TSMC Production Costs Set To Rise

Krenum

Krenum

Fully [H]
Joined
Apr 29, 2005
Messages
19,193
Chip manufacturing set to rise, raising the CPU / GPU costs even more than they are now!

https://asia.nikkei.com/Business/Te...kes-chip-prices-up-to-20-amid-supply-shortage

"Contract chipmaking giant Taiwan Semiconductor Manufacturing Co. told clients Wednesday that it plans to raise prices on products by as much as 20%, in what would be the company's steepest single increase".
RTUt.gif
 
And Harry and Lloyd had a hand in it?

Edit: N/M. Too dependent on product technology.
 
Last edited:
Probably to fund that huge fab they are building in Arizona, or did that get mothballed ?
 
Gross. 20% increase in manufacturing cost usually results in a more significant increase in retail. Unless of course if the manufacturers absorb the increase... bwahahahahahaha
 
[Spectre] said:
And the land of diminishing returns is real.
Click to expand...
But now that IBM has that 3nm node up and running TSMC can’t let themselves fall behind. That and they want to cut a year off the build time on their US based fab so I’ve been reading.
 
Lakados said:
But now that IBM has that 3nm node up and running TSMC can’t let themselves fall behind. That and they want to cut a year off the build time on their US based fab so I’ve been reading.
Click to expand...
IBM still has FABs? Or is that R&D or equipment designs?
 
Lakados said:
They fab their own equipment in-house and license out tech for others.
Click to expand...
So, just being like sure, they don't have a fab of their own that will produce these technologies and designs?
 
[Spectre] said:
So, just being like sure, they don't have a fab of their own that will produce these technologies and designs?
Click to expand...
They have a small fab they use for making their own stuff internally but no they don't build stuff for others, they design the tech that the companies who make the equipment that TSMC and the likes purchase.

There was an article floating around back earlier this year about Intel and IBM supposedly partnering up to get their 3nm process out the door at a more rapid pace since IBM's partnership with Global Foundries has fallen apart. Just have to dig that out from the archives somewhere for references.

Found it, it was 2nm not 3nm
https://www.forbes.com/sites/linley...-could-revive-intel-fab-tech/?sh=157102fc4426

Intel’s manufacturing woes have been widely reported, but the company has a secret weapon to solve them: A new alliance with IBM. That company’s recent announcement of the world’s first two-nanometer (2nm) transistor provides a roadmap for how Intel can regain its manufacturing mojo and improve the competitiveness of its processors.

TSMC plans to put into production next year a 3nm FinFET technology, but the foundry will switch to GAAFET for its 2nm process, which is likely to enter production in 2024 or 2025. IBM’s 2nm process offers greater density (more transistors per square millimeter) and lower power than TSMC’s 3nm, so it will likely be similar to that company’s 2nm. Although IBM no longer builds chips, it expects its partners will bring the 2nm technology into production around the end of 2024, about the same time that TSMC reaches that level.
 
Last edited:
According to WSJ 10% for more advanced process that Nvidia / Amd use:

Taiwan Semiconductor Manufacturing Co. plans to increase the prices of its most advanced chips by roughly 10%, while less advanced chips used by customers like auto makers will cost about 20% more, these people said. The higher prices will generally take effect late this year or next year, the people said.
 
imsirovic5 said:
According to WSJ 10% for more advanced process that Nvidia / Amd use:

Taiwan Semiconductor Manufacturing Co. plans to increase the prices of its most advanced chips by roughly 10%, while less advanced chips used by customers like auto makers will cost about 20% more, these people said. The higher prices will generally take effect late this year or next year, the people said.
Click to expand...
And somehow cars will cost 20% more too lol
 
drklu said:
Probably to fund that huge fab they are building in Arizona, or did that get mothballed ?
Click to expand...
No, plans for that fab are apparently on a strict timeline, as it is supposedly to become the primary fab for producing Apple's A-series chips starting in 2024. TSMC plans to hire up to 1,900 people by 2023.
 
"The 100 billion dollar investment in expansion has fueled qualms about potential loss in profitability."

😆 Potential loss in profitability? Rrriigghhtt and ive got some oceanfront property in Phoenix ill sell ya real cheap ;) Hell even i couldve come up with a better excuse to raise prices!
TSMC is only booked solid for what? The next 50 years!?
 
hititnquitit said:
"The 100 billion dollar investment in expansion has fueled qualms about potential loss in profitability."

😆 Potential loss in profitability? Rrriigghhtt and ive got some oceanfront property in Phoenix ill sell ya real cheap ;) Hell even i couldve come up with a better excuse to raise prices!
TSMC is only booked solid for what? The next 50 years!?
Click to expand...
TSMC in its most profitable year to date made themselves 17.6B, so this 100B investment in expansion would take 6 years to pay off assuming they put everything they make into paying that off, which they can’t meaning it’s going to take them probably closer to 20 years to pay that down. They are “only” worth some 90B right now so this investment is literally doubling their current size. It’s a pretty big deal
 
Lakados said:
TSMC in its most profitable year to date made themselves 17.6B, so this 100B investment in expansion would take 6 years to pay off assuming they put everything they make into paying that off, which they can’t meaning it’s going to take them probably closer to 20 years to pay that down. They are “only” worth some 90B right now so this investment is literally doubling their current size. It’s a pretty big deal
Click to expand...
Of course it is! Its a 100b!
There is zero chance of a loss in profitability was my point. China is quite literally the only entity that can slow their growth. Without going political, not even they are that foolish. Profitability is going to continue to skyrocket with every new fab. Japan and Europe being the next targets, they are huge profit insulators both literally and figuratively. TSMC is doing fiiine, fine and will continue to experience unparalleled growth. Demand certainly isn't going to wane.
Even when Intel finally gets their fabs up and running to current standards, TSMC has so many (BIG)well established customers Intel will have far to much to prove to make many inroads. They will be forced to pick up TSMC scraps.
The only thing that can hold TSMC back is TSMC.
 
hititnquitit said:
Of course it is! Its a 100b!
There is zero chance of a loss in profitability was my point. China is quite literally the only entity that can slow their growth. Without going political, not even they are that foolish. Profitability is going to continue to skyrocket with every new fab. Japan and Europe being the next targets, they are huge profit insulators both literally and figuratively. TSMC is doing fiiine, fine and will continue to experience unparalleled growth. Demand certainly isn't going to wane.
Even when Intel finally gets their fabs up and running to current standards, TSMC has so many (BIG)well established customers Intel will have far to much to prove to make many inroads. They will be forced to pick up TSMC scraps.
The only thing that can hold TSMC back is TSMC.
Click to expand...
Well they don't have the means to pay out 100B in cash, so that means loans and loans that big have interest so that means a significant portion of their profits are going to be paying that down which in turn means a loss of profits. There is also the very strong likelyhood that by the time their new investments are online Intel will have caught up, their IBM partnership for 2nm is expected to be online for 2024, and as those are going to be contracted fabs for production TSMC is going to be facing increasing pressure in 2 years time so that means they will have to lower prices at that stage to compete and they won't be able to get the same rates they are now.

Besides the biggest hike is on the older nodes for Auto Manufacturing that they have the least profits on, that they were happy not really using until the automakers lobbied the governments who strongarmed them into opening them back up cutting wafer supply from AMD and the other higher-paying clients. So might as well charge them a premium if they are gonna be dicks about it.
 
Lakados said:
Well they don't have the means to pay out 100B in cash, so that means loans and loans that big have interest so that means a significant portion of their profits are going to be paying that down which in turn means a loss of profits. There is also the very strong likelyhood that by the time their new investments are online Intel will have caught up, their IBM partnership for 2nm is expected to be online for 2024, and as those are going to be contracted fabs for production TSMC is going to be facing increasing pressure in 2 years time so that means they will have to lower prices at that stage to compete and they won't be able to get the same rates they are now.

Besides the biggest hike is on the older nodes for Auto Manufacturing that they have the least profits on, that they were happy not really using until the automakers lobbied the governments who strongarmed them into opening them back up cutting wafer supply from AMD and the other higher-paying clients. So might as well charge them a premium if they are gonna be dicks about it.
Click to expand...
I hear you and while the payments sound daunting they wouldn't have taken on more debt than they are capable of paying off quickly. I don't know the specifics obviously but we all know they are too smart to take on a bad loan.
They could build 5 or 6 more fabs and would mayyybe be able to meet demand in 10 years. The current chip shortage is a magnification of where demand is and it will only continue to increase massively for a very long time to come. Which of course means bigger and bigger profits for those at the crest of the wave. TSMC is in the perfect position to capitalize and they will. They generally don't make mistakes.
The Intel-IBM partnership sounds impressive but i have my reservations as to Intels ability to scale such a small node so quickly. Even with IBMs help im doubtful. Delays have kind of become synonymous with anything cutting edge and Intel unfortunately. Im not saying they wont get there, im sure they will. Targets just have a habit of moving for them(as we all know). Intel having the ability to undercut TSMC? I dont have much faith in that either. Unless they are going to be willing to take massive losses for the first couple of years(not out of the question) i cant see them forcing price cuts right away. Intel is hugely profitable but i don't know if they want to kick off a brand new node taking losses if they feel they are competitive. Most likely they will come in right with TSMC and try to sell their tech or quality first. If that fails they will probably start price cutting. By that time TSMC will have wrapped up so much of the market(and mind share), will companies be willing to take a chance (Nvidia/Samsung come to mind) with Intels new unproven node? Whoo, i don't know. I just don't see businesses flocking to Intel.
Which is why i said they will likely be relegated to feeding off of TSMCs scraps. At least until they can show some proof that their node is comparable. But like most things tech related we get to hurry up and wait to see how it all shakes out. Weeee!

Ah the automakers. I agree, they made their bed...
They panicked. Plain and simple. All that overhead collecting dust, unions and investors banging on the door. Yeah, i feel no sympathy for them. Suffer like the rest of us or USE your clout to do something about it rather than look to the gov to bail you out.
Folks in need of a car are getting the raw end of the deal but that was going to happen no matter what.
 
hititnquitit said:
I hear you and while the payments sound daunting they wouldn't have taken on more debt than they are capable of paying off quickly. I don't know the specifics obviously but we all know they are too smart to take on a bad loan.
They could build 5 or 6 more fabs and would mayyybe be able to meet demand in 10 years. The current chip shortage is a magnification of where demand is and it will only continue to increase massively for a very long time to come. Which of course means bigger and bigger profits for those at the crest of the wave. TSMC is in the perfect position to capitalize and they will. They generally don't make mistakes.
The Intel-IBM partnership sounds impressive but i have my reservations as to Intels ability to scale such a small node so quickly. Even with IBMs help im doubtful. Delays have kind of become synonymous with anything cutting edge and Intel unfortunately. Im not saying they wont get there, im sure they will. Targets just have a habit of moving for them(as we all know). Intel having the ability to undercut TSMC? I dont have much faith in that either. Unless they are going to be willing to take massive losses for the first couple of years(not out of the question) i cant see them forcing price cuts right away. Intel is hugely profitable but i don't know if they want to kick off a brand new node taking losses if they feel they are competitive. Most likely they will come in right with TSMC and try to sell their tech or quality first. If that fails they will probably start price cutting. By that time TSMC will have wrapped up so much of the market(and mind share), will companies be willing to take a chance (Nvidia/Samsung come to mind) with Intels new unproven node? Whoo, i don't know. I just don't see businesses flocking to Intel.
Which is why i said they will likely be relegated to feeding off of TSMCs scraps. At least until they can show some proof that their node is comparable. But like most things tech related we get to hurry up and wait to see how it all shakes out. Weeee!

Ah the automakers. I agree, they made their bed...
They panicked. Plain and simple. All that overhead collecting dust, unions and investors banging on the door. Yeah, i feel no sympathy for them. Suffer like the rest of us or USE your clout to do something about it rather than look to the gov to bail you out.
Folks in need of a car are getting the raw end of the deal but that was going to happen no matter what.
Click to expand...
IBM already has their 2nm process functional inside their labs. The are in the scale up process now. They fully expect to hit their 2024 timetable with the two new fabs that Intel is making for them. Intel’s timetable for deployment gets their 2nm in production a whopping 4-6 months ahead of TSMC’s which based on current testing are about on par with each other.

TSMC though can easily pay it off, that 100B gets them 3 new fabs on their 5nm or 3nm processes. And yeah that will make it easier to pay off that 100B, but it’s still going to take 5-10 years and Investors are immediately concerned with quarterly results with long term gains secondary. I mean If they take half their current annual profits and put it back towards the loan at their new price structure that would take them 10 years. To most that sounds reasonable but all investors are going to hear is “hey for the next decade we’re only half as profitable”
 
Last edited:
Lakados said:
IBM already has their 2nm process functional inside their labs. The are in the scale up process now. They fully expect to hit their 2024 timetable with the two new fabs that Intel is making for them. Intel’s timetable for deployment gets their 2nm in production a whopping 4-6 months ahead of TSMC’s which based on current testing are about on par with each other.

TSMC though can easily pay it off, that 100B gets them 3 new fabs on their 5nm or 3nm processes. And yeah that will make it easier to pay off that 100B, but it’s still going to take 5-10 years and Investors are immediately concerned with quarterly results with long term gains secondary. I mean If they take half their current annual profits and put it back towards the loan at their new price structure that would take them 10 years. To most that sounds reasonable but all investors are going to hear is “hey for the next decade we’re only half as profitable”
Click to expand...
This is true and endemic among investors. At some point, I wonder if all of the short-term gains at the expense of long-term viability in many industries will result in catastrophe for the market as a whole. A group of motivated investors can wreak havoc on a business, regardless of its fiscal health, if they smell a faster return on their investment.
 
You must log in or register to reply here.
Back
Top