Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/moor-insights-%E2%80%9C-if-you-spin-out-ifs-before-design-ifs-is-healthy-it-will-fail%E2%80%9D.20928/
        )

    [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
)

Moor Insights: “[if] you spin out IFS before Design & IFS is healthy, it will fail”

B

Brady

Member
“Let me be crystal freaking clear: you spin IFS out before the Design company and IFS is healthy, it will fail. The only caveat here is taking it private…”

”Well why do you say that: … I ran corporate marketing at AMD while spinning off Globalfoundries… talk about two sick puppies… AMD was locked into long-term wafer agreements (that they never kept)… they had to do that to keep Globalfoundries going… the only reason that didn’t fail was because Mudabala who was bankrolling GF and internalized all the losses”.

100% agree. In another thread a “Plan B” for Intel was discussed. Right now, there is no Plan B that is successful as a leading edge foundry. It’s only “Globalfoundries, Intel Edition”.

At ~8:15
 
What's stopping Samsung or TSMC from just going after the remainder of GloFo's business, as innovation has greatly slowed at GF? (i.e. is GF a dead man walking or could they still be a viable business in 10 years time?)
 
samwilde said:
The only reason Mudabala sank all that money into Globalfoundries was that they originally thought they would be getting a fab in Abu Dhabi. We know how well that went. Do not expect a sugar daddy like that to show up for Intel.
Click to expand...
The only possible sugar daddy I can think of is Japanese government: Combine Rapidus with Intel’s foundry
 
Brady said:
“Let me be crystal freaking clear: you spin IFS out before the Design company and IFS is healthy, it will fail. The only caveat here is taking it private…”

”Well why do you say that: … I ran corporate marketing at AMD while spinning off Globalfoundries… talk about two sick puppies… AMD was locked into long-term wafer agreements (that they never kept)… they had to do that to keep Globalfoundries going… the only reason that didn’t fail was because Mudabala who was bankrolling GF and internalized all the losses”.

100% agree. In another thread a “Plan B” for Intel was discussed. Right now, there is no Plan B that is successful as a leading edge foundry. It’s only “Globalfoundries, Intel Edition”.

At ~8:15
Click to expand...

I follow their YouTube channel, and I often feel that both of them are too optimistic about Intel and trust what Intel says. This suddenly reminds me of the realtor who helped us get our first house many years ago. She warned us back then that we can like a house, but we shouldn't fall in love with it.

Love can be very dangerous when investing and evaluating business strategies.
 
hist78 said:
I'm wondering why Intel and Japanese government didn't do it to begin with. Instead Rapidus went to IBM, a more risky choice.
Click to expand...
Isn’t that strictly because IBM had a 2nm process (and established process R&D at Albany Nanotech) that was a jumping off point?

hist78 said:
I follow their YouTube channel, and I often feel that both of them are too optimistic about Intel and trust what Intel says. This suddenly reminds me of the realtor who helped us get our first house many years ago. She warned us back then that we can like a house, but we shouldn't fall in love with it.

Love can be very dangerous when investing and evaluating business strategies.
Click to expand...
Surely.

That being said, I feel it’s necessary to listen to different perspectives from people with experience. I felt this take is a good counter balance to what I perceive as the prevailing “knives out” takes regarding Intel right now.

I personally agree with this take because as I’ve said in other threads the turnaround of Intel was never projected to be complete in 2024/2025. These years were going intended to be the inflection (better known at the bottom…).

One thing I find interesting reading this forum is seeing so many industry insiders talking about how the foundry industry isn’t for the faint of heart, but also want Intel to give up while they’re doing the hard work of trying to make it happen.

Whom is faint of heart in this case?
 
Brady said:
Isn’t that strictly because IBM had a 2nm process (and established process R&D at Albany Nanotech) that was a jumping off point?


Surely.

That being said, I feel it’s necessary to listen to different perspectives from people with experience. I felt this take is a good counter balance to what I perceive as the prevailing “knives out” takes regarding Intel right now.

I personally agree with this take because as I’ve said in other threads the turnaround of Intel was never projected to be complete in 2024/2025. These years were going intended to be the inflection (better known at the bottom…).

One thing I find interesting reading this forum is seeing so many industry insiders talking about how the foundry industry isn’t for the faint of heart, but also want Intel to give up while they’re doing the hard work of trying to make it happen.

Whom is faint of heart in this case?
Click to expand...
I think the concern that some people are raising is that Intel does not have the technological and manufacturing knowhow to be a foundry and that they are losing so much money that they won’t have the financial resources to pull it off. Were it not for geopolitical concerns, the easiest strategy would have been to sell the technology to SMIC and rely on PRC resources (financial and people) to create a true competitor to TSMC.
 
vslee said:
US gov't likely wouldn't let the Japanese take over Intel anyways. 7-Eleven yes, Intel no.
Click to expand...
It might be far more cost effective for the US taxpayer though.

After all, Japan has a far better success record of government backing for industry than the US does. And a pretty good record in manufactuing and fabs. And the stomach for this sort of all-in long term investment (generally lacking in the US and Europe).

It's a left field suggestion, but certainly an interesting one. And at a time when the range of options on the table for Intel seems to be expanding rather than shrinking (which in itself is not usually a good sign - if Intel is really through its corporate turnaround, the options would be shrinking). If it came to a choice between IFS folding, merging with GloFo, ending up with Samsung or being run in Japan, Japan seems like the least worst option to me.
 
osnium said:
I think the concern that some people are raising is that Intel does not have the technological and manufacturing knowhow to be a foundry and that they are losing so much money that they won’t have the financial resources to pull it off. Were it not for geopolitical concerns, the easiest strategy would have been to sell the technology to SMIC and rely on PRC resources (financial and people) to create a true competitor to TSMC.
Click to expand...
Easy strategy, and the exact outsourcing that yielded the end of the US’s manufacturing capabilities.

You do get to the heart of it though: to bail out of foundry now (and splitting foundry out right now is in effect bailing, IMO) is the easiest way for Wall Street to make money. For sure.
 
Brady said:
Easy strategy, and the exact outsourcing that yielded the end of the US’s manufacturing capabilities.

You do get to the heart of it though: to bail out of foundry now (and splitting foundry out right now is in effect bailing, IMO) is the easiest way for Wall Street to make money. For sure.
Click to expand...
It’s how the US’ economic and political structure is setup now. Allow other countries to do hard work that requires competence and commitment that is not present in the US. Rather than trying to swim against currents, best to go with the flow and shutting down the fabs and selling the technology to China or Japan is the easiest way for the owners of the company to recoup their investment. The US based employees may not like it but they are not and should not be a consideration in this matter. They did not execute in 10nm and they need to suffer the consequences.
 
tooLongInEDA said:
It might be far more cost effective for the US taxpayer though.

After all, Japan has a far better success record of government backing for industry than the US does. And a pretty good record in manufactuing and fabs. And the stomach for this sort of all-in long term investment (generally lacking in the US and Europe).

It's a left field suggestion, but certainly an interesting one. And at a time when the range of options on the table for Intel seems to be expanding rather than shrinking (which in itself is not usually a good sign - if Intel is really through its corporate turnaround, the options would be shrinking). If it came to a choice between IFS folding, merging with GloFo, ending up with Samsung or being run in Japan, Japan seems like the least worst option to me.
Click to expand...
As I was looking at how the Japaneses subsidized their steel industry it became pretty clear to me that their approach is to lower the tax burden on the industry they are supporting thus making more funds available for a successful industry player to reinvest. The beauty of the Japanese system is that it rewards investments in the growth of the industry while self selecting specific businesses that are successful. If a business isn't turning a profit, they don't have funds to reinvest and take advantage of the tax incentives.

Unfortunately, the focus on tax incentives makes the Japanese approach a non-starter in the US despite the fact that it has been proven to work very well. The narrative in the US is that we need to tax the greedy corporations to ensure they pair their "fair" share. The result is the only real support that the government can give an industry is to throw money at it which doesn't inherently require an industry player to strive to be successful to see the benefit of the government support.
 
osnium said:
It’s how the US’ economic and political structure is setup now. Allow other countries to do hard work that requires competence and commitment that is not present in the US. Rather than trying to swim against currents, best to go with the flow and shutting down the fabs and selling the technology to China or Japan is the easiest way for the owners of the company to recoup their investment. The US based employees may not like it but they are not and should not be a consideration in this matter. They did not execute in 10nm and they need to suffer the consequences.
Click to expand...
The issue with 10nm was not the fault of the employees failing to execute. Intel missed the 22nm launch by about a quarter, They missed the 14nm launch by about 2 quarters. Management decided that they needed to make up this gap and designed 10nm to scale much more aggressively than the traditional 2x scaling (essentially trying to skip a node) which is why Intel 10 was comparable to TSMC 7nm when it launched. Skipping a node is a proven recipe for disaster as you don't get the benefits of incremental learning that come from the traditional cadence. But Intel, being Intel, couldn't learn from their past problems and decided to try and bite off even more despite the challenges they saw on their earlier process node. To add insult to injury, the decision was made to do this without the use of EUV. All of these were management decisions, not a failure to execute by the rank and file employee. The employees were given a task that couldn't be achieved in the timeline management set. Yet you want to hold them responsible for the failure of leadership?
 
osnium said:
You might have had better leadership on the foundry side but Tower’s newest node was 45 nm so there might not have been much benefit from the pure technology side.
Click to expand...
I think Intel's challenges arise from a lack of foundry knowledge/leadership, not the technology side. As a result I'm inclined to think that Intel would be managing the transition better. Not to mention another established income stream couldn't hurt given Intel's current financial issues.
 
Artificer60 said:
The issue with 10nm was not the fault of the employees failing to execute. Intel missed the 22nm launch by about a quarter, They missed the 14nm launch by about 2 quarters. Management decided that they needed to make up this gap and designed 10nm to scale much more aggressively than the traditional 2x scaling (essentially trying to skip a node) which is why Intel 10 was comparable to TSMC 7nm when it launched. Skipping a node is a proven recipe for disaster as you don't get the benefits of incremental learning that come from the traditional cadence. But Intel, being Intel, couldn't learn from their past problems and decided to try and bite off even more despite the challenges they saw on their earlier process node. To add insult to injury, the decision was made to do this without the use of EUV. All of these were management decisions, not a failure to execute by the rank and file employee. The employees were given a task that couldn't be achieved in the timeline management set. Yet you want to hold them responsible for the failure of leadership?
Click to expand...
Management also comes under US based employees and I think it’s the collective failure of the entire group. I do agree that management is probably more to blame on an individual basis. The broader point remains that there doesn’t seem to be much of a market for a foundry that struggles with delivering on time.
 
You must log in or register to reply here.
Back
Top