Few would dispute it is in the best interest of the United States, for both economic and national security reasons, to build a sizeable domestic semiconductor manufacturing capacity. But despite lofty promises, Washington is falling far behind in its commitment to this mission, which is set to become a total failure if serious changes are not made.
At the center of this is the CHIPS and Science Act, unveiled in 2022 with much fanfare and passed by Congress with strong support from the Biden administration. The act was supposed to build (or rebuild) U.S. indigenous semiconductor manufacturing so that any disruption or suffocation of the global chip supply chain would not endanger the American economy and industry.
The urgency is real for America with risks involved. Taiwan’s semiconductor giant TSMC produces more than half the world’s chips. With clients including Apple and Nvidia, its chips are used in everything from vehicles and smartphones to supercomputer AI.
As tensions and the risk of war between China and Taiwan increase, pundits and industry experts alike warn that near-total dependency on Taiwan-based productions could put the American economy and security at serious risk.
Despite TSMC being the crown jewel of its economy, Taiwan under the Washington-appeasing Tsai Ing-wen government, greenlighted TSMC to move some productions overseas — a move that generated considerable criticism within Taiwan. TSMC started the construction of a major plant in Phoenix, Ariz. in 2021.
The plant would produce 5 nanometer, or 5nm, chips; among the world’s most advanced chips, only behind TSMC’s flagship 3nm back home in Taiwan. TSMC increased the scale of its investment in 2022 to $40 billion in response to the CHIPS Act and promising signals from Washington.
But just three years later, TSMC’s plant in Arizona is lagging. Not only has it completely missed its originally promised construction completion mark of 2024, but reportedly, volume production of the advanced chips won’t start until at least 2027 or 2028.
On the surface, the media explained the Arizona plant’s repeated delays and troubles as a case of a foreign company failing to acclimate to the American labor environment and work culture. TSMC management and Chairman Mark Liu also noted that American workers were averse to working overtime and could not work as hard as their Taiwanese counterparts. It is also alleged that American workers failed to adhere to the strict and demanding schedules of the chip production environment.
But such discussion touches on only the surface and not the root causes. TSMC’s setback in America had as much to do with U.S. political leadership’s failure to put their money (and heart) where their mouth is.
As of 2024, zero grants from the CHIPS Act have been awarded to TSMC despite it being the one company that all observers agreed matters the most. In fact, only two projects by small American companies have received tiny grants from the U.S. government as part of the CHIPS Act, neither of which makes advanced chips like TSMC.
This is a far cry from the CHIPS Act’s sponsors’ original promise that the U.S. would mobilize all resources available to chip makers who can pump out their products on American soil. The lackluster support to TSMC proves Washington seems interested in supporting only token projects that do not meaningfully contribute to the mission of chip self-sufficiency.
In comparison, TSMC’s plants in other countries are far more successful. In February, TSMC opened the first stage of its new plant in Kumamoto, Japan. Despite starting a full year behind the Arizona plant in 2022, the Kumamoto plant was built in under two years and is now operational, with a second-stage plant beginning construction this year, which would feature production of 6/7nm chips on par with what TSMC’s Arizona plant promised.
TSMC’s success in Kumamoto is rooted in the Japanese government’s firm commitment to supporting the project. Tokyo has already given billions in U.S. dollars subsidizing TSMC for the first Kumamoto plant and is on track to subsidize billions more for its second plant, the construction of which is already well underway.
The Kumamoto plant received whole-of-government support from not just Tokyo but also the city government, which subsidized the construction of infrastructures and surrounding environmental necessities. This is a sharp contrast to Arizona’s TSMC plant, which has received little support from the state government.
An unconfirmed report from Bloomberg earlier this month claimed Washington is finally set to give TSMC the first taste of subsidies reportedly around $5 billion. This is a step in the right direction, but even if all of these reported subsidies are actually delivered, they would still fall short of the $8.2 billion the Japanese government gave to TSMC’s plants in Kumamoto.
Given that the U.S. CHIPS Act authorized a staggering $280 billion in funding, one must wonder why the administration has been so meager in funding what should have been the project’s flagship program.
If the U.S. is serious about building chip self-sufficiency, it must invest serious commitment and do more than talk big about promises and vision. It should be noted Japan is not the only country besides the U.S. that has wooed TSMC to its door; Germany is also on track to subsidize TSMC to build its first plant in Europe.
Time is running out for the U.S. in the competition for chips, and the consequences of losing would be catastrophic economically and for U.S. national security interests.
Paul Huang is a research fellow at the Taiwan Public Opinion Foundation and a visiting fellow at Fairbank Center for Chinese Studies at Harvard University. Follow him on X @PaulHuangReport.
Jason Hsu is an Edward Mason Fellow at Harvard Kennedy School at Harvard University and a former legislator of Taiwan. Follow him on X @augama.