Why Trump’s tech dreams are quietly crumbling

In an opinion column, Gillian Tett argues that the Trump administration's push to bring chip production back to the US is faltering, largely because of a shortage of engineers. Former TSMC chief scientist Philip Wong, now at Stanford, told her only 17% of his former PhD students work at chip-making companies; two-thirds of his graduates are non-American, and most stayed in the US. The White House has doubled down on the 2022 Chips Act (which Trump called 'horrible, horrible' in 2025), taken a 10% government stake in Intel and imposed big tariffs on chipmaking equipment, and TSMC has pledged more than $250bn of new fab investment. But the industry body CCIA has warned Trump that tariffs could raise the cost of building fabs and the AI supply chain by $450bn between 2026 and 2030, a McKinsey-backed industry report puts the engineer shortfall at up to 157,000 ('Engineers represent the binding constraint'), and a congressional report says labour and supply-chain shortages forced TSMC and Micron to delay fabs in Arizona and New York. Visa curbs and the expulsion of some South Korean workers add to the problem, in her view. Counterweights she includes: US strengths remain innovation, chip design and deployment backed by deep capital markets, and some efforts are under way, such as Arizona community colleges training engineers. She suggests accepting that supply chains are global and championing skilled immigration. The column runs days before Tuesday's midterms; it is Tett's argument, not a policy assessment.