It took time for America’s semiconductor industry to implode. It floated. Over the course of three decades, chip manufacturing quietly relocated offshore as businesses sought lower labor costs, more favorable tax regimes, and lower costs. The United States used to produce almost 40% of the world’s semiconductors, but by the time Washington began to pay attention, that percentage had dropped to just 12%. In the meantime, Taiwan was manufacturing over 90% of the world’s most sophisticated chips. When COVID-19 revealed the true fragility of global supply chains, that imbalance—which for years seemed like a far-off economic issue—suddenly appeared to be a national security crisis.
Washington’s response to that realization was the CHIPS and Science Act, which President Biden signed into law in August 2022. It was one of the biggest industrial policy initiatives the United States had tried in decades, with about $280 billion in authorized funding and $52.7 billion actually appropriated. Offering grants, tax credits, and subsidies to entice chipmakers to locate their operations in the United States was the basic idea. As it happened, the execution would be much messier.
It’s difficult to avoid experiencing a mixture of anxiety and ambition when looking at the numbers. The legislation set aside $13 billion for workforce development and research, $39 billion for direct manufacturing incentives, and a 25 percent tax credit for qualified capital expenditures. It approved billions more for STEM education, biotechnology, clean energy research, and quantum computing in addition to semiconductors. It had an impact on NASA. The Department of Energy was affected. The National Science Foundation was affected. To be honest, it’s still unclear whether all those threads can be connected in a logical way.
The initial outcomes were noisy. Companies began announcing investments at an almost unbelievable rate within months of the passage. Intel pledged to invest up to $100 billion in Ohio. The largest foreign direct investment in a greenfield project in American history, TSMC, the Taiwanese behemoth that produces chips for Apple, AMD, Nvidia, and almost everyone else, announced it was increasing its investment in Arizona from $12 billion to $40 billion. A $20 billion factory close to Syracuse was announced by Micron. The announcements continued to come in, adding up to what appeared to be a true industrial revival on paper.
And yet. According to TSMC, construction costs in the United States were four to five times higher than in Taiwan. There were few skilled workers. TSMC postponed opening in Arizona from 2026 to 2028, citing a variety of reasons, including deep cultural differences between its American hires and Taiwanese engineers and bureaucratic friction. Subcontractors were accused of stealing wages, safety procedures were neglected, and union organizers and a business more used to working in a completely different labor environment were at odds. Rebuilding a semiconductor industry isn’t quite as simple as flipping a switch, despite the fact that these issues weren’t necessarily fatal.
Additionally, the CHIPS Act received criticism from unanticipated sources. It’s a blank check for corporations, according to Bernie Sanders. The effective dominance of TSMC and a few large fabless companies over the entire American chip market, according to an antitrust think tank, was not challenged. Environmental analysts noted that the law made remarkably little mention of the massive amounts of water and chemical waste produced during the chip manufacturing process. The Act’s approach to labor development, according to Brookings workforce researchers, was disjointed and inadequately supportive of the communities that would actually surround these facilities.
As all of this was happening, Europe swiftly responded with its own version, the European Chips Act, allocating about €43 billion by the end of the decade. Since then, the EU has proposed a Chips Act 2.0 with the goals of strengthening relationships between manufacturers and the industries that actually use chips, reducing reliance on outside suppliers, and expediting permitting to within a year. The global competition for semiconductor capacity is undoubtedly getting more intense, though it’s unclear if Europe can close its own gap.
Beneath all the complications, there is something truly important going on here. Apple chips are reportedly being produced at a TSMC facility in Phoenix with yields that are higher than those of some Taiwan facilities. This information would have seemed nearly unimaginable a short time ago. Thousands of research grants and billions of dollars in private funding have been made possible by the NSF’s recently established Technology, Innovation and Partnerships Directorate. Purdue established the nation’s first degree program specifically focused on semiconductors. These are not minor issues.
It’s still unclear if the CHIPS Act will be remembered as the time when America successfully regained its industrial footing or as a well-meaning endeavor that fell short of what policy could accomplish on its own. The number of factories is increasing. The money is coming in. However, the difficulties are genuine.
