you more business.
If you have no spare capacity, you cannot bid to steal clients from competitors since you have insufficient facilities to actually fulfill your new contractual obligations. Nvidia wanted to use TSMC for the 3000 series, TSMC couldn't commit sufficient timely wafers for the price it was willing to pay, so Nvidia went with Samsung's inferior 8nm.

In an at least somewhat healthy industry, someone is always ready and able to take new orders on short notice without extortionate pricing.