Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/eetimes-clickbait-tsmc-trims-expansion-plans-as-outlook-dims.16384/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

EETimes Clickbait: TSMC Trims Expansion Plans as Outlook Dims

Daniel Nenni

Daniel Nenni

Admin
Staff member
Years ago EETimes was the gold standard for semiconductor news. I read it every week and it was a crowning achievement when I got an article published. Today EETimes is a shell of its former self, sadly.

So tell me, does this title match the investor call or even the article it represents? I guess the author is trying to stay relevant but that is not the legacy I would be looking for at his age.

My guess is that TSMC will be back to the 20% ish growth rate in 2023 which is hardly a dim outlook.

 
I think the wording of the article is a bit funny, however it's saying the outlook is dimmer vs the end of q1 forecast, not that it's a dim outlook overall. (Our sun is still pretty bright despite being dimmer than lots of other stars).

The US fed is going to drive down demand pretty hard at some point, and it looks like China has demand issues as well given the constant lockdowns. (not being able to accept a Western vaccine means they're stuck in a cycle of no herd immunity and no effective vaccine). Europe is in the same boat fiscally though mainly due to energy costs.. All of these are going to persist through the rest of this year and likely 2023.

At some point these changes have to reduce demand for TSMC products at least vs current forecasts. Though I'm also 100% certain TSMC will take a longer view and can forecast better than any of it's customers and probably competitors.

So yes I agree with it sounding click baity as the title was somewhere between useful and "know this one word to save the universe". But I think demand is going to reduce over the next several quarters.
 
Xebec said:
So yes I agree with it sounding click baity as the title was somewhere between useful and "know this one word to save the universe". But I think demand is going to reduce over the next several quarters.
Click to expand...

Demand has to decrease, there is no doubt about that. The question is will it be a hard or soft landing? I say soft for the foundries but that does not get easy clicks. The sky is not falling. The semiconductor industry is very diverse and much more crash resistant than ever before, absolutely, so no, history will not repeat itself as easily this time.

TSMC has the best data for forecasting based on hundreds of customers and 30+ years of servicing them. Covid was a new twist but I think TSMC handled it expertly. Listen closely to TSMC and cross check with the ecosystem. And now with live events back it is much easier to get a semiconductor sanity check. BY the way, EDA and IP are still growing double digits based on reporting companies and that does not include the China EDA/IP companies so it is actually higher.

One thing you should remember is that today's TSMC is not like TSMC of 10 years ago. Today's TSMC is VERY competitive. I would describe CC Wei's leadership style as "Speaking softly while carrying a very big stick". This is in stark contrast to Pat Gelsinger, Tom Caulfield, and Samsung leadership. CC Wei has positioned TSMC to be a competitive force in the foundry business. Look at the foundry landscape and tell me who would survive if TSMC took a scorched earth approach? GF doesn't make money today in the best of times. What if TSMC targeted GF customers directly? Is there anything GF does that TSMC can't? And even if there is can GF survive based on just that? Would Samsung be in the foundry business without the support of their memory business? The same goes for Intel, no IDM business no foundry.

The best market today for TSMC competitive foundries is the "Not TSMC" business. QCOM, NVIDIA, and others have kept Samsung Foundry on life support for many years to keep TSMC in check. The same can be said for GF and UMC. SMIC has a lock on the China market or they would be in the same boat. But when push comes to shove fabless semiconductor and system companies need a trusted supplier of wafers.
 
Daniel Nenni said:
Demand has to decrease, there is no doubt about that. The question is will it be a hard or soft landing? I say soft for the foundries but that does not get easy clicks. The sky is not falling. The semiconductor industry is very diverse and much more crash resistant than ever before, absolutely, so no, history will not repeat itself as easily this time.

TSMC has the best data for forecasting based on hundreds of customers and 30+ years of servicing them. Covid was a new twist but I think TSMC handled it expertly. Listen closely to TSMC and cross check with the ecosystem. And now with live events back it is much easier to get a semiconductor sanity check. BY the way, EDA and IP are still growing double digits based on reporting companies and that does not include the China EDA/IP companies so it is actually higher.

One thing you should remember is that today's TSMC is not like TSMC of 10 years ago. Today's TSMC is VERY competitive. I would describe CC Wei's leadership style as "Speaking softly while carrying a very big stick". This is in stark contrast to Pat Gelsinger, Tom Caulfield, and Samsung leadership. CC Wei has positioned TSMC to be a competitive force in the foundry business. Look at the foundry landscape and tell me who would survive if TSMC took a scorched earth approach? GF doesn't make money today in the best of times. What if TSMC targeted GF customers directly? Is there anything GF does that TSMC can't? And even if there is can GF survive based on just that? Would Samsung be in the foundry business without the support of their memory business? The same goes for Intel, no IDM business no foundry.

The best market today for TSMC competitive foundries is the "Not TSMC" business. QCOM, NVIDIA, and others have kept Samsung Foundry on life support for many years to keep TSMC in check. The same can be said for GF and UMC. SMIC has a lock on the China market or they would be in the same boat. But when push comes to shove fabless semiconductor and system companies need a trusted supplier of wafers.
Click to expand...

I think "Not TSMC" is also one of Intel Foundry Service's strategies.
 
Daniel Nenni said:
Years ago EETimes was the gold standard for semiconductor news. I read it every week and it was a crowning achievement when I got an article published. Today EETimes is a shell of its former self, sadly.

So tell me, does this title match the investor call or even the article it represents? I guess the author is trying to stay relevant but that is not the legacy I would be looking for at his age.

My guess is that TSMC will be back to the 20% ish growth rate in 2023 which is hardly a dim outlook.

Click to expand...

The article's title and the writing can lead people to think TSMC delaying or reducing its 2022 Capex is due to the decreasing market demand.

This is not what TSMC stated reason. In both Q1 2022 and Q2 2022 earnings conference calls, TSMC repeatedly described the challenges to get manufacturing equipment delivered on schedule. TSMC explained that it's caused by worldwide supply chain constraints the equipment suppliers are facing.

While decreasing demand may partially explain the delay of TSMC 2022 Capex spending, but it's the author own opinion, not TSMC's. He needs to spell it out.
 
hist78 said:
The article's title and the writing can lead people to think TSMC delaying or reducing its 2022 Capex is due to the decreasing market demand.
This is not what TSMC stated reason. In both Q1 2022 and Q2 2022 earnings conference calls, TSMC repeatedly described the challenges to get manufacturing equipment delivered on schedule. TSMC explained that it's caused by worldwide supply chain constraints the equipment suppliers are facing.
While decreasing demand may partially explain the delay of TSMC 2022 Capex spending, but it's the author own opinion, not TSMC's. He needs to spell it out.
Click to expand...

I also heard at Semicon West that the equipment people are getting preferential treatment at TSMC right now to get chips for the equipment that they are waiting for. It's an ecosystem.
 
Daniel Nenni said:
Years ago EETimes was the gold standard for semiconductor news. I read it every week and it was a crowning achievement when I got an article published. Today EETimes is a shell of its former self, sadly.

So tell me, does this title match the investor call or even the article it represents? I guess the author is trying to stay relevant but that is not the legacy I would be looking for at his age.

My guess is that TSMC will be back to the 20% ish growth rate in 2023 which is hardly a dim outlook.

Click to expand...
IMO, Daniel, EETimes has always been more than a bit of a gossip rag. If that was the gold standard, the standards were a lot lower years ago.
 
Daniel Nenni said:
Demand has to decrease, there is no doubt about that. The question is will it be a hard or soft landing? I say soft for the foundries but that does not get easy clicks. The sky is not falling. The semiconductor industry is very diverse and much more crash resistant than ever before, absolutely, so no, history will not repeat itself as easily this time.
Click to expand...

Agreed, for now.

For my education - when the world economy struggled in 2000, what exactly happened with the foundry businesses of the time? Did it experience a drop about a year later before recovering or did it never really drop?

That seems like an analog to today; IT organizations (and consumers) over-invested in software and equipment for a threat that never materialized, then there was a real slow down in IT equipment purchases for a couple of years after since everything was already modernized. I feel like Covid had some of that - people buying the latest and greatest for a giant shift away from the office. I feel like large scale corporations that buy a lot of silicon are going to shift back to a 4-5 year refresh cycle and have some ‘lean years’ after doing mass purchasing in 2020-2021. (Where consumers are unpredictable but also bought a bunch of ‘first time’ equipment in that same timeframe).

Daniel Nenni said:
What if TSMC targeted GF customers directly? Is there anything GF does that TSMC can't? And even if there is can GF survive based on just that? Would Samsung be in the foundry business without the support of their memory business? The same goes for Intel, no IDM business no foundry
Click to expand...

This is an avenue I had not considered; Thank you for that. TSMC could indeed go very aggressive against some of the other foundries, especially at that 28nm node they’re building up for where things are relatively portable.

The only devils advocate point I can think of is that it takes a few years to pivot designs - so while I’m sure they have designs that will ‘appear out of thin air’ that were competed a while ago, if demand for N5 and N7 suddenly dropped beyond expectations it would take some time to re-fill that capacity. I’m also not familiar with how locked in the client agreements are - which could make this a moot point for the next 6 quarters anyway..
 
Talking about the clickbait, that EE Time article is actually not that bad. At least it is not funny and silly enough compare to a recent hardwaretimes.com article written by its cofounder Areej (thanks to @lilo777 mentioned it in another thread). To avoid further spreading false information through search engine algorithm, I won't provide the exact link here. Instead I will attach the screenshot to this post.

The author claims Taiwan's semiconductor companies, including TSMC, are investing US$119 billion to build 20 fabs in China!

How can that happen under so many embargo, sanctions, and restrictions enforced by US, Taiwan, Japan, and European countries on China? It won't pass a simple fact check. It turned out those 20 fabs are being built in Taiwan, not China.

Screenshot_1-m.png
Screenshot_2.png
 
Last edited:
Daniel Nenni said:
I also heard at Semicon West that the equipment people are getting preferential treatment at TSMC right now to get chips for the equipment that they are waiting for. It's an ecosystem.
Click to expand...
Do you mean if equipment companies promise to prioritize equipment shipment to TSMC, they will receive chips to build equipment earlier?

I thought they buy chips from disty.
 
VCT said:
Do you mean if equipment companies promise to prioritize equipment shipment to TSMC, they will receive chips to build equipment earlier?

I thought they buy chips from disty.
Click to expand...

TSMC is probably just prioritizing which they commonly do for customers in need. Chips for equipment companies is very low volume. Automotive is an example of prioritizing. TSMC’s automotive content is up 14% in Q2 2022, trying to catch up from Covid delays.
 
Last edited:
  • Like
Reactions: VCT
You must log in or register to reply here.
Back
Top