Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-to-manufacture-auto-chips.14053/
        )

    [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] => 2020570
            [XFI] => 1050070
        )

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

Intel to manufacture auto chips?

count

Active member

Does this seem realistic to anyone here? It's not so easy to port a design over and start pumping out chips on a totally different process in 6-9 months. Nor is it easy to reconfigure a production line. This seems like a PR stunt to me.
 

hist78

Active member
It might be very difficult for those existing products but "might" have a slim chance for new products. And it depends on how you define "start producing" and what kind of product it is.

Also I'm not sure if any major auto parts suppliers dare to cut their order quantity at foundries and move it to Intel at this moment. Any foundries' customers will be very happy see they walk away from the waiting line.
 
Last edited:

Daniel Nenni

Admin
Staff member

Does this seem realistic to anyone here? It's not so easy to port a design over and start pumping out chips on a totally different process in 6-9 months. Nor is it easy to reconfigure a production line. This seems like a PR stunt to me.

No porting because most automotive designs are not FinFETs. Intel will be pushing their 22nm or even 14nm processes and those are FinFETs. Getting new automotive designs is certainly possible, even probable, once the Intel processes are automotive ISO 26262 certified. So let's say HVM automotive designs optimistically in 2025?
 

Portland

Active member
Intel is spreading themselves thin and burning out their employees in an unsustainable process as companies like nvidia are encroaching on their server market.
 

Daniel Nenni

Admin
Staff member
Intel is spreading themselves thin and burning out their employees in an unsustainable process as companies like nvidia are encroaching on their server market.

In general yes but in this case Intel is focusing foundry business on specific market segments which is smart. Intel is obviously chasing Chips ACT money so that may be a worthy vertical to chase. Unfortunately, any automotive foundry business revenue is very far out. Much farther out than the life span of an Intel CEO.
 

count

Active member
No porting because most automotive designs are not FinFETs. Intel will be pushing their 22nm or even 14nm processes and those are FinFETs. Getting new automotive designs is certainly possible, even probable, once the Intel processes are automotive ISO 26262 certified. So let's say HVM automotive designs optimistically in 2025?
2025 is a realistic timeline. Intel is out there saying they will produce auto chips in 6-9 months which to me smells like expired baloney.
 

Daniel Nenni

Admin
Staff member
2025 is a realistic timeline. Intel is out there saying they will produce auto chips in 6-9 months which to me smells like expired baloney.

Agreed, and Intel's expired baloney will stink up their foundry business, which requires absolute trust, absolutely.
 

Jozo035

Member
No porting because most automotive designs are not FinFETs. Intel will be pushing their 22nm or even 14nm processes and those are FinFETs. Getting new automotive designs is certainly possible, even probable, once the Intel processes are automotive ISO 26262 certified. So let's say HVM automotive designs optimistically in 2025?
Well, these days even FEM, gateway... modules (used to control seat position, mirrors, windows...) uses 16/12 nm silicon. Infotainment, ADAS... needs 10/7nm, because people hates "laggy screens"...

Actually, automotive suppliers (TI, ST, NXP...) are in transition to 5nm...

And this is in my opinion something, lot of people missing here... Automotive is not something where Intel can just dump pieces of old unutilized capacity. It is actually very competitive area, technologically advanced, high volume, high reliability...
 

Daniel Nenni

Admin
Staff member
Well, these days even FEM, gateway... modules (used to control seat position, mirrors, windows...) uses 16/12 nm silicon. Infotainment, ADAS... needs 10/7nm, because people hates "laggy screens"...

Actually, automotive suppliers (TI, ST, NXP...) are in transition to 5nm...

And this is in my opinion something, lot of people missing here... Automotive is not something where Intel can just dump pieces of old unutilized capacity. It is actually very competitive area, technologically advanced, high volume, high reliability...

You are right, thanks. TSMC has done a good job of auto qualifying FinFET processes. I know TSMC has qualified down to 7nm for ISO26262 and AECQ-100 but I have not heard that about 5nm.

Certainly ADAS is FinFET, probably at the leading edge though so probably not a great place for Intel to start the automotive qualification.

Price is also an issue with automotive which is one of the reasons why the 200mm fabs are doing so well at the trailing edge. I know TSMC has qualified down to 7nm for ISO26262 and AECQ-100 but I have not heard that about 5nm.
 
Top