Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/what-new-fabs-will-go-broke-first.18014/
        )

    [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
)

What new fabs will go broke first?

Arthur Hanson

Well-known member
With all the fab building and expansions going on it's easy to see supply outrunning demand. Also many fabs will not be able to produce the AI/ML chips or others that will be in demand. Any thoughts on what fabs under construction will run into trouble first? Also, what about the supply of skilled labor to man the new fabs? Even for the best new people, I see a couple of years before peak efficiency is achieved and many years of losses up front. Will we end up with a few new giga fabs and some economic money furnaces? Will like the real gold rush will the suppliers of the picks and shovels, the equipment suppliers be some of the biggest winners?
 
With all the fab building and expansions going on it's easy to see supply outrunning demand. Also many fabs will not be able to produce the AI/ML chips or others that will be in demand. Any thoughts on what fabs under construction will run into trouble first? Also, what about the supply of skilled labor to man the new fabs? Even for the best new people, I see a couple of years before peak efficiency is achieved and many years of losses up front. Will we end up with a few new giga fabs and some economic money furnaces?
TSMC only builds with customer commitments and management is truly the best in the industry. They will be fine. I think the real losers will be the memory guys and Samsung in particular. They just want to build build build and don’t have a client base or product to support it atm. Maybe things will change if TSMC stumbles repeatedly but that’s hard to imagine.
 
Empty fabs is not a new thing. Rather than going broke there will be fab shells without equipment. The biggest concern in my mind is staffing. The semiconductor industry needs to work on recruiting at the university level. The problem is you can't work from home so it will be tough to get the XYZ generation to show up. And yes I work from home now so who am I to talk. No way would I go back to a fab and wear a bunny suit all day. Sometimes I work from my sailboat inside the Golden Gate so no complaints here.
 
Empty fabs is not a new thing. Rather than going broke there will be fab shells without equipment. The biggest concern in my mind is staffing. The semiconductor industry needs to work on recruiting at the university level. The problem is you can't work from home so it will be tough to get the XYZ generation to show up. And yes I work from home now so who am I to talk. No way would I go back to a fab and wear a bunny suit all day. Sometimes I work from my sailboat inside the Golden Gate so no complaints here.

I believe there will be a clawback clause if a fab is built with some government subsidies but stays below the performance requirements, such as an empty shell or low utilization.
 
There’s a whole new, very diverse generation learning how to put the bunny suit on. They’re focused. I don’t foresee big issues there.

Fabs have been fully automated since about 2008, with the software refined enough so that humans weren’t needed for most lots to flow. Work from home is routine since 2020, for non-exempts. The fab isn’t exactly empty of people, but it’s not what old timers remember.
 
There’s a whole new, very diverse generation learning how to put the bunny suit on. They’re focused. I don’t foresee big issues there.

Fabs have been fully automated since about 2008, with the software refined enough so that humans weren’t needed for most lots to flow. Work from home is routine since 2020, for non-exempts. The fab isn’t exactly empty of people, but it’s not what old timers remember.

That is true. There are far fewer people inside a fab these days, especially a TSMC fab as they are highly automated. There are still quite a few people behind the scenes especially on the equipment side and operations. Still, people have to show up for fab jobs which is not the case for some industries. I do see more companies going back to the office here in Silicon Valley. The parking lots are filling up once again and traffic is stifling.
 
There’s a whole new, very diverse generation learning how to put the bunny suit on. They’re focused. I don’t foresee big issues there.

Fabs have been fully automated since about 2008, with the software refined enough so that humans weren’t needed for most lots to flow. Work from home is routine since 2020, for non-exempts. The fab isn’t exactly empty of people, but it’s not what old timers remember.
Also, most of the people who are working "at the fab" never have to wear a bunny suit. They are doing tool/yield/recipe analysis from their office. (plus management) This also means a lot of this can be done remotely.
 
I think hybrid is the future for fabs‘ exempts, with 2 days WFH, 3 days on site. It could go away, but it’s a costless, valuable benefit. It would hurt morale more than the benefit costs.

Some issues though: 1) Earning the same whether onsite or not, is if fair to those stuck onsite 2) The Elon Musk “hardcore” model with fewer people working very hard, onsite, in agile sprints, is probably more productive 3) The real test of hybrid WFH is when there is a downturn, which could be here now or soon
 
Back
Top