Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-to-outsource-atom-xeon-based-socs-to-tsmc.13334/
        )

    [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
)

Intel to Outsource 'Atom & Xeon Based SoCs' to TSMC

Ah, LinkedIn, the ultimate spy tool. I mentioned this before, Intel will use TSMC for low cost and low power chips to make room for HPC silicon at the Intel fabs. Intel will save CAPEX and be on a level process playing field with AMD. Let the chip hunger games begin!

The question is: Will Intel manufacturing step up and give TSMC some decent competition or will Intel go very fablite?
 
Bob Swan said on CNBC that Intel will increase its outsourcing from the current 20% to 80% of capacity. Interesting to say the least.
 
Bob Swan said on CNBC that Intel will increase its outsourcing from the current 20% to 80% of capacity. Interesting to say the least.
Do you have a link for that? I can't find it. 80% seems very high based on what I have been hearing around the ecosystem.

Here is a cut/paste from the Bob Swan Q&A at the Credit Suisse 24th Annual Technology Conference (Transcript) on 12/01/2020 (Tuesday):

John Pitzer
Well, Bob, just my last question on the manufacturing front, before moving on to some of the product cycles. We know how the IDM business model works through the P&L. We also know how outsourcing non-CPU transistors work through the P&L because you're doing a lot of that today, 20-plus percent of your revenue stream today is outsourced. I'm kind of curious to the extent that this becomes an outsourcing of CPU transistors. How does that work its way through the P&L from a gross margin, op margin, free cash flow perspective? Is it a meaningful change from some of the longer term targets that you've discussed in the past?

Robert Swan
Well, I mean, first, as you said, we do a reasonable amount of leveraging third-party foundries today in the 15% to 20% range. And over time, I expect that with a more disaggregated design that for that to increase and give us more flexibility in both how to deliver products, but how to deliver differentiated economics within an IDM model, but in more flexible IDM model. But I think simply put the more we leverage a third-party foundry, number one, all else equal, gross margins come down a little bit.

Number two, all else equal the capital that we need to deploy for how volume manufacturing comes down a little bit and the effectiveness of return on capital, the challenge is making sure return on capital growth doesn't shrink in the evaluation of the magnitude of, if, when and what we leveraged third-party foundries. And that again is part of this evaluation, as we think about 2003, is what – how do those trade-offs manifest themselves in terms of market share, ASP capture and the cost of the product that we make or buy relative to the capital that we need to employ. So those are the inherent trade-offs that we try to manage to get to the right answer for 2023, 2024 timeframe.
 
Xeon is a crown jewel of Intel. Very likely they will pick somewhat dated dies, or low-end, or embedded for production in Taiwan.
 
Do you have a link for that? I can't find it. 80% seems very high based on what I have been hearing around the ecosystem.

Here is a cut/paste from the Bob Swan Q&A at the Credit Suisse 24th Annual Technology Conference (Transcript) on 12/01/2020 (Tuesday):

John Pitzer
Well, Bob, just my last question on the manufacturing front, before moving on to some of the product cycles. We know how the IDM business model works through the P&L. We also know how outsourcing non-CPU transistors work through the P&L because you're doing a lot of that today, 20-plus percent of your revenue stream today is outsourced. I'm kind of curious to the extent that this becomes an outsourcing of CPU transistors. How does that work its way through the P&L from a gross margin, op margin, free cash flow perspective? Is it a meaningful change from some of the longer term targets that you've discussed in the past?

Robert Swan
Well, I mean, first, as you said, we do a reasonable amount of leveraging third-party foundries today in the 15% to 20% range. And over time, I expect that with a more disaggregated design that for that to increase and give us more flexibility in both how to deliver products, but how to deliver differentiated economics within an IDM model, but in more flexible IDM model. But I think simply put the more we leverage a third-party foundry, number one, all else equal, gross margins come down a little bit.

Number two, all else equal the capital that we need to deploy for how volume manufacturing comes down a little bit and the effectiveness of return on capital, the challenge is making sure return on capital growth doesn't shrink in the evaluation of the magnitude of, if, when and what we leveraged third-party foundries. And that again is part of this evaluation, as we think about 2003, is what – how do those trade-offs manifest themselves in terms of market share, ASP capture and the cost of the product that we make or buy relative to the capital that we need to employ. So those are the inherent trade-offs that we try to manage to get to the right answer for 2023, 2024 timeframe. Even I was surprised by the levels stated.
Dan, I doubt there is a link for that, but I had my wife listen to it closely on DVR and explain the ramifications to her. I'm in the process of teaching her the market and critical announcements like this I have her listen to and then give me her view and what she thinks the ramifications will be. I want her to be able to be independent if anything did happen to me. So I listened to the presentation about three times in total and paused it after each pronouncement to go over it with my wife.
 
Tom's just updated the article saying the TSMC job posting has been removed. The Seoul jobs remain posted.

Swan has repeatedly stated "third-party foundries" so perhaps Samsung was always included in Intel's plans?
 
Atom, whatever but Xeon that's something.

10 nm products from Israel are probably the best in the world. There the only ones the can do scale at that scope now. America should but they would have to end some careers.

Two observation if an Intel product uses they same process as amd or mediatek why not buy a competitor. Second is shipping, Intel spends a lot on shipping and zim has a reputation for a high tech niche, shipping semiconductors is different than clothes. I think what's going on zim has had to restructure because of record demand and Intel has to use the same shipping as apple or their competitors.
 
I would not take that bet because it could certainly go that way. The same way as IBM.

IBM is a great comparison in many ways. IBM ruled the mainframe era in much the same way is Intel ruled the PC era. And IBMs stumbles in the PC business are analogous to Intel's stumbles in the mobile business. And then, once PC architectures started to invade the data center market, IBM couldn't compete. Now we have mobile architectures invading the PC market and Intel doesn't have a response. It's classic disruption playing exactly like you'd predict from the Christensen/Innovators Dilemma model.
 
Back
Top