Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/qualcomm-ceo-cristiano-amon-said-he-sees-an-opportunity-to-partner-with-intel-and-its-foundry-service.14351/
        )

    [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
)

Qualcomm CEO Cristiano Amon said he sees an opportunity to partner with Intel and its foundry service.

Daniel Nenni

Admin
Staff member
Yes, QCOM is on my list of possible Intel foundry customers now that Intel is out of the SoC and modem business. They still may compete in other smaller QCOM market segments however (ADAS and Datacenter). The problem is that QCOM uses multiple foundries to leverage pricing so I'm not sure how Intel can compete with Samsung or even TSMC. Also, QCOM uses leading edge processes for their high volume chips (SoC and Modems) and Intel is behind with 7nm and lacks the EUV throughput to accommodate QCOM. It would be a fair fight though because QCOM is one tough foundry customer, absolutely!

Certainly worth discussing though, thoughts?

 
I think it's a pity inquiry for a struggling CEO. We'll know more after the quarterly report but it isn't looking good.
 
>QCOM uses leading edge processes for their high volume chips

is Intel 10nm good enough?

Intel 10nm is comparable to TSMC 7nm so it's not bad. Once Intel has 10nm space in their fabs for foundry customers it will be several process nodes behind TSMC and Samsung. QCOM needs to be first to a new process to stay competitive with Mediatek and Apple. In fact, Mediatek has passed QCOM in the smartphone race and is now the #1 SoC chip company. Mediatek of course is close partners with TSMC while QCOM plays the foundry field looking for the best pricing. Go Mediatek!!!!
 
Once Intel has 10nm space in their fabs for foundry customers it will be several process nodes behind TSMC and Samsung.

I don't get why the CEO is pushing the foundry idea. It seems like work wasted.
 
I don't get why the CEO is pushing the foundry idea. It seems like work wasted.
Probably to get US Government money.

When Intel entered the foundry business 10+ years ago my first thought was that they would buy a big ASIC company to jumpstart the foundry business. This would allow Intel to have complete control over process technology, IP, and packaging ensuring secrecy. Intel is all about secrecy. My pick at the time was eSilicon which is now part of Marvel. Or maybe OpenSilicon which is now part of SiFive (OpenFive).
 
Intel quarterly revenue is projected to be $17 billion a downward trend. Govt money is nice but they there are priorities.

I don't buy the sifive takeover was for the foundry narrative. I'm no lead mechanic or test engineer but I do work around large automated machines. My guess is as nodes shrink machine learning and artificial intelligence will become essential and Intel is thinking ahead. CPU are a solid maker and Intel isn't going away. It may have a few down years coming but I think they'll re-emerge. Or maybe Intel will fail and disappear.
 
I was actually curious why Qualcomm CEO Cristiano Amon would make those comments. The reason being TSMC appears to be widening its process technology lead on Samsung. Based on MY interpretation of Scotten's graphs, it would appear Samsung's (still in development) 3nm GAAFET is roughly equivalent/slightly inferior to current TSMC's 5nm node. MediaTek is rumored to be unveiling its first 4nm chipset in Q4 2021 using TSMC's 4nm node. If that happens, the question may become whether Qualcomm can even afford to go back to Samsung Foundry now or in the future. There is no doubt that the US/China trade war has been hugely beneficial for MediaTek. MediaTek has eclipsed Qualcomm to be the top smartphone chipset vendor in 2020 (and most likely in 2021 as well). As MediaTek has grown, it's becoming a bigger and more important client for TSMC. Unlike Qualcomm, it has stayed with TSMC and most likely will get access to the latest nodes (behind Apple of course). This may force Qualcomm to continue using TSMC to match the performance/power/yield advantages the latest TSMC nodes offer over Samsung.

Unlike Nvidia, Qualcomm technology lead apparently isn't wide enough for Apple, Google, Xiaomi, Nokia, Oppo, etc. to not consider using MediaTek or their own custom-designed chips. Put simply, Qualcomm can't afford to lose its technological edge in processors and modems to MediaTek. By all accounts, MediaTek is behind Qualcomm but catching up fast. It doesn't appear to me that Qualcomm can afford to use Samsung Foundry for its high-end chips too much longer.

Full disclosure: I'm a shareholder of both Qualcomm and TSMC.
 
Intel 10nm is comparable to TSMC 7nm so it's not bad. Once Intel has 10nm space in their fabs for foundry customers it will be several process nodes behind TSMC and Samsung. QCOM needs to be first to a new process to stay competitive with Mediatek and Apple. In fact, Mediatek has passed QCOM in the smartphone race and is now the #1 SoC chip company. Mediatek of course is close partners with TSMC while QCOM plays the foundry field looking for the best pricing. Go Mediatek!!!!
It's a very strange world indeed when Samsung A32 phone uses Mediatek. Oh the conversations that must have been had at Samsung senior level.
A shot across the bow of Qualcomm, who has been neglecting the low end chips.
Can I see Intel coming in with a competitive 22nm Finfet foundry process for low end smartphone chips? Yes I can. Would it help differentiate Qualcomm at the low end, vs. Mediatek? Yes it would.
 
Back
Top