TSMC May Increase Wafer Pricing by 10% in 2025

[fansink]

Due to the high demand for advanced processors in consumer electronics and high performance computing, TSMC is looking to hike its wafer pricing for all types of customers in 2025, according to a Morgan Stanley note to clients cited by Eric Jhonsa , an investor.

TSMC may increase wafer pricing by 10% for 2025: Report

Chips are getting more expensive.

www.tomshardware.com

 

[fansink]

TSMC Decides To Slash 7nm Pricing Amid Weakening Demand, 3nm Costs Rises Up By 10%

TSMC has decided to raise 3nm chip pricing by a decent margin but has also heavily slashed 7nm process prices amid weakening demand.

www.google.com

 

[Scotten Jones]

Historically when utilization goes down the foundries drop prices to fill the fabs, and when utilization goes up prices go up. Since the Pandemic TSMC's attitudes seems to be raise prices every year no matter what. The 3nm ramp is spoken for but they have unused capacity at all nodes 5nm and larger. the 7nm cut makes sense but they have an even bigger problem at 5nm. 5nm hit 100% utilized at the end of 2022 but then fell off a lot and is still underutilized.

 

[BruceA]

Historically when utilization goes down the foundries drop prices to fill the fabs, and when utilization goes up prices go up. Since the Pandemic TSMC's attitudes seems to be raise prices every year no matter what. The 3nm ramp is spoken for but they have unused capacity at all nodes 5nm and larger. the 7nm cut makes sense but they have an even bigger problem at 5nm. 5nm hit 100% utilized at the end of 2022 but then fell off a lot and is still underutilized.

N3 F18b has high depreciation, N5 Fab18a, is mostly depreciated and Fab15 N7 depreciated enables a very flexible pricing approach that enable high margins even if underutilized. The numbers speak for themselves the amazing foundry model that is the one Intel hope to transition to in 5 years

 

[Scotten Jones]

N3 F18b has high depreciation, N5 Fab18a, is mostly depreciated and Fab15 N7 depreciated enables a very flexible pricing approach that enable high margins even if underutilized. The numbers speak for themselves the amazing foundry model that is the one Intel hope to transition to in 5 years

None of the "N5 Fab 18a" equipment is depreciated yet, your "mostly depreciated" statement is wrong.

When equipment is still depreciating that makes it even more important to keep the fab full. Any price above variable cost helps absorb fixed cost so you drop prices to at least absorb some of the cost, otherwise the fixed costs kills your bottom line.

I am not saying TSMC isn't a great foundry, I am just surprised they aren't cutting prices at 5nm to fill the fab. It may just reflect that they think they have all the available 5nm business anyway but it is the kind of thing that will make customers look harder for alternatives.

 

[nghanayem]

I am not saying TSMC isn't a great foundry, I am just surprised they aren't cutting prices at 5nm to fill the fab. It may just reflect that they think they have all the available 5nm business anyway but it is the kind of thing that will make customers look harder for alternatives.

TSMC did talk about converting some of their N5 capacity to N3 production. So maybe they think there is enough demand on N3 for the couple of N3 phases+1-2 converted N5 phases to solve the problem? Could also be that they expect alot of trailing customers on N4RF and the like. However given how poorly utilized N7 fabs are now that even low end APs are moving to 5”nm” class technology, that later possibility seems unlikely.

As for depression TSMC needs to start depreciating as soon as they start running non too qual material. So presumably F18a tooling should be depreciated this year as they started risk production/running customer cert material in mid 2019, and presumably they were up and running development wafers post F12 transfer in like 2018/19 (I dont really know how early TSMC hands over process development to HVM1, so dates may be wrong). Did you make the comment because ROC tax laws allow depreciation not to start until TSMC starts making revenue producing wafers?

 

[BruceA]

None of the "N5 Fab 18a" equipment is depreciated yet, your "mostly depreciated" statement is wrong.

When equipment is still depreciating that makes it even more important to keep the fab full. Any price above variable cost helps absorb fixed cost so you drop prices to at least absorb some of the cost, otherwise the fixed costs kills your bottom line.

Fab 18a is N5 and N4 and been running years!! I am curious P1 has been running for three years, the last wave of capacity brought up in P7 has largely been idle, but the 100k capacity has been running. What creative finance allows you to run production tools for one to three years and meet your NONE statement ?

You have confused Fab18b which is N3 and is depreciating now.

 

[Scotten Jones]

Fab 18a is N5 and N4 and been running years!! I am curious P1 has been running for three years, the last wave of capacity brought up in P7 has largely been idle, but the 100k capacity has been running. What creative finance allows you to run production tools for one to three years and meet your NONE statement ?

You have confused Fab18b which is N3 and is depreciating now.

TSMC uses five year depreciation for equipment as disclosed in their annual report. Depreciation doesn't start until equipment is in production, I am not confusing anything, 100% of TSMC's 5nm equipment is still depreciating!

The other thing to keep in mind is they started phase 1 and then brought up phase 2, and phase 3. In each case they put some capacity on-line and then added to it over time. If you put 10k wpm on line, and then 6 months later add another 10k wpm, each group depreciates on a different time line. I track every fab by phase and by capacity over time.

Phase 1 and 2 will still be depreciating at the end of this year, at the end of next year roughly half the equipment will be depreciated in each. For phase 3 all the equipment will still be depreciating at the end of next year. Plus eventually Arizona will come on-line with new 5/4nm equipment that will start depreciating then.

 

[Scotten Jones]

TSMC did talk about converting some of their N5 capacity to N3 production. So maybe they think there is enough demand on N3 for the couple of N3 phases+1-2 converted N5 phases to solve the problem? Could also be that they expect alot of trailing customers on N4RF and the like. However given how poorly utilized N7 fabs are now that even low end APs are moving to 5”nm” class technology, that later possibility seems unlikely.

As for depression TSMC needs to start depreciating as soon as they start running non too qual material. So presumably F18a tooling should be depreciated this year as they started risk production/running customer cert material in mid 2019, and presumably they were up and running development wafers post F12 transfer in like 2018/19 (I dont really know how early TSMC hands over process development to HVM1, so dates may be wrong). Did you make the comment because ROC tax laws allow depreciation not to start until TSMC starts making revenue producing wafers?

Depreciation starts on equipment when it enters production and then TSMC depreciates equipment for five years per their annual report.

The other thing is that TSMC brought up Fab 18 phase 1 and then ramped it up over time, and then phase 2 and ramp and then phase 3 and ramp. If you put 10k wpm on-line it starts depreciating, if you put another 10k wpm on-line it starts depreciating when it comes on line so there is a whole portfolio of equipment that came on-line on different dates and becomes fully depreciated on different dates. I track every fab and fab phase individually and for each phase have an initial and up to 12 upgrades states all depreciating on different time lines.

None of TSMC's 5nm will be fully depreciated this year, at the end of 2025 maybe 1/3 of the 5nm equipment fleet in the fab 18 phases will be fully depreciated but at the same time 5/4nm equipment will be coming on-line in Arizona, so worldwide maybe 1/4 of the equipment will be fully depreciated. It will be over five more years before all of TSMC's worldwide 5nm equipment fleet is fully depreciated.

This is a very complex subject to track and I have spent over two decades immersed in this.

 

[Scotten Jones]

TSMC did talk about converting some of their N5 capacity to N3 production. So maybe they think there is enough demand on N3 for the couple of N3 phases+1-2 converted N5 phases to solve the problem? Could also be that they expect alot of trailing customers on N4RF and the like. However given how poorly utilized N7 fabs are now that even low end APs are moving to 5”nm” class technology, that later possibility seems unlikely.

As for depression TSMC needs to start depreciating as soon as they start running non too qual material. So presumably F18a tooling should be depreciated this year as they started risk production/running customer cert material in mid 2019, and presumably they were up and running development wafers post F12 transfer in like 2018/19 (I dont really know how early TSMC hands over process development to HVM1, so dates may be wrong). Did you make the comment because ROC tax laws allow depreciation not to start until TSMC starts making revenue producing wafers?

Another comment on this is there are tax rules and then what companies use for reporting, those are often not the same and they may keep two sets of books. So while Taiwan allows 3 year depreciation for taxes, TSMC uses 5 years for reporting and their public financial statements are based on this.

 

[Daniel Nenni]

Historically when utilization goes down the foundries drop prices to fill the fabs, and when utilization goes up prices go up. Since the Pandemic TSMC's attitudes seems to be raise prices every year no matter what. The 3nm ramp is spoken for but they have unused capacity at all nodes 5nm and larger. the 7nm cut makes sense but they have an even bigger problem at 5nm. 5nm hit 100% utilized at the end of 2022 but then fell off a lot and is still underutilized.

The big change that I see is lack of foundry competition at 5nm and below. I also feel that the TSMC ecosystem has been undervalued in the past. The entire ecosystem follows TSMC at no cost to TSMC or the TSMC customer. Other foundries have to pay IP companies to port and it is mostly done at the request/cost of the customer.

 

[nghanayem]

Depreciation starts on equipment when it enters production and then TSMC depreciates equipment for five years per their annual report.

The other thing is that TSMC brought up Fab 18 phase 1 and then ramped it up over time, and then phase 2 and ramp and then phase 3 and ramp. If you put 10k wpm on-line it starts depreciating, if you put another 10k wpm on-line it starts depreciating when it comes on line so there is a whole portfolio of equipment that came on-line on different dates and becomes fully depreciated on different dates. I track every fab and fab phase individually and for each phase have an initial and up to 12 upgrades states all depreciating on different time lines.

By this logic did intel 4/3 tooling in D1 that was not used for intel 7 production only start depreciating once intel sold it's first MTL chip to an OEM, rather than when they turned those tools over to the development team's pilot line? My understanding from my engineering economics course in college was that R&D tooling counts as "in production" as soon as you start running value creating experiments. If my understanding is correct, why wouldn't that apply to F18's N5 pilot line pre first revenue generating wafer out?

 

[hist78]

None of the "N5 Fab 18a" equipment is depreciated yet, your "mostly depreciated" statement is wrong.

When equipment is still depreciating that makes it even more important to keep the fab full. Any price above variable cost helps absorb fixed cost so you drop prices to at least absorb some of the cost, otherwise the fixed costs kills your bottom line.

I am not saying TSMC isn't a great foundry, I am just surprised they aren't cutting prices at 5nm to fill the fab. It may just reflect that they think they have all the available 5nm business anyway but it is the kind of thing that will make customers look harder for alternatives.

TSMC started N5 risk production in April 2019 at its Fab18 in Tainan Taiwan. That means around April 2024 or a little bit latter certain part of TSMC N5 equipment had completed its 5-year depreciation schedule. Any newer equipment or equipment put into service later will definitely reach the end of 5-year depreciation schedule on a later date.

 

[hist78]

TSMC uses five year depreciation for equipment as disclosed in their annual report. Depreciation doesn't start until equipment is in production, I am not confusing anything, 100% of TSMC's 5nm equipment is still depreciating!

The other thing to keep in mind is they started phase 1 and then brought up phase 2, and phase 3. In each case they put some capacity on-line and then added to it over time. If you put 10k wpm on line, and then 6 months later add another 10k wpm, each group depreciates on a different time line. I track every fab by phase and by capacity over time.

Phase 1 and 2 will still be depreciating at the end of this year, at the end of next year roughly half the equipment will be depreciated in each. For phase 3 all the equipment will still be depreciating at the end of next year. Plus eventually Arizona will come on-line with new 5/4nm equipment that will start depreciating then.

For tax purpose, typically once an equipment is "placed in service", the depreciation will start as soon as possible. It's not required that the equipment needs to generate revenue or be running in order to record the depreciation expense. I think TSMC will start the equipment depreciation as soon as its risk production started. Additionally for internal accounting purpose the equipment depreciation may start even earlier to truly reflect the useful life of the machine. This may be different from the depreciation reported to the taxing authorities.

 

[Daniel Nenni]

TSMC did talk about converting some of their N5 capacity to N3 production. So maybe they think there is enough demand on N3 for the couple of N3 phases+1-2 converted N5 phases to solve the problem? Could also be that they expect alot of trailing customers on N4RF and the like. However given how poorly utilized N7 fabs are now that even low end APs are moving to 5”nm” class technology, that later possibility seems unlikely.

Hopefully someone will ask this question on the next investor call. I would guess they are.

Samsung was late on 5nm so TSMC got the lion's share out of the gate but now Intel and Samsung have TSMC N5 alternatives so the NOT TSMC market is open for business. For 3nm it is still only TSMC thus far and I think 2nm will be the same, 90%+ TSMC out of the gate. I'm talking about wafer volumes not press releases.

 

[fansink]

For 3nm it is still only TSMC thus far and I think 2nm will be the same, 90%+ TSMC out of the gate. I'm talking about wafer volumes not press releases.

Exactly

 

[hist78]

Hopefully someone will ask this question on the next investor call. I would guess they are.

Samsung was late on 5nm so TSMC got the lion's share out of the gate but now Intel and Samsung have TSMC N5 alternatives so the NOT TSMC market is open for business. For 3nm it is still only TSMC thus far and I think 2nm will be the same, 90%+ TSMC out of the gate. I'm talking about wafer volumes not press releases.

On the other hand, TSMC N5 fab equipment is starting fully depreciated and some of them may already complete the 5-year depreciation schedule. It puts enormous pricing pressure on those Not-TSMC 5nm competitors.

 

[Scotten Jones]

By this logic did intel 4/3 tooling in D1 that was not used for intel 7 production only start depreciating once intel sold it's first MTL chip to an OEM, rather than when they turned those tools over to the development team's pilot line? My understanding from my engineering economics course in college was that R&D tooling counts as "in production" as soon as you start running value creating experiments. If my understanding is correct, why wouldn't that apply to F18's N5 pilot line pre first revenue generating wafer out?

I can tell when equipment starts to become fully depreciated from TSMC's financials and some analysis techniques I have. when they report their quarterly results I get a very strong signal if anything drops off for depreciation. Sometimes after the fact I find out they started depreciation a quarter or maybe two before I thought they did. I generally go by when they start reporting revenue and again that may be off by a quarter or two max, that is my experience over decades of nodes.

If you think about it, taking in enormous depreciation costs in your P&L without offsetting revenue would be a huge drag on profitability.

 

[Scotten Jones]

TSMC started N5 risk production in April 2019 at its Fab18 in Tainan Taiwan. That means around April 2024 or a little bit latter certain part of TSMC N5 equipment had completed its 5-year depreciation schedule. Any newer equipment or equipment put into service later will definitely reach the end of 5-year depreciation schedule on a later date.

In my experience they start depreciating equipment right around when they start reporting revenue, within one or two quarters.

 

[Scotten Jones]

For tax purpose, typically once an equipment is "placed in service", the depreciation will start as soon as possible. It's not required that the equipment needs to generate revenue or be running in order to record the depreciation expense. I think TSMC will start the equipment depreciation as soon as its risk production started. Additionally for internal accounting purpose the equipment depreciation may start even earlier to truly reflect the useful life of the machine. This may be different from the depreciation reported to the taxing authorities.

I don't base anything on what they report to tax authorities.