Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/morris-chang-globalization-is-almost-dead-and-free-trade-is-almost-dead-speech.17215/
        )

    [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
)

Morris Chang "Globalization is almost dead and free trade is almost dead" Speech

Daniel Nenni

Admin
Staff member
TSMC Arizona Celebration.png


Morris Chang's speech at the TSMC Arizona fab opening ceremony last week:

When I started TSMC back in 1987, I had a dream. Probably because of my background, which up to that point, was primarily America. Probably because of my background, my dream was to build fabs in the United States.

So, eight years from our start up, we started in 1987, and in 1995, we broke ground in a town called Camas, which is in the state of Washington, just on the border of Oregon, In fact, it's very close to Portland, Oregon. We called it Wafertech. It was a well-conceived semiconductor factory. Its technology was completely up to date at that time. It was, I thought, a dream fulfilled.

However, we ran into cost problems. We ran into people problems. We ran into cultural problems. And before long, the dream fulfilled became a nightmare fulfilled. It took up several years to untangle ourselves from the nightmare. And I decided that I needed to postpone the dream, just postpone, just postpone it.

25 years passed and we have a new chairman, Mark Liu. And he happens to share my dream! Now you see a partial, not yet fulfilled, but progress of this dream. Incidentally, this ceremony was called the first 'tool-in'. Nobody outside Taiwan understood what that meant, so now it’s called 'opening ceremony'. It is a Taiwan custom. I did not understand what the name meant at first, but now, I’ve heard it so many times, I’ve been at 'first tool-in' ceremonies so many times, I now understand what it means. Basically it means, the end of the beginning for a factory.

The romance of the beginning is gone! The initial excitement is gone! A lot of hard work remains, a lot of hard work. The twenty-some years in the past have witnessed a big change in the world. A big geopolitical situation. Globalization is almost dead. Free trade is almost dead. And a lot of people still wish they would come back, but I really don’t think they will be back for a while.

In the meantime, because of the change in political situation, the new dream, well…it’s the old dream revived, has the help of the US government, the federal government, the state government, the local government. Not only that, we did learn from our experience earlier, and we are far more prepared now. And we hired almost 600 engineers here a year and a half ago, we sent them to Taiwan, and they were under training in Taiwan for one year to a year and a half. In the meantime, about the same number of Taiwan engineers underwent training in Taiwan also.

So before we see a single wafer, we have more than a thousand people being trained. This I think is a very good sign that we are prepared. It’s a very good sign that my dream of 25 years ago will now be fulfilled by Mark. Anyway I came here specifically to see the end of the beginning and to wish TSMC the best, in the full expectation that we are going to have success. And it will be a very meaningful success as Mark has just said. And my dream lives. Thank you very much.
 
I like Morris. He gets things done and says it like it is. No bull.

Too bad Ronald Lee Ermey passed away. We would have made a great trainer.
 
Well, that's a start, at least now Morris knows which state Camas is in. I'm not sure I would consider a facility 41 miles from Hillsboro, Oregon, where most of the existing semiconductor workforce lives, situated in a small town with a population (back then) of less than 25,000, and is only accessible by a two-lane road on the Washington side of the Columbia river, or a narrow bridge from the interstate highway on the Oregon side, well-conceived.

And I wonder... just looking at a map this morning, I see that it is a 50 mile, 1 hour 10 minute commute from Chandler to the TSMC construction site. Interesting. At least it's on an interstate highway this time. Or is there a semiconductor workforce he can hire from living in northern Phoenix?
 
In Taiwan towns are formed around the fabs in pretty close proximity. Maybe that will happen in AZ? The last time I visited the TSMC fabs in China they were also pretty isolated. Maybe that has changed.

My first fab job was not far from where I lived. As it turns out that was a bad thing since the ground water was tainted. We moved 50 miles away and I drove the 1-2 hours to Silicon Valley 5 days a week for 30+ years without question.
 
In Taiwan towns are formed around the fabs in pretty close proximity. Maybe that will happen in AZ? The last time I visited the TSMC fabs in China they were also pretty isolated. Maybe that has changed.

My first fab job was not far from where I lived. As it turns out that was a bad thing since the ground water was tainted. We moved 50 miles away and I drove the 1-2 hours to Silicon Valley 5 days a week for 30+ years without question.
Perhaps I'm over-sensitive to long commutes. Early in my career, for over four years, I had a 52 mile, each way, commute in Ohio. It was very annoying, even though it was all interstate highway. Admittedly, Ohio has cold and snowy/icy winters, and Phoenix doesn't, and those winters had me staying overnight a couple of times in my office. But Phoenix traffic is far worse than anything was in Ohio in the 1980s.

I know a lot of people in The Valley with 1-2 hour commutes. San Diego too.
 
In Taiwan towns are formed around the fabs in pretty close proximity. Maybe that will happen in AZ? The last time I visited the TSMC fabs in China they were also pretty isolated. Maybe that has changed.

My first fab job was not far from where I lived. As it turns out that was a bad thing since the ground water was tainted. We moved 50 miles away and I drove the 1-2 hours to Silicon Valley 5 days a week for 30+ years without question.
Yeah I wouldn't blame you with the environmental condition in the valley.
 
Reading this speech makes me reflect on why Global Foundries gave up.

Did their top management lack the same 'dream'? Were their ambitions that different?
 
They ran out of money and had no customers on the leading edge. They needed extra money from the oil barons just to build 15k WSPM, with their only customer being IBM. Not only was it the right choice to bail from the Moore's law rat race, but just like with AMD going fabless it was the only choice. Which is a real shame. I (and just about everyone besides Samsung and TSMC upper management) would love if the leading edge had 3 foundry players (and hopefully 4 if intel can pull their gambit off).
 
| They ran out of money and had no customers on the leading edge.

AMD, IBM & Qualcomm were serious customers and really wanted GF to succeed at 7nm. (Lisa Su visited GF and was very clear about this.) But -- GF's investor ran out of patience and refused to fund at some point. There is a reason GF changed their CEO ~2017. Almost immediately, GF started to pull back on a number of initiatives. (they built a new fab in China ... the building was complete ... canceled just before installing fab tools) From an outsiders' perspective -- I think GF was concerned they would not get through that period.

The chip shortage & government motivation has really helped GF.
 
| They ran out of money and had no customers on the leading edge.

AMD, IBM & Qualcomm were serious customers and really wanted GF to succeed at 7nm. (Lisa Su visited GF and was very clear about this.) But -- GF's investor ran out of patience and refused to fund at some point. There is a reason GF changed their CEO ~2017. Almost immediately, GF started to pull back on a number of initiatives. (they built a new fab in China ... the building was complete ... canceled just before installing fab tools) From an outsiders' perspective -- I think GF was concerned they would not get through that period.

The chip shortage & government motivation has really helped GF.
I doubt the AMD line. These wafer agreements are signed years in advance. GF would have been asking for more than a measly 15k if Qualcomm and AMD were going to be throwing them anything more than small volumes as a second and in the case of Qualcomm third source supplier. Of course just my observations though. If you've heard straight from the any of the horses' mouths than you would be far more well informed than my speculation.
 
| They ran out of money and had no customers on the leading edge.

AMD, IBM & Qualcomm were serious customers and really wanted GF to succeed at 7nm. (Lisa Su visited GF and was very clear about this.) But -- GF's investor ran out of patience and refused to fund at some point. There is a reason GF changed their CEO ~2017. Almost immediately, GF started to pull back on a number of initiatives. (they built a new fab in China ... the building was complete ... canceled just before installing fab tools) From an outsiders' perspective -- I think GF was concerned they would not get through that period.

The chip shortage & government motivation has really helped GF.
If they were really serious wouldn't they have put up some guarantees?

Such as precommitment to buy X wafers at Y price, assuming Z conditions.

And if they did... why would GF investors turn down guaranteed profits?
 
Last edited:
And if they did... why would GF investors turn down guaranteed profits?
Even when GF was the only supplier for AMD, IBM, and did business with other firms GF was hemorrhaging money. Why spend a boatload of money finishing, ramping, and supporting 7nm when your best case scenario was still losing money. Specialty nodes allow GF to charge more for their nodes that they already made, and with a relatively modest investment. From a business perspective it was the obvious move to ditch 7nm even if they had a sizable customer base (and my understanding is they didn’t even have that). Real shame I would have loved to see how gf7 evolved, and would have loved to see gf3 gaa (like Samsung they were skipping 5nm for 3nm).
 
What's wrong with GF14?
Nothing. Doesn’t change the fact that the original GF 14nm was cancelled, and that as a company GF was still losing money even with their licensed Samsung node. The way I see it GF just didn’t have enough scale and at the time their tech wasn’t good enough to justify a premium over cheaper higher capacity TSMC and Samsung nodes that came to market before GF’s own offerings.
 
If they were really serious wouldn't they have put up some guarantees?

Such as precommitment to buy X wafers at Y price, assuming Z conditions.

And if they did... why would GF investors turn down guaranteed profits?

IBM actually did commit to 7nm GF. (paying $1.5B during the spinoff). But GF still canceled. This is why IBM is suing GF.

Give Lisa Su a ton of credit for having a backup plan and quickly jumping to TSMC. This is her best move as CEO. She didn't bet her company's future on GF.

This isn't simple stuff. There is a reason Intel delayed & delayed 10nm and 7nm.
 
And I wonder... just looking at a map this morning, I see that it is a 50 mile, 1 hour 10 minute commute from Chandler to the TSMC construction site. Interesting. At least it's on an interstate highway this time. Or is there a semiconductor workforce he can hire from living in northern Phoenix?
What makes you think someone lives in Chandler and will commute to TSMC long-term? Intel employees living right next to the Ocotillo fab who jump ship? (TSMC is 45-50 minutes from there, not 1hr 10min; if you're looking for distances from downtown Chandler that's probably not a good reference point.)

It seems just as likely if someone has ties to a specific home, they live in Phoenix or Tempe, in which case it's not as far. Or they don't have ties to a specific home, and are willing to move. North Phoenix is nice, closer to Flagstaff & Sedona, just a few degrees cooler because of the increase in altitude.
 
IBM actually did commit to 7nm GF. (paying $1.5B during the spinoff). But GF still canceled. This is why IBM is suing GF.

Give Lisa Su a ton of credit for having a backup plan and quickly jumping to TSMC. This is her best move as CEO. She didn't bet her company's future on GF.

This isn't simple stuff. There is a reason Intel delayed & delayed 10nm and 7nm.

I won't count that IBM's $1.5 billion payment to Globalfoundries totally as a prepayment for 7nm. At that time IBM can't find a suitable buyer for its semiconductor manufacturing division. Several companies, including TSMC, decided to skip the opportunity. It was that bad.

The $1.5 billion was a bonus IBM put into the deal to convince Globalfoundries to take over a division that IBM had no intention to keep.
 
Dear GlobalFoundries:

Congratulations figuring out that triple/quadruple patterning wasn't worth it so early in the process. Please consider adding backside connections for the power supplies in your 14nm process, and put effort into interposers. I suspect this will be very popular.
 
What makes you think someone lives in Chandler and will commute to TSMC long-term? Intel employees living right next to the Ocotillo fab who jump ship?
If you're looking for experienced engineers, that is exactly what I was thinking. Chandler and the surrounding communities are where most of the Intel AZ people I know live.
(TSMC is 45-50 minutes from there, not 1hr 10min; if you're looking for distances from downtown Chandler that's probably not a good reference point.)
I did several map plots from various residential areas in the south, and during the morning rush hour (7:30am) I noticed most of the routes were about 1hr 10min, mostly due to bottlenecks heading north, like I-17.
It seems just as likely if someone has ties to a specific home, they live in Phoenix or Tempe, in which case it's not as far. Or they don't have ties to a specific home, and are willing to move. North Phoenix is nice, closer to Flagstaff & Sedona, just a few degrees cooler because of the increase in altitude.
For highly mobile employees, like those who rent homes or apartments but don't have children in school or spouses who also work near the current residence, I agree. There's also the matter of whether or not TSMC will be looking for experienced people, or looking mostly for recent college grads and those early in their careers. Prior to about the year 2000, many established US high-tech companies preferred to hire RCGs, and that might very well be part of TSMC's culture (I have no information), which would much reduce my commute questions.
 
Back
Top