This week, Huang Hansen ( Philip Wong ), who once served as TSMC's chief scientist and is now a professor at Stanford University, compiled statistics on the career paths of his former doctoral students.
He found that only 17% of the students went on to work in companies that actually manufacture chips, such as Intel, Micron, Samsung, SK Hynix, and TSMC; the majority of the rest work in chip design or chip application companies, such as Apple, NVIDIA, AMD, Google, Pinterest, or other institutions like McKinsey.
“They all possess the capability to work in dust-free workshops and start from a blank silicon wafer to produce usable transistors or memory devices,” he said, “but most people ultimately chose other industries.” As for the reasons, he explained that it’s because “the salaries are higher, there are better career development paths, greater job mobility, and the working environment is also more attractive.”
He pointed out that in places like South Korea, semiconductor manufacturing companies are actually the most favored employers by job seekers.
Well-known columnist Gillian Tett believes that American politicians should pay attention to this phenomenon before the mid-term elections on Tuesday; investors who have been distracted by the performance of tech stocks and the upcoming Anthropic listings also need to be mindful.
After Trump returned to power, he promised to relocate chip production back to the United States in order to support the development of AI. Tett believes that, in theory, this goal sounds very promising.
The White House has continued the "Chip Act" introduced by the Democratic Party in 2022, providing subsidies for domestic chip production, despite Trump describing this act as "terrible" in 2025. The U.S. government has also taken a stake in Intel, holding a 10% share, and at the same time imposes high import tariffs on semiconductor production equipment to encourage domestic factory construction.
At the same time, companies such as TSMC have pledged to make additional investments of over $250 billion in wafer factory projects in the United States. Industrial policy seems to be back in vogue, which is also a point that Trump has heavily promoted. However, research by economists Ian Fletcher and Marc Fasteau indicates that for industrial policy to be effective, it must possess synergy, integrity, and longevity; Trump's policy lacks all three of these elements.
Looking at infrastructure again, American officials stated at this week's Global Chips Conference held at the University of Oxford that they plan to improve the supporting infrastructure. However, the event was also filled with numerous complaints: insufficient power supply, severe delays in approval processes, and a huge shortage of workers in wafer factories.
A recent report jointly authored by McKinsey and industry organizations warns that engineers represent the biggest bottleneck, with a total labor shortage of up to 157,000 people. The report states, "If targeted intervention measures are not taken, the talent shortage could lead to delays in the commissioning of new wafer factories." A congressional report also mentioned that shortages in labor and the supply chain have forced TSMC and Micron to postpone their previously announced plans to build factories in Arizona and New York.
Tett has analyzed several feasible solutions based on this.
The first approach is to emulate Asia by implementing wise, long-term, and coherent industrial policies. However, this conflicts with traditional American concepts and the federal governance structure. Taking Arizona as an example, local officials are making efforts to cultivate engineers through community colleges, but it is understood that coordinating competing chip companies is extremely difficult.
The second approach is to recruit overseas engineers. Reports indicate that this channel has already been operating in a low-profile manner. However, Trump's base strongly opposes this practice; the government has not only tightened visa regulations but also deported some Korean technical workers, which is quite unusual.
The third approach is to encourage foreign engineers studying at American universities to stay in the U.S. for their career development after graduation. This is indeed happening quietly: among Huang Hansen's doctoral students, two-thirds are not American citizens, but the majority of them choose to remain in the U.S. A report from the University of California, Berkeley, indicates that this phenomenon is quite common, and these international students contribute significantly to America's technological innovation. However, the MAGA camp also resists this path.
Thus, there remains the fourth option: The White House needs to recognize an obvious reality—technology supply chains have long been highly globalized, and achieving complete self-sufficiency is almost impossible; it should openly support the immigration of highly skilled talents, cross-border trade, and cooperation with allies.
Responsible Editor: Chen Yujia
