Skip to main content

TSMC founder and former CEO says US chipmaking will be 'exercise in futility'

Intel Core i9-12900K
Intel Core i9-12900K (Image credit: Rich Edmonds / Windows Central)

What you need to know

  • TSMC's founder and former CEO Morris Chang believes that efforts to manufacture semiconductors in the United States will be a "very expensive exercise in futility."
  • Chang pointed to the difference in labor and manufacturing costs in the United States compared to Taiwan.
  • The U.S. government and several large tech companies plan to invest billions of dollars to chipmaking within the United States.

TSMC's founder and former CEO Morris Chang believes that semiconductor production in the United States will struggle to compete with manufacturing in Taiwan. Chang called recent efforts to increase semiconductor manufacturing in the U.S. a "very expensive exercise in futility." He argued that making chips in the U.S. will cost more, causing the U.S. industry to fail to compete with TSMC and other companies around the world.

Chang shared his thoughts at the Brookings Institution think tank (PDF) (via The Register).

The TSMC founder pointed to high manufacturing costs as the main reason U.S.-based semiconductor manufacturing will fail to compete. He highlighted the struggles of TSMC's factory in Oregon. "The same product, the Oregon cost, is about 50 percent more than the Taiwan cost," explained Chang.

It's worth noting that TSMC has plans for a massive chipmaking factory in Arizona. Chang notes that the decision to make that plant was made by the current chairman of TSMC, since Chang was almost retired at the time.

TSMC isn't the only organization to invest heavily in U.S.-based manufacturing. Intel announced plans for what it claims will be the "largest silicon manufacturing location on the planet" in January 2022. The company is also constructing chip factories in Arizona and investing heavily in domestic manufacturing of semiconductors.

The U.S government is considering a bill that will incentive semiconductor production within the United States that will provide over $50 billion in funding. These plans are part of an ongoing effort to have the U.S. compete with China in the semiconductor industry. The United States also has to compete with tech giants in Taiwan, such as TSMC.

These efforts will not succeed, according to Chang. "The recent effort of the U.S. to increase onshore manufacturing of semiconductors, right now you're talking about spending only tens of billions of dollars of money of subsidy," said Chang. "Well, it's not going to be enough. I think it will be a very expensive exercise in futility. The U.S. will increase onshore manufacturing of semiconductors somewhat. But all of that will be very high-cost increase, high unit cost. It will be noncompetitive in the world markets where you compete with factories like TSMC."

Sean Endicott
News Writer and apps editor

Sean Endicott brings nearly a decade of experience covering Microsoft and Windows news to Windows Central. He joined our team in 2017 as an app reviewer and now heads up our day-to-day news coverage. If you have a news tip or an app to review, hit him up at (opens in new tab).

  • Sounds like whistling past the graveyard.
    Semiconductor fabs are very energy intensive and we are moving into a two tier energy market with the NorthAM energy costs at $70 a barrel or less and the outside at double if they're lucky. He also seems to forget fabs are heavily automated so labor costs are a small part of the costs. Biggest part is the ASML EUV machines. TSMC is also subject to the never ending threats from China. And Taiwan has the second lowest fertility rate and has for over a decade. (China is lowest.) Cheap labor is a declining advantage.
    The US is near the top for the developed nations. Plus there's immigration, legal and otherwise. Times have changed since his days at the top.
  • Building a strong Airforce is also not profitable. We do it as a matter of national security. During covid, we had to rely on China to send us masks. Somebody asked the question, when it comes to issues of the health of our citizens, do we really want to totally rely on our enemy to be our only supplier? America doesn't need to be the leader in manufacturing ...but we do need to have some capacity and the ability to scale up in the event of a global situation without leaving us completely crippled and dependent.
  • "do we really want to totally rely on our enemy to be our only supplier?" Um, this isn't Newsmax. China still supplies the world with face masks. And US companies switched to making masks very quickly. Would have happened faster if Trump didn't dither, but that's another story. Also, Taiwan isn't China. Again, not Newsmax.
  • Newsmax? Hadn't heard of them. Might be worth checking out. 😇
    Try GEOPOP on youtube. Zeihan in particular brings up some very interesting (verifiable facts).
    Don't have to agree with his conclusions to appreciate there is far more going on that the media (of any stripe) reports.
  • Newsmax! Ironically rhymes with facts! Almost! As for GeoPop, honestly, I wouldn't take them too seriously either. It's great that YouTube gives journalism huge new opportunities, but not every channel is worth it.
  • The trick is to listen to everything and blindly believe nothing.
    Crosscheck and verify. Facts will emerge, bias and opinion will fade.
  • "He also seems to forget fabs are heavily automated so labor costs are a small part of the costs." Absolutely true. But most importantly, in a country with high labor costs (like the US), all kinds of businesses are going to be more capital intensive. It's the very definition of development. And the more expensive the labor, the more skilled it tends to be. It's not money thrown away. What *is* money thrown away is tax dollars given to huge, profitable companies who would set up in the US anyway because it makes business sense. Chang is actually (wrongly) making a case for government subsidies, since he thinks US-based factories *don't* make business sense!
  • They make sense today--they'll make a lot more sense by the time they're running.
    He also misses the point that the subsidies aren't to get the plants built--as you said, they woukd get built anyway-- but to ensure they plants are built to the next gen spec not to the current, barely started generation. That will ensure parity with whatever TSMC and SAMSUNG are doing with *their* government subsidies.
    What's good for the goose and all that... Fabs are a strategic chokepoint for many industrids much like energy and food and since the pandemic killed globalization (it's taking its time laying down in the grave) the new model is "build where you sell".
  • I don't think the subsidies will change much in terms of the investment decisions. They've already decided what kind of chip factory they want. Subsidies may make one place (very slightly) more lucrative than the next best option, but evidence on that is weak because companies will build where it makes sense for them (geography + infrastructure +local labor supply). Globalization isn't even close to dead. Seriously, GeoPop is not worth your time.
  • Look to the trains of stalled cargo ships and the companies quietly leaving China.
    The fertiizer shortages. Followed by food shortages. Those that have will hoard, those that don't will suffer.
    It's started. And Geopop is, as its name says, a pop source for geopolitical data.
    Think of it as a starting point for beginners. More serious questers will follow up the data and see where the winds blow. Try this: Look at china and taiwan, factor in tbe failure of zero covid and the closure of shanghai, and the oil crisis brewing. Any company relying on chinese manufacturing is in it deep. It won't get better soon. The ones with localized supply chains will thrive. Microsoft, for one, is all over: Mexico, Malaysia, US. They'll survive. The reasons for all the plants the TSMC founder frets over is precisely because everybody expects things to get better. And it's not just Semiconductors: Tesla Giga Berlin and GigaAustin are hedges against the Chinese problems. The next shoe to drop: Congress reinstates the old oil export restrictions to keep US oil production local. That used to be the case until very recently (circa 2017-ish). The US used to be the main proponent of globalization. Once US circles the wagons it all cracks. Seriously, there's a bigger game afoot than fab subsidies.
  • globalization is a farce, it never created a wide decentralized network of suppliers otherwise we wouldn’t be in this supply chain mess, it just shifted most of America’s manufacturing power to China. God forbid we ever do get into a hot war with them.
  • It costs TSMC more because they have to offer higher salaries to steal talent in Oregon and Arizona from Intel and other Semiconductor industries. They also do nothing to nurture talent from the College/Technical school level. Employees that have left are finding oppressive working conditions and almost instantly find the move not worth it all. They can't cane their workers in America for more productivity.
  • And after all of the money that US companies poured into offshore fabs, all for the sake of profit...
  • "all for the sake of profit..." Um, yes, they're called "for profit organizations."
  • Exactly.
    It made sense under the internationalist, Breton Woods global model.
    But that too is on its last legs.
    The future is not the past with a different calendar.
  • Seems like they're scared of more competition than anything. Why say that and then proceed to have multiple factories domestically and then say we cant or will fail.....really silly. So the US is only allowed to make offshore stuff or get from other people (according to tsmc)...?
  • Just remember that it's not according to the company. It's according to Chang, "TSMC's founder and former CEO." I'm sure he's had an illustrious career, but he sounds like a dinosaur now. As the article states, TSMC is building factories in the US, and they've never been doing better.
  • Too many taxes, and money is oriented for Walstreet.
  • Actually, no. it is headed for Austin.
    Not even Silivalley.
    Again, different times. Wall Street was the 60's just as SiliValley was the 80's.
    The 2030's belong to Texas, from Austin to Houston to Boca Chica.
  • Former says one thing, actual TSMC rulers build US plants.
    Of course they built facilities in America în exchange for "protection" against China,and by being coerced by imperialist corpo-banksters from US.
  • The "banksters" have no say in this. Global supply chains were *their* millieu.
    The next age belongs to engineers. Efficiency will be king. Hence, automation to the Nth. Gigafactories and Microfactories. AI all over to make up for the lack of warm bodies as the population explosion of tbe 60's peters out and global population begins to stabilize. The driving force behind de-globalization is the pandemic and, by now the endemic, Covid. Suppliers *need* to be close to their customers or risk being replaced by somebody closer come the inevitable shutdowns. It is all part ofliving with the waves of diseases headed our way. Covid is but the first of many.
    This was known 20 years ago but the IdiotPoliticians™ refused to believe it.
  • The banksters / corpo mafia have all the say in this,and basically ,in every aspect of this planet. You just chose not to believe it or aknowledge it. They who control the money supply,control your life,and mine,remotly.
    The whole society is in debt to them,a debt that cant ever be repaid. Open your eyes and stop fantasing about supply chains,and efficiency. The truth is we are Slaves. They dictate,and smaller fishes comply. TSMC is just a young tiny fish to the big whales who control money. Efficiency in capitalism ? Dont make me laugh. It's a system who encourage waste,planned obsolence,and continous massive consume,in order for it to keep breathing.
  • All sixties thinking, Rip. *This* is what the 21st century looks like: That is not what bankers create; that is what engineers with a free hand build.
    See is you can spot the (few) humans. Similarly, have you watched the teardown videos of XBOXes vs Playstations?
    There's a reason XBOXes are more available: the things are modularized and designfor automated assembly. Samsung does the same with their phones. Different age, different rules.
    Wall street is an afterthought.
  • I'm surprised he has forgotten one key thing - these fabs don't supply directly to consumers. They supply directly to companies who then have to compete in the global chain. Therefore in order to meet demand, these companies will swallow some of the cost. In regards to other costs, these will be offset by both increased unit cost and by other fabs already up and running (as no new wafer manufacturer can start up due to absurd costs, patents etc). As that effectively reduces the tax burden. Then there's the sleazy way off avoiding taxes. Finally, increased fabs will ultimately mean more demand for raw materials. Which by the conventional norms of supply and demand - prices will go up.
  • Both true, to a point.
    Once demand and prices go up enough, alternatives will be found. And once those are mainstreamed, the associated ramp up costs will be covered and prices will stabilize. EV batteries are a good example: lithium is abundant but most sites weren't profitable at traditional demand levels. At new er procrs, places like the Dalton sea become viable. The batteries use nickel which is getting expensive, so companies are using iron instead. Sodium batteries are next. Other chemistries are moving from labs to production.
    High prices are great drivers of change. And one of the most critical "raw materials" is energy. And Texas is one of the worlds biggest energy producers. Those plants are going to Texas (instead of SiliValley) because of this. One factoid about the newest tech many forget: moving to small processes means smaller chips. This applies not just to monster SOC's but to more mundane chips. More chips to the wafer offsets the materials costs a bit. Also (and we're seing this in Austin and Arizona) suppliers will set up shop near the fabs. So subsidizing a fab brings in other, non-subsidized, plants. Like for making silicon. Or parts for the horrendously complicated ASML EUV lithography boxes. Lost in the panic over Fabs is the bottlenecks ASML deals with. And a string majority of the critical parts ASML needs are american. And ASML is borced to grow to meet the demand for EUV. They might have to supplement their assembly facility in the Netherlands with another in the US. TBD, that. Supply chains run both ways. Suppliers and consumers.
    Getting close to consumers adds value. Increasingly so.
    Getting close to critical suppliers too.
    And if you have a place that does both, you go there. Those subsidies have good logic behind tbem. It's just 2025 logic, not 2005 logic.
  • Isn't it just a matter of time before China is dropping bombs on their factories? They will not be much competition after that happens.
  • Not China.
    One reason China is looking to invade is TSMC.
    But yes, those fabs are a high value target in a Taiwan conflict.
    Would only take a few TOMAHAWKS or a single special forces opertion to ensure China doesn't get them. Geopolitics is cold hearted.