Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/intel-reportedly-sticking-with-tsmc-as-part-of-new-long-term-partnership-strategy-amid-underwhelming-18a-yields.22245/
        )

    [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] => 2021770
            [XFI] => 1050270
        )

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

Intel reportedly sticking with TSMC as part of new long-term partnership strategy amid underwhelming 18A yields

fansink

fansink

Well-known member
Intel is seemingly shifting its strategy of no longer outsourcing wafers to third parties to instead embracing a longer partnership with TSMC. This could be due to the unfavourable yields from Intel’s own 18A process.

It was recently reported that Intel may be pushing back its production of the next gen Panther Lake mobile processors due to unsatisfactory yields from the in-house 18A semiconductor manufacturing process. Now, according to the latest Intel earnings call, the company has announced a long-term partnership with TSMC for outsourcing wafers used for manufacturing its processors. The call, transcribed by Seeking Alpha, was between Intel’s John Pitzer and Morgan Stanley’s Joe Moore.

Intel has been struggling with bringing semiconductor manufacturing in-house as 18A, the process developed by Intel Foundry, has not presented favourable yields. TSMC, on the other hand, has a much more experienced and efficient 3nm and 2nm manufacturing process. According to the earnings call between Intel’s Corporate Vice President of Investor Relations and the MD, Semiconductor Industry Analyst at Morgan Stanley, Intel has been outsourcing about 30% of its wafers, most of it to TSMC, for a while now. However, till about a year ago, Intel wanted to reduce that number down to zero.

Now, it looks like Intel has a new strategy in place which brings about a longer partnership with TSMC. “We think it's always good to have at least some of our wafers with TSMC. They're a great supplier. It creates a good competition between them and Intel Foundry,” Pitzer reportedly added.

The earnings call and new strategy comes shortly after a known industry analyst, Ming Chi Kuo, shared that Intel’s Panther Lake mobile processors production has been pushed back due to the underwhelming performance of 18A process. This meant that the company would not be able to ship Panther Lake notebooks till 2026, missing out on 2025 Q4 sales, negatively impacting 2H25 revenue, profits, and supply chain trust.

The strategy shift may suggest that Intel has realized it is better to stick with TSMC for longer, but it is unclear how this will affect the Panther Lake processors that were supposed to be the first mobile SoCs developed on 18A.


 
I can see multiple reasons why this makes sense.

1. Intel has had a customer relationship with TSMC for decades. It might be that for the short to intermediate term TSMC will be able to make a lower bid for low volume chips (hundreds of thousands to low millions) than IFS can do.
2. Since industry IP is always targeted for TSMC first (it will be the de facto development fab environment for the IP industry the foreseeable future), Intel chips that use external IP may get to market quicker by using TSMC.
3. Intel's development team will want to have the option to use the best fab process for a given chip, even if the best is a TSMC process. Why give competitors like AMD an advantage just because IFS is in-house? Would a rational and non-emotional BoD want to hobble a product team just because IFS exists?
4. If you want to really benchmark the competition as your customers would see them, in a lot of ways, the best way to do it is be one of their competition's customers.
 
Last edited:
re: the Clearwater Forest delay reason from that transcript:

18A looks good. It really is an advanced packaging issue, where we’re doing hybrid bonding for the first time. The good news is we fixed that issue already, but it has caused a delay as to when we’re going to officially launch the product. But we do expect to ship samples in the second half of the year to customers. So hopefully your channel checks will start to hear some good things about Clearwater as we ship those samples.
 
blueone said:
I can see multiple reasons why this makes sense.

1. Intel has had a customer relationship with TSMC for decades. It might be that for the short to intermediate term TSMC will be able to make a lower bid for low volume chips (hundreds of thousands to low millions) than IFS can do.
2. Since industry IP is always targeted for TSMC first (it will the de facto development fab environment for the IP industry the foreseeable future), Intel chips that use external IP may get to market quicker by using TSMC.
3. Intel's development team will want to have the option to use the best fab process for a given chip, even if the best is a TSMC process. Why give competitors like AMD an advantage just because IFS is in-house? Would a rational and non-emotional BoD want to hobble a product team just because IFS exists?
4. If you want to really benchmark the competition as your customers would see them, in a lot of ways, the best way to do it is be one of their competition's customers.
Click to expand...
They need to put everything at Ifs To make it profitable not otherway around I agree they should choose the best process for product but for volume should be IFS.

NVL is already leaked except for Halo SKU everything is IFS mostly
 
I thought Pitzer's comment on Foundry break even was pretty clear. It is the transition from Intel7 to 18A that is going to drive most of the cost structure improvement that will get Intel Foundry to break even.

"I mean, we talked about sort of a target of giving Intel foundry to op profit breakeven by the end of twenty twenty seven. That’s mostly on the back of Intel products and actually mostly on the transition from moving our mix from Intel seven wafers to Intel 18A. As we’ve talked about in the past, as we make that transition, you will see ASP per wafer go up three times faster than cost per wafer.

And it’s really what’s driving us to a breakeven op profit point as we exit 2027."
 
Artificer60 said:
I thought Pitzer's comment on Foundry break even was pretty clear. It is the transition from Intel7 to 18A that is going to drive most of the cost structure improvement that will get Intel Foundry to break even.

"I mean, we talked about sort of a target of giving Intel foundry to op profit breakeven by the end of twenty twenty seven. That’s mostly on the back of Intel products and actually mostly on the transition from moving our mix from Intel seven wafers to Intel 18A. As we’ve talked about in the past, as we make that transition, you will see ASP per wafer go up three times faster than cost per wafer.

And it’s really what’s driving us to a breakeven op profit point as we exit 2027."
Click to expand...
Very clear. He said 18A is on track and reaffirmed the 2027 timeline for IFS to break even.
 
Xebec said:
re: the Clearwater Forest delay reason from that transcript:

18A looks good. It really is an advanced packaging issue, where we’re doing hybrid bonding for the first time. The good news is we fixed that issue already, but it has caused a delay as to when we’re going to officially launch the product. But we do expect to ship samples in the second half of the year to customers. So hopefully your channel checks will start to hear some good things about Clearwater as we ship those samples.
Click to expand...


It's strategic to blame anything other that 18a yields, such as some obscure packaging issue.

Bottom line, Pat said they'd be nearly weened off TSMC by now, but in typical Intel fashion:

As John Pitzer, Intel stated, "So I do think Michelle has been given a bit more agency to go off and make decisions around what she thinks is best for the product roadmap and for longer term market share there. And I think the biggest sort of manifestation of that today is she probably has more agency to use TSMC longer and for more than she might have had six or nine months ago. And to be clear, we have roughly 30% of our wafers outsourced today. That is probably a high watermark for us. But to the extent that I think a year ago we were talking about trying to get that to zero as quickly as possible, that’s no longer the strategy."

Who’ll be shocked if, a year from now, Intel’s outsourcing to TSMC hits 40%? Anyone? Bueller?
 
fansink said:
It's strategic to blame anything other that 18a yields, such as some obscure packaging issue.

Bottom line, Pat said they'd be nearly weened off TSMC by now, but in typical Intel fashion:

As John Pitzer, Intel stated, "So I do think Michelle has been given a bit more agency to go off and make decisions around what she thinks is best for the product roadmap and for longer term market share there. And I think the biggest sort of manifestation of that today is she probably has more agency to use TSMC longer and for more than she might have had six or nine months ago. And to be clear, we have roughly 30% of our wafers outsourced today. That is probably a high watermark for us. But to the extent that I think a year ago we were talking about trying to get that to zero as quickly as possible, that’s no longer the strategy."

Who’ll be shocked if, a year from now, Intel’s outsourcing to TSMC hits 40%? Anyone? Bueller?
Click to expand...
They said 15-20%. Given the short time frame - 1 year, it should not be that hard to 'sort the wheat from the chaff' to derive a prediction. But in 3, 5,10 years, no one knows. There are many external factors. If there is a geo-conflict, the outsource percentage could be dropped to 0%.
 
XYang2023 said:
They said 15-20%. Given the short time frame - 1 year, it should not be that hard to 'sort the wheat from the chaff' to derive a prediction. But in 3, 5,10 years, no one knows. There are many external factors. If there is a geo-conflict, the outsource percentage could be dropped to 0%.
Click to expand...


For nearly a decade, Intel’s words rarely match its actions, and ever fewer still blindly follow their lead. GL
 
fansink said:
For nearly a decade, Intel’s words rarely match its actions, and ever fewer still blindly follow their lead. GL
Click to expand...
10 year ago, 5 year ago, 1 year go, and even 3 month ago doesn't mean much. The company, meaning Intel Corporate, is a different company. IF is largely an independent unit. For IF, it is also important for them to adhere their expected timeline to earn trust for external customers.

Hence, the best prediction should be based on the most recent information not information sometime ago. In fact, basing decade old information can also be regarded blindly follow one's belief.

In the end, there is no right or wrong at this moment, what matters is what will actually happen in one year's time. What we are having now are just best estimates and various believes. The estimates and believes get adjusted as time goes.
 
fansink said:
It's strategic to blame anything other that 18a yields, such as some obscure packaging issue.

Bottom line, Pat said they'd be nearly weened off TSMC by now, but in typical Intel fashion:

As John Pitzer, Intel stated, "So I do think Michelle has been given a bit more agency to go off and make decisions around what she thinks is best for the product roadmap and for longer term market share there. And I think the biggest sort of manifestation of that today is she probably has more agency to use TSMC longer and for more than she might have had six or nine months ago. And to be clear, we have roughly 30% of our wafers outsourced today. That is probably a high watermark for us. But to the extent that I think a year ago we were talking about trying to get that to zero as quickly as possible, that’s no longer the strategy."

Who’ll be shocked if, a year from now, Intel’s outsourcing to TSMC hits 40%? Anyone? Bueller?
Click to expand...
40% don't think so as even 30% is very high for Intel themselves.

I would tell you that they will outsource to TSMC but not this scale they would definitely lower it to 15-20%.
 
Last edited:
Artificer60 said:
I thought Pitzer's comment on Foundry break even was pretty clear. It is the transition from Intel7 to 18A that is going to drive most of the cost structure improvement that will get Intel Foundry to break even.

"I mean, we talked about sort of a target of giving Intel foundry to op profit breakeven by the end of twenty twenty seven. That’s mostly on the back of Intel products and actually mostly on the transition from moving our mix from Intel seven wafers to Intel 18A. As we’ve talked about in the past, as we make that transition, you will see ASP per wafer go up three times faster than cost per wafer.

And it’s really what’s driving us to a breakeven op profit point as we exit 2027."
Click to expand...
This scenario doesn't have any external customer, it's just cutting Intel 7 from Intel Product.
 
XYang2023 said:
Should there be Intel 12 with UMC?
Click to expand...

That's a new addition, seems close to Intel 16. Maybe an expected small contribution?

Intel 7 seems was in Foundry plan in earlier self-assessment (below), now about to be gone like 20A, leaving a hole.

www.servethehome.com

Intel Foundry Operating Model Shown with Path to Process Leadership

Intel showed its financials as Intel Foundry and Intel Products while also pointing to a potentially better future
www.servethehome.com www.servethehome.com

Intel-Self-Assessment-on-18A-and-14A-Competitive-Versus-TSMC.jpg
 
You must log in or register to reply here.
Back
Top