Unintended consequences of sanctions

[Fred Chen]

Unintended consequences of sanctions - Econlib

In recent months, the US government has tried to make it more difficult for China to access cutting edge chip technology from American firms, as well as from our allies. Back in March, Ben Thompson discussed one consequence of delinking from China: This point applies to semiconductors broadly...

www.econlib.org

Unintended consequences of sanctions​

By: Scott Sumner
In recent months, the US government has tried to make it more difficult for China to access cutting edge chip technology from American firms, as well as from our allies. Back in March, Ben Thompson discussed one consequence of delinking from China:

This point applies to semiconductors broadly: as long as China needs U.S. technology or TSMC manufacturing, it is heavily incentivized to not take action against Taiwan; when and if China develops its own technology, whether now or many years from now, that deterrence is no longer a factor.

In October, Thompson expanded upon that warning:

China meanwhile, has had good reason to keep TSMC around, even as it built up its own trailing edge fabs: the country needs cutting edge chips, and TSMC makes them. However, if those chips are cut off, then what use is TSMC to China? . . . the more that China builds up its chip capabilities — even if that is only at trailing nodes — the more motivation there is to make TSMC a target, not only to deny the U.S. its advanced capabilities, but also the basic chips that are more integral to everyday life than we ever realized.

Bloomberg points out that the US is heavily reliant on Taiwan’s chips, and that even the new TSMC factory being built in Phoenix will not change that reality:

TSMC’s Arizona fabs will produce 600,000 wafers annually. That sounds impressive, but it’s really not. The Taiwanese company topped 14.2 million last year and is on track to churn out 15.4 million 12-inch wafers loaded with chips this year. If it keeps the same average capacity growth of 8.1% it achieved over the past five years, Arizona will account for just 2.85% of its 21 million annual global output in 2026.
That’s a drop in the bucket, not a game changer.

In my view, this exercise in industrial policy merely papers over the fundamental problem, which is a lack of talent:

TSMC founder Morris Chang has poured cold water on the US ability to compete.
“There’s a lack of manufacturing talents to begin with,” Chang told The Brookings Institution earlier this year. “We did it at the urging of the US government, and we felt that we should do it.”

If the government were serious about competing with China they’d be changing our immigration laws to attract more engineering talent, not building white elephants in the Arizona desert:

Research, development, planning and operations will all remain in Taiwan. Should Beijing decide to attack, those functions will cease at least temporarily, if not permanently. This would mean cutting TSMC Arizona off from all the crucial know-how it needs to run the minuscule capacity it has on US soil.

 

[nghanayem]

I generally agree, but a few critiques. Semis are important don't get me wrong, but it is a minor economic inconvenience compared to the disaster that would be caused from the West placing sanctions even a fifth as draconian was what was put on Russia. For the PRC develops working EUV tool 20 years almost seems optimistic, given they don't even have functional dry DUV tools more than a decade since their release despite having numerous examples to reverse engineer. And so let's say after 20 years they do it. Hooray! Now we just need another decade to get to hyper-NA with DSA. Oh and by the way we still need to develop a parallel EDA/IP ecosystem from the rest of the world. Even with the sanctions I see TSMC's N-2 being more practical than SMIC's best for many years to come.

As for the talent I don't totally buy it. Many of TSMC's upper echelons and researchers were and continue to be educated in the US. Half of intel's fabs and all of their node and packaging R&D is in the US. Micron has a substantial footprint and all of their R&D in the US. And there is also the swarm of trailing edge fabs from Samsung, to TI, to GF, and Skywater. The US's population is also 14.5x that of the ROC. Granted the US economy is much more service based than the ROC; there is still a very large manufacturing and DOD veteran talent pool to draw skilled technicians from, as well as far more universities to spew out engineers. If the statement was "at Samsung Korea wages there is no engineering talent in the US", then I'd agree fully. Unlike in the ROC or ROK, TSMC and the Chaebol are not the only employers in the US and as such they must compete for the brightest scientist and engineers.

I sincerely hope Micron and Intel continue to execute and prove to the naysayers that American engineers and technicians are just as intelligent and capable than their Asian counterparts, as well as providing further evidence that Americans are still capable of manufacturing excellence (as if the US chemicals industry or American/Foreign owned auto factories in the US don't already show this).

 

[Fred Chen]

I think Morris was referring to the lack of an entity like TSMC in the US, so the heavy dependence on TSMC in Taiwan is felt not only by China but also by the US.

 

[hist78]

Unintended consequences of sanctions - Econlib

In recent months, the US government has tried to make it more difficult for China to access cutting edge chip technology from American firms, as well as from our allies. Back in March, Ben Thompson discussed one consequence of delinking from China: This point applies to semiconductors broadly...

www.econlib.org

Unintended consequences of sanctions​

By: Scott Sumner
In recent months, the US government has tried to make it more difficult for China to access cutting edge chip technology from American firms, as well as from our allies. Back in March, Ben Thompson discussed one consequence of delinking from China:

In October, Thompson expanded upon that warning:

Bloomberg points out that the US is heavily reliant on Taiwan’s chips, and that even the new TSMC factory being built in Phoenix will not change that reality:

In my view, this exercise in industrial policy merely papers over the fundamental problem, which is a lack of talent:

If the government were serious about competing with China they’d be changing our immigration laws to attract more engineering talent, not building white elephants in the Arizona desert:

After reading the article you referring to, I'm not sure what exactly the unintended consequences (plural) of the sanctions that Scott Sumner tried to point out?

 

[Fred Chen]

After reading the article you referring to, I'm not sure what exactly the unintended consequences (plural) of the sanctions that Scott Sumner tried to point out?

He might be referring to both TSMC Arizona and TSMC Taiwan being affected.

 

[blueone]

History documents that the US begins to compete best when its leaders and the general population become convinced we have fallen behind in a field of critical importance which can threaten our national security or economic leadership. Anyone who knows US history understands this is a national phobia which episodically reappears. Frustratingly, sometimes when the US achieves a strong leadership position we get lazy and stupid, so we get ourselves in a new predicament, sometimes in the same fields in which the phobia originally whipped us into action. (Space vehicle launch capability comes to mind.) We certainly have sat by while we lost leadership, or even high placement, in numerous areas of manufacturing, semiconductors being just one. Whenever I read Morris's statements lately, they make me chuckle. Keep on talking, Morris. Tell us how inept and inadequate we are. You just might help wake us up. For all of the hubbub around the CHIPS Act, IMO we still haven't really woken up about semiconductors yet.